X-Git-Url: https://git.whamcloud.com/?p=fs%2Flustre-release.git;a=blobdiff_plain;f=lustre%2Ftests%2Fsanity.sh;h=6aadd4f7da555f1337eb1b992ca629f307519852;hp=4e7ee5436c39e5493ce1c0c9daa19c8bf9072c3f;hb=e99b9dfa326292514eae0dbf4a96b3fb3855dceb;hpb=4cc21ecbcee4d29017fe5d400f3a88321d65d58d diff --git a/lustre/tests/sanity.sh b/lustre/tests/sanity.sh index 4e7ee54..6aadd4f 100644 --- a/lustre/tests/sanity.sh +++ b/lustre/tests/sanity.sh @@ -7,8 +7,8 @@ set -e ONLY=${ONLY:-"$*"} -# bug number for skipped test: 2108 9789 3637 9789 3561 12653 12653 13310 10764 -ALWAYS_EXCEPT=" 42a 42b 42c 42d 45 65a 65e 74b 75 $SANITY_EXCEPT" +# bug number for skipped test: 13297 2108 9789 3637 9789 3561 12622 12653 12653 5188 10764 16260 +ALWAYS_EXCEPT=" 27u 42a 42b 42c 42d 45 51d 65a 65e 68b 75 119d $SANITY_EXCEPT" # bug number for skipped test: 2108 9789 3637 9789 3561 5188/5749 1443 #ALWAYS_EXCEPT=${ALWAYS_EXCEPT:-"27m 42a 42b 42c 42d 45 68 76"} # UPDATE THE COMMENT ABOVE WITH BUG NUMBERS WHEN CHANGING ALWAYS_EXCEPT! @@ -27,8 +27,8 @@ case `uname -r` in *) error "unsupported kernel" ;; esac -SRCDIR=`dirname $0` -export PATH=$PWD/$SRCDIR:$SRCDIR:$SRCDIR/../utils:$PATH:/sbin +SRCDIR=$(cd $(dirname $0); echo $PWD) +export PATH=$PATH:/sbin TMP=${TMP:-/tmp} @@ -45,8 +45,7 @@ LCTL=${LCTL:-lctl} MCREATE=${MCREATE:-mcreate} OPENFILE=${OPENFILE:-openfile} OPENUNLINK=${OPENUNLINK:-openunlink} -RANDOM_READS=${RANDOM_READS:-"random-reads"} -TOEXCL=${TOEXCL:-toexcl} +READS=${READS:-"reads"} TRUNCATE=${TRUNCATE:-truncate} MUNLINK=${MUNLINK:-munlink} SOCKETSERVER=${SOCKETSERVER:-socketserver} @@ -68,7 +67,7 @@ SAVE_PWD=$PWD CLEANUP=${CLEANUP:-:} SETUP=${SETUP:-:} TRACE=${TRACE:-""} -LUSTRE=${LUSTRE:-`dirname $0`/..} +LUSTRE=${LUSTRE:-$(cd $(dirname $0)/..; echo $PWD)} . $LUSTRE/tests/test-framework.sh init_test_env $@ . ${CONFIG:=$LUSTRE/tests/cfg/${NAME}.sh} @@ -110,7 +109,7 @@ fi check_and_setup_lustre DIR=${DIR:-$MOUNT} -[ -z "`echo $DIR | grep $MOUNT`" ] && echo "$DIR not in $MOUNT" && exit 99 +assert_DIR LOVNAME=`lctl get_param -n llite.*.lov.common_name | tail -n 1` OSTCOUNT=`lctl get_param -n lov.$LOVNAME.numobd` @@ -126,11 +125,11 @@ rm -rf $DIR/[Rdfs][0-9]* # $RUNAS_ID may get set incorrectly somewhere else [ $UID -eq 0 -a $RUNAS_ID -eq 0 ] && error "\$RUNAS_ID set to 0, but \$UID is also 0!" -check_runas_id $RUNAS_ID $RUNAS +check_runas_id $RUNAS_ID $RUNAS_ID $RUNAS build_test_filter -if [ "${ONLY}" = "MOUNT" ] ; then +if [ "${ONLY}" = "MOUNT" ] ; then echo "Lustre is up, please go on" exit fi @@ -143,8 +142,8 @@ echo # add a newline after mke2fs. umask 077 -OLDDEBUG="`sysctl -n lnet.debug 2> /dev/null`" -sysctl -w lnet.debug=-1 2> /dev/null || true +OLDDEBUG="`lctl get_param -n debug 2> /dev/null`" +lctl set_param debug=-1 2> /dev/null || true test_0() { touch $DIR/$tfile $CHECKSTAT -t file $DIR/$tfile || error @@ -159,6 +158,12 @@ test_0b() { } run_test 0b "chmod 0755 $DIR =============================" +test_0c() { + $LCTL get_param mdc.*.import | grep "state: FULL" || error "import not FULL" + $LCTL get_param mdc.*.import | grep "target: $FSNAME-MDT" || error "bad target" +} +run_test 0c "check import proc =============================" + test_1a() { mkdir $DIR/d1 mkdir $DIR/d1/d2 @@ -458,6 +463,14 @@ test_17f() { } run_test 17f "symlinks: long and very long symlink name ========================" +test_17g() { + mkdir -p $DIR/$tdir + LONGSYMLINK="$(dd if=/dev/zero bs=4095 count=1 | tr '\0' 'x')" + ln -s $LONGSYMLINK $DIR/$tdir/$tfile + ls -l $DIR/$tdir +} +run_test 17g "symlinks: really long symlink name ===============================" + test_18() { touch $DIR/f ls $DIR || error @@ -515,6 +528,7 @@ run_test 21 "write to dangling link ============================" test_22() { WDIR=$DIR/$tdir + mkdir -p $WDIR chown $RUNAS_ID $WDIR (cd $WDIR || error "cd $WDIR failed"; $RUNAS tar cf - /etc/hosts /etc/sysconfig/network | \ @@ -526,9 +540,12 @@ test_22() { run_test 22 "unpack tar archive as non-root user ===============" test_23() { - mkdir $DIR/d23 - $TOEXCL $DIR/d23/f23 - $TOEXCL -e $DIR/d23/f23 || error + mkdir -p $DIR/$tdir + local file=$DIR/$tdir/$tfile + + openfile -f O_CREAT:O_EXCL $file || error "$file create failed" + openfile -f O_CREAT:O_EXCL $file && + error "$file recreate succeeded" || true } run_test 23 "O_CREAT|O_EXCL in subdir ==========================" @@ -570,7 +587,7 @@ test_24d() { run_test 24d "mkdir .../R4/{f,g}; rename .../R4/f .../R4/g =====" test_24e() { - echo '-- cross directory renames --' + echo '-- cross directory renames --' mkdir $DIR/R5{a,b} touch $DIR/R5a/f mv $DIR/R5a/f $DIR/R5b/g @@ -625,7 +642,7 @@ test_24j() { $CHECKSTAT -a $DIR/R10/f || error $CHECKSTAT -a $DIR/R10/g || error } -run_test 24j "source does not exist ============================" +run_test 24j "source does not exist ============================" test_24k() { mkdir $DIR/R11a $DIR/R11a/d @@ -781,6 +798,7 @@ run_test 26e "unlink multiple component recursive symlink ======" # recursive symlinks (bug 7022) test_26f() { + mkdir -p $DIR/$tdir mkdir $DIR/$tdir/$tfile || error "mkdir $DIR/$tdir/$tfile failed" cd $DIR/$tdir/$tfile || error "cd $DIR/$tdir/$tfile failed" mkdir -p lndir/bar1 || error "mkdir lndir/bar1 failed" @@ -821,7 +839,7 @@ run_test 27c "create two stripe file f01 =======================" test_27d() { mkdir -p $DIR/d27 - $SETSTRIPE $DIR/d27/fdef 0 -1 0 || error "lstripe failed" + $SETSTRIPE -c0 -i-1 -s0 $DIR/d27/fdef || error "lstripe failed" $CHECKSTAT -t file $DIR/d27/fdef || error "checkstat failed" dd if=/dev/zero of=$DIR/d27/fdef bs=4k count=4 || error } @@ -917,25 +935,32 @@ reset_enospc() { [ "$1" ] && FAIL_LOC=$1 || FAIL_LOC=0 mkdir -p $DIR/d27/nospc rmdir $DIR/d27/nospc - sysctl -w lustre.fail_loc=$FAIL_LOC + do_nodes $(comma_list $(osts_nodes)) lctl set_param fail_loc=$FAIL_LOC } exhaust_precreations() { - OSTIDX=$1 - OST=$(lctl get_param -n lov.${LOVNAME}.target_obd | - grep ${OSTIDX}": " | \ - awk '{print $2}' | sed -e 's/_UUID$//') + local OSTIDX=$1 + local MDSIDX=$(get_mds_dir "$DIR/d27") + echo OSTIDX=$OSTIDX MDSIDX=$MDSIDX + + local OST=$(lfs osts | grep ${OSTIDX}": " | awk '{print $2}' | sed -e 's/_UUID$//') + local MDT_INDEX=$(lfs df | grep "\[MDT:$((MDSIDX - 1))\]" | awk '{print $1}' | \ + sed -e 's/_UUID$//;s/^.*-//') + # on the mdt's osc - last_id=$(lctl get_param -n osc.*${OST}-osc-MDT0000.prealloc_last_id) - next_id=$(lctl get_param -n osc.*${OST}-osc-MDT0000.prealloc_next_id) + local last_id=$(do_facet mds${MDSIDX} lctl get_param -n osc.*${OST}-osc-${MDT_INDEX}.prealloc_last_id) + local next_id=$(do_facet mds${MDSIDX} lctl get_param -n osc.*${OST}-osc-${MDT_INDEX}.prealloc_next_id) - mkdir -p $DIR/d27/${OST} - $SETSTRIPE $DIR/d27/${OST} -i $OSTIDX -c 1 + echo ${OST}-osc-${MDT_INDEX}.prealloc_last_id=$last_id + echo ${OST}-osc-${MDT_INDEX}.prealloc_next_id=$next_id + + mkdir -p $DIR/d27 + $SETSTRIPE $DIR/d27 -i $OSTIDX -c 1 #define OBD_FAIL_OST_ENOSPC 0x215 - sysctl -w lustre.fail_loc=0x215 + do_facet ost$((OSTIDX + 1)) lctl set_param fail_loc=0x215 echo "Creating to objid $last_id on ost $OST..." - createmany -o $DIR/d27/${OST}/f $next_id $((last_id - next_id + 2)) - lctl get_param -n osc.*${OST}-osc-MDT0000.prealloc* | grep '[0-9]' + createmany -o $DIR/d27/${OST}-f $next_id $((last_id - next_id + 2)) + do_facet mds${MDSIDX} lctl get_param osc.*${OST}-osc-${MDT_INDEX}.prealloc* | grep '[0-9]' reset_enospc $2 } @@ -949,7 +974,8 @@ exhaust_all_precreations() { test_27n() { [ "$OSTCOUNT" -lt "2" ] && skip "too few OSTs" && return - remote_mds && skip "remote MDS" && return + remote_mds_nodsh && skip "remote MDS with nodsh" && return + remote_ost_nodsh && skip "remote OST with nodsh" && return reset_enospc rm -f $DIR/d27/f27n @@ -963,7 +989,8 @@ run_test 27n "create file with some full OSTs ==================" test_27o() { [ "$OSTCOUNT" -lt "2" ] && skip "too few OSTs" && return - remote_mds && skip "remote MDS" && return + remote_mds_nodsh && skip "remote MDS with nodsh" && return + remote_ost_nodsh && skip "remote OST with nodsh" && return reset_enospc rm -f $DIR/d27/f27o @@ -979,18 +1006,19 @@ run_test 27o "create file with all full OSTs (should error) ====" test_27p() { [ "$OSTCOUNT" -lt "2" ] && skip "too few OSTs" && return - remote_mds && skip "remote MDS" && return + remote_mds_nodsh && skip "remote MDS with nodsh" && return + remote_ost_nodsh && skip "remote OST with nodsh" && return reset_enospc rm -f $DIR/d27/f27p - $MCREATE $DIR/d27/f27p || error - $TRUNCATE $DIR/d27/f27p 80000000 || error - $CHECKSTAT -s 80000000 $DIR/d27/f27p || error + $MCREATE $DIR/d27/f27p || error "mcreate failed" + $TRUNCATE $DIR/d27/f27p 80000000 || error "truncate failed" + $CHECKSTAT -s 80000000 $DIR/d27/f27p || error "checkstat failed" exhaust_precreations 0 0x80000215 - echo foo >> $DIR/d27/f27p || error - $CHECKSTAT -s 80000004 $DIR/d27/f27p || error + echo foo >> $DIR/d27/f27p || error "append failed" + $CHECKSTAT -s 80000004 $DIR/d27/f27p || error "checkstat failed" reset_enospc } @@ -998,7 +1026,8 @@ run_test 27p "append to a truncated file with some full OSTs ===" test_27q() { [ "$OSTCOUNT" -lt "2" ] && skip "too few OSTs" && return - remote_mds && skip "remote MDS" && return + remote_mds_nodsh && skip "remote MDS with nodsh" && return + remote_ost_nodsh && skip "remote OST with nodsh" && return reset_enospc rm -f $DIR/d27/f27q @@ -1018,7 +1047,8 @@ run_test 27q "append to truncated file with all OSTs full (should error) ===" test_27r() { [ "$OSTCOUNT" -lt "2" ] && skip "too few OSTs" && return - remote_mds && skip "remote MDS" && return + remote_mds_nodsh && skip "remote MDS with nodsh" && return + remote_ost_nodsh && skip "remote OST with nodsh" && return reset_enospc rm -f $DIR/d27/f27r @@ -1048,18 +1078,17 @@ test_27t() { # bug 10864 run_test 27t "check that utils parse path correctly" test_27u() { # bug 4900 - [ "$OSTCOUNT" -lt "2" ] && skip "too few OSTs" && return - remote_mds && skip "remote MDS" && return - - #define OBD_FAIL_MDS_OSC_PRECREATE 0x13d + [ "$OSTCOUNT" -lt "2" ] && skip "too few OSTs" && return + remote_mds_nodsh && skip "remote MDS with nodsh" && return - sysctl -w lustre.fail_loc=0x13d +#define OBD_FAIL_MDS_OSC_PRECREATE 0x139 + do_facet $SINGLEMDS lctl set_param fail_loc=0x139 mkdir -p $DIR/d27u createmany -o $DIR/d27u/t- 1000 - sysctl -w lustre.fail_loc=0 + do_facet $SINGLEMDS lctl set_param fail_loc=0 TLOG=$DIR/$tfile.getstripe - $GETSTRIPE $DIR/d27u > $TLOG + $GETSTRIPE $DIR/d27u > $TLOG OBJS=`awk -vobj=0 '($1 == 0) { obj += 1 } END { print obj;}' $TLOG` unlinkmany $DIR/d27u/t- 1000 [ $OBJS -gt 0 ] && \ @@ -1069,7 +1098,8 @@ run_test 27u "skip object creation on OSC w/o objects ==========" test_27v() { # bug 4900 [ "$OSTCOUNT" -lt "2" ] && skip "too few OSTs" && return - remote_mds && skip "remote MDS" && return + remote_mds_nodsh && skip "remote MDS with nodsh" && return + remote_ost_nodsh && skip "remote OST with nodsh" && return exhaust_all_precreations @@ -1078,15 +1108,15 @@ test_27v() { # bug 4900 touch $DIR/$tdir/$tfile #define OBD_FAIL_TGT_DELAY_PRECREATE 0x705 - sysctl -w lustre.fail_loc=0x705 + lctl set_param fail_loc=0x705 START=`date +%s` for F in `seq 1 32`; do touch $DIR/$tdir/$tfile.$F done - sysctl -w lustre.fail_loc=0 + lctl set_param fail_loc=0 FINISH=`date +%s` - TIMEOUT=`sysctl -n lustre.timeout` + TIMEOUT=`lctl get_param -n timeout` [ $((FINISH - START)) -ge $((TIMEOUT / 2)) ] && \ error "$FINISH - $START >= $TIMEOUT / 2" @@ -1112,7 +1142,9 @@ test_27w() { # bug 10997 } run_test 27w "check lfs setstripe -c -s -i options =============" -test_28() { +# createtest also checks that device nodes are created and +# then visible correctly (#2091) +test_28() { # bug 2091 mkdir $DIR/d28 $CREATETEST $DIR/d28/ct || error } @@ -1124,14 +1156,32 @@ test_29() { touch $DIR/d29/foo log 'first d29' ls -l $DIR/d29 - LOCKCOUNTORIG=`lctl get_param -n ldlm.namespaces.*mdc*.lock_count` - LOCKUNUSEDCOUNTORIG=`lctl get_param -n ldlm.namespaces.*mdc*.lock_unused_count` - [ -z $"LOCKCOUNTORIG" ] && echo "No mdc lock count" && return 1 + + declare -i LOCKCOUNTORIG=0 + for lock_count in $(lctl get_param -n ldlm.namespaces.*mdc*.lock_count); do + let LOCKCOUNTORIG=$LOCKCOUNTORIG+$lock_count + done + [ $LOCKCOUNTORIG -eq 0 ] && echo "No mdc lock count" && return 1 + + declare -i LOCKUNUSEDCOUNTORIG=0 + for unused_count in $(lctl get_param -n ldlm.namespaces.*mdc*.lock_unused_count); do + let LOCKUNUSEDCOUNTORIG=$LOCKUNUSEDCOUNTORIG+$unused_count + done + log 'second d29' ls -l $DIR/d29 log 'done' - LOCKCOUNTCURRENT=`lctl get_param -n ldlm.namespaces.*mdc*.lock_count` - LOCKUNUSEDCOUNTCURRENT=`lctl get_param -n ldlm.namespaces.*mdc*.lock_unused_count` + + declare -i LOCKCOUNTCURRENT=0 + for lock_count in $(lctl get_param -n ldlm.namespaces.*mdc*.lock_count); do + let LOCKCOUNTCURRENT=$LOCKCOUNTCURRENT+$lock_count + done + + declare -i LOCKUNUSEDCOUNTCURRENT=0 + for unused_count in $(lctl get_param -n ldlm.namespaces.*mdc*.lock_unused_count); do + let LOCKUNUSEDCOUNTCURRENT=$LOCKUNUSEDCOUNTCURRENT+$unused_count + done + if [ "$LOCKCOUNTCURRENT" -gt "$LOCKCOUNTORIG" ]; then lctl set_param -n ldlm.dump_namespaces "" error "CURRENT: $LOCKCOUNTCURRENT > $LOCKCOUNTORIG" @@ -1306,27 +1356,27 @@ run_test 31m "link to file: the same, non-existing, dir===============" test_32a() { echo "== more mountpoints and symlinks =================" [ -e $DIR/d32a ] && rm -fr $DIR/d32a - mkdir -p $DIR/d32a/ext2-mountpoint + mkdir -p $DIR/d32a/ext2-mountpoint mount -t ext2 -o loop $EXT2_DEV $DIR/d32a/ext2-mountpoint || error - $CHECKSTAT -t dir $DIR/d32a/ext2-mountpoint/.. || error + $CHECKSTAT -t dir $DIR/d32a/ext2-mountpoint/.. || error $UMOUNT $DIR/d32a/ext2-mountpoint || error } run_test 32a "stat d32a/ext2-mountpoint/.. =====================" test_32b() { [ -e $DIR/d32b ] && rm -fr $DIR/d32b - mkdir -p $DIR/d32b/ext2-mountpoint + mkdir -p $DIR/d32b/ext2-mountpoint mount -t ext2 -o loop $EXT2_DEV $DIR/d32b/ext2-mountpoint || error ls -al $DIR/d32b/ext2-mountpoint/.. || error $UMOUNT $DIR/d32b/ext2-mountpoint || error } run_test 32b "open d32b/ext2-mountpoint/.. =====================" - + test_32c() { [ -e $DIR/d32c ] && rm -fr $DIR/d32c - mkdir -p $DIR/d32c/ext2-mountpoint + mkdir -p $DIR/d32c/ext2-mountpoint mount -t ext2 -o loop $EXT2_DEV $DIR/d32c/ext2-mountpoint || error - mkdir -p $DIR/d32c/d2/test_dir + mkdir -p $DIR/d32c/d2/test_dir $CHECKSTAT -t dir $DIR/d32c/ext2-mountpoint/../d2/test_dir || error $UMOUNT $DIR/d32c/ext2-mountpoint || error } @@ -1334,9 +1384,9 @@ run_test 32c "stat d32c/ext2-mountpoint/../d2/test_dir =========" test_32d() { [ -e $DIR/d32d ] && rm -fr $DIR/d32d - mkdir -p $DIR/d32d/ext2-mountpoint + mkdir -p $DIR/d32d/ext2-mountpoint mount -t ext2 -o loop $EXT2_DEV $DIR/d32d/ext2-mountpoint || error - mkdir -p $DIR/d32d/d2/test_dir + mkdir -p $DIR/d32d/d2/test_dir ls -al $DIR/d32d/ext2-mountpoint/../d2/test_dir || error $UMOUNT $DIR/d32d/ext2-mountpoint || error } @@ -1344,10 +1394,10 @@ run_test 32d "open d32d/ext2-mountpoint/../d2/test_dir =========" test_32e() { [ -e $DIR/d32e ] && rm -fr $DIR/d32e - mkdir -p $DIR/d32e/tmp - TMP_DIR=$DIR/d32e/tmp - ln -s $DIR/d32e $TMP_DIR/symlink11 - ln -s $TMP_DIR/symlink11 $TMP_DIR/../symlink01 + mkdir -p $DIR/d32e/tmp + TMP_DIR=$DIR/d32e/tmp + ln -s $DIR/d32e $TMP_DIR/symlink11 + ln -s $TMP_DIR/symlink11 $TMP_DIR/../symlink01 $CHECKSTAT -t link $DIR/d32e/tmp/symlink11 || error $CHECKSTAT -t link $DIR/d32e/symlink01 || error } @@ -1355,20 +1405,20 @@ run_test 32e "stat d32e/symlink->tmp/symlink->lustre-subdir ====" test_32f() { [ -e $DIR/d32f ] && rm -fr $DIR/d32f - mkdir -p $DIR/d32f/tmp - TMP_DIR=$DIR/d32f/tmp - ln -s $DIR/d32f $TMP_DIR/symlink11 - ln -s $TMP_DIR/symlink11 $TMP_DIR/../symlink01 + mkdir -p $DIR/d32f/tmp + TMP_DIR=$DIR/d32f/tmp + ln -s $DIR/d32f $TMP_DIR/symlink11 + ln -s $TMP_DIR/symlink11 $TMP_DIR/../symlink01 ls $DIR/d32f/tmp/symlink11 || error ls $DIR/d32f/symlink01 || error } run_test 32f "open d32f/symlink->tmp/symlink->lustre-subdir ====" test_32g() { - TMP_DIR=$DIR/$tdir/tmp + TMP_DIR=$DIR/$tdir/tmp mkdir -p $TMP_DIR $DIR/${tdir}2 - ln -s $DIR/${tdir}2 $TMP_DIR/symlink12 - ln -s $TMP_DIR/symlink12 $TMP_DIR/../symlink02 + ln -s $DIR/${tdir}2 $TMP_DIR/symlink12 + ln -s $TMP_DIR/symlink12 $TMP_DIR/../symlink02 $CHECKSTAT -t link $TMP_DIR/symlink12 || error $CHECKSTAT -t link $DIR/$tdir/symlink02 || error $CHECKSTAT -t dir -f $TMP_DIR/symlink12 || error @@ -1378,10 +1428,10 @@ run_test 32g "stat d32g/symlink->tmp/symlink->lustre-subdir/${tdir}2" test_32h() { rm -fr $DIR/$tdir $DIR/${tdir}2 - TMP_DIR=$DIR/$tdir/tmp - mkdir -p $TMP_DIR $DIR/${tdir}2 - ln -s $DIR/${tdir}2 $TMP_DIR/symlink12 - ln -s $TMP_DIR/symlink12 $TMP_DIR/../symlink02 + TMP_DIR=$DIR/$tdir/tmp + mkdir -p $TMP_DIR $DIR/${tdir}2 + ln -s $DIR/${tdir}2 $TMP_DIR/symlink12 + ln -s $TMP_DIR/symlink12 $TMP_DIR/../symlink02 ls $TMP_DIR/symlink12 || error ls $DIR/$tdir/symlink02 || error } @@ -1389,17 +1439,17 @@ run_test 32h "open d32h/symlink->tmp/symlink->lustre-subdir/${tdir}2" test_32i() { [ -e $DIR/d32i ] && rm -fr $DIR/d32i - mkdir -p $DIR/d32i/ext2-mountpoint + mkdir -p $DIR/d32i/ext2-mountpoint mount -t ext2 -o loop $EXT2_DEV $DIR/d32i/ext2-mountpoint || error touch $DIR/d32i/test_file - $CHECKSTAT -t file $DIR/d32i/ext2-mountpoint/../test_file || error + $CHECKSTAT -t file $DIR/d32i/ext2-mountpoint/../test_file || error $UMOUNT $DIR/d32i/ext2-mountpoint || error } run_test 32i "stat d32i/ext2-mountpoint/../test_file ===========" test_32j() { [ -e $DIR/d32j ] && rm -fr $DIR/d32j - mkdir -p $DIR/d32j/ext2-mountpoint + mkdir -p $DIR/d32j/ext2-mountpoint mount -t ext2 -o loop $EXT2_DEV $DIR/d32j/ext2-mountpoint || error touch $DIR/d32j/test_file cat $DIR/d32j/ext2-mountpoint/../test_file || error @@ -1409,8 +1459,8 @@ run_test 32j "open d32j/ext2-mountpoint/../test_file ===========" test_32k() { rm -fr $DIR/d32k - mkdir -p $DIR/d32k/ext2-mountpoint - mount -t ext2 -o loop $EXT2_DEV $DIR/d32k/ext2-mountpoint + mkdir -p $DIR/d32k/ext2-mountpoint + mount -t ext2 -o loop $EXT2_DEV $DIR/d32k/ext2-mountpoint mkdir -p $DIR/d32k/d2 touch $DIR/d32k/d2/test_file || error $CHECKSTAT -t file $DIR/d32k/ext2-mountpoint/../d2/test_file || error @@ -1420,7 +1470,7 @@ run_test 32k "stat d32k/ext2-mountpoint/../d2/test_file ========" test_32l() { rm -fr $DIR/d32l - mkdir -p $DIR/d32l/ext2-mountpoint + mkdir -p $DIR/d32l/ext2-mountpoint mount -t ext2 -o loop $EXT2_DEV $DIR/d32l/ext2-mountpoint || error mkdir -p $DIR/d32l/d2 touch $DIR/d32l/d2/test_file @@ -1431,10 +1481,10 @@ run_test 32l "open d32l/ext2-mountpoint/../d2/test_file ========" test_32m() { rm -fr $DIR/d32m - mkdir -p $DIR/d32m/tmp - TMP_DIR=$DIR/d32m/tmp - ln -s $DIR $TMP_DIR/symlink11 - ln -s $TMP_DIR/symlink11 $TMP_DIR/../symlink01 + mkdir -p $DIR/d32m/tmp + TMP_DIR=$DIR/d32m/tmp + ln -s $DIR $TMP_DIR/symlink11 + ln -s $TMP_DIR/symlink11 $TMP_DIR/../symlink01 $CHECKSTAT -t link $DIR/d32m/tmp/symlink11 || error $CHECKSTAT -t link $DIR/d32m/symlink01 || error } @@ -1442,10 +1492,10 @@ run_test 32m "stat d32m/symlink->tmp/symlink->lustre-root ======" test_32n() { rm -fr $DIR/d32n - mkdir -p $DIR/d32n/tmp - TMP_DIR=$DIR/d32n/tmp - ln -s $DIR $TMP_DIR/symlink11 - ln -s $TMP_DIR/symlink11 $TMP_DIR/../symlink01 + mkdir -p $DIR/d32n/tmp + TMP_DIR=$DIR/d32n/tmp + ln -s $DIR $TMP_DIR/symlink11 + ln -s $TMP_DIR/symlink11 $TMP_DIR/../symlink01 ls -l $DIR/d32n/tmp/symlink11 || error ls -l $DIR/d32n/symlink01 || error } @@ -1453,11 +1503,11 @@ run_test 32n "open d32n/symlink->tmp/symlink->lustre-root ======" test_32o() { rm -fr $DIR/d32o $DIR/$tfile - touch $DIR/$tfile - mkdir -p $DIR/d32o/tmp - TMP_DIR=$DIR/d32o/tmp - ln -s $DIR/$tfile $TMP_DIR/symlink12 - ln -s $TMP_DIR/symlink12 $TMP_DIR/../symlink02 + touch $DIR/$tfile + mkdir -p $DIR/d32o/tmp + TMP_DIR=$DIR/d32o/tmp + ln -s $DIR/$tfile $TMP_DIR/symlink12 + ln -s $TMP_DIR/symlink12 $TMP_DIR/../symlink02 $CHECKSTAT -t link $DIR/d32o/tmp/symlink12 || error $CHECKSTAT -t link $DIR/d32o/symlink02 || error $CHECKSTAT -t file -f $DIR/d32o/tmp/symlink12 || error @@ -1471,15 +1521,15 @@ test_32p() { log 32p_2 rm -f $DIR/$tfile log 32p_3 - touch $DIR/$tfile + touch $DIR/$tfile log 32p_4 - mkdir -p $DIR/d32p/tmp + mkdir -p $DIR/d32p/tmp log 32p_5 - TMP_DIR=$DIR/d32p/tmp + TMP_DIR=$DIR/d32p/tmp log 32p_6 - ln -s $DIR/$tfile $TMP_DIR/symlink12 + ln -s $DIR/$tfile $TMP_DIR/symlink12 log 32p_7 - ln -s $TMP_DIR/symlink12 $TMP_DIR/../symlink02 + ln -s $TMP_DIR/symlink12 $TMP_DIR/../symlink02 log 32p_8 cat $DIR/d32p/tmp/symlink12 || error log 32p_9 @@ -1558,7 +1608,7 @@ test_34b() { run_test 34b "O_RDONLY opening file doesn't create objects =====" test_34c() { - [ ! -f $DIR/f34 ] && test_34a + [ ! -f $DIR/f34 ] && test_34a $CHECKSTAT -s $TEST_34_SIZE $DIR/f34 || error $OPENFILE -f O_RDWR $DIR/f34 $GETSTRIPE $DIR/f34 2>&1 | grep -q "no stripe info" && error @@ -1567,7 +1617,7 @@ test_34c() { run_test 34c "O_RDWR opening file-with-size works ==============" test_34d() { - [ ! -f $DIR/f34 ] && test_34a + [ ! -f $DIR/f34 ] && test_34a dd if=/dev/zero of=$DIR/f34 conv=notrunc bs=4k count=1 || error $CHECKSTAT -s $TEST_34_SIZE $DIR/f34 || error rm $DIR/f34 @@ -1649,7 +1699,7 @@ test_36f() { DATESTR="Dec 20 2000" mkdir -p $DIR/$tdir #define OBD_FAIL_OST_BRW_PAUSE_BULK 0x214 - sysctl -w lustre.fail_loc=0x80000214 + lctl set_param fail_loc=0x80000214 date; date +%s cp /etc/hosts $DIR/$tdir/$tfile sync & # write RPC generated with "current" inode timestamp, but delayed @@ -1668,10 +1718,12 @@ test_36f() { run_test 36f "utime on file racing with OST BRW write ==========" test_36g() { - remote_ost && skip "remote OST" && return + remote_ost_nodsh && skip "remote OST with nodsh" && return + + mkdir -p $DIR/$tdir export FMD_MAX_AGE=`do_facet ost1 lctl get_param -n obdfilter.*.client_cache_seconds 2> /dev/null | head -n 1` FMD_BEFORE="`awk '/ll_fmd_cache/ { print $2 }' /proc/slabinfo`" - touch $DIR/d36/$tfile + touch $DIR/$tdir/$tfile sleep $((FMD_MAX_AGE + 12)) FMD_AFTER="`awk '/ll_fmd_cache/ { print $2 }' /proc/slabinfo`" [ "$FMD_AFTER" -gt "$FMD_BEFORE" ] && \ @@ -1691,9 +1743,15 @@ test_37() { run_test 37 "ls a mounted file system to check old content =====" test_38() { - o_directory $DIR/$tfile + local file=$DIR/$tfile + touch $file + openfile -f O_DIRECTORY $file + local RC=$? + local ENOTDIR=20 + [ $RC -eq 0 ] && error "opened file $file with O_DIRECTORY" || true + [ $RC -eq $ENOTDIR ] || error "error $RC should be ENOTDIR ($ENOTDIR)" } -run_test 38 "open a regular file with O_DIRECTORY ==============" +run_test 38 "open a regular file with O_DIRECTORY should return -ENOTDIR ===" test_39() { touch $DIR/$tfile @@ -1745,9 +1803,9 @@ start_writeback() { # in 2.6, restore /proc/sys/vm/dirty_writeback_centisecs, # dirty_ratio, dirty_background_ratio if [ -f /proc/sys/vm/dirty_writeback_centisecs ]; then - echo $WRITEBACK_SAVE > /proc/sys/vm/dirty_writeback_centisecs - echo $BG_DIRTY_RATIO_SAVE > /proc/sys/vm/dirty_background_ratio - echo $DIRTY_RATIO_SAVE > /proc/sys/vm/dirty_ratio + sysctl -w vm.dirty_writeback_centisecs=$WRITEBACK_SAVE + sysctl -w vm.dirty_background_ratio=$BG_DIRTY_RATIO_SAVE + sysctl -w vm.dirty_ratio=$DIRTY_RATIO_SAVE else # if file not here, we are a 2.4 kernel kill -CONT `pidof kupdated` @@ -1760,15 +1818,15 @@ stop_writeback() { trap start_writeback EXIT # in 2.6, save and 0 /proc/sys/vm/dirty_writeback_centisecs if [ -f /proc/sys/vm/dirty_writeback_centisecs ]; then - WRITEBACK_SAVE=`cat /proc/sys/vm/dirty_writeback_centisecs` - echo 0 > /proc/sys/vm/dirty_writeback_centisecs - echo 0 > /proc/sys/vm/dirty_writeback_centisecs + WRITEBACK_SAVE=`sysctl -n vm.dirty_writeback_centisecs` + sysctl -w vm.dirty_writeback_centisecs=0 + sysctl -w vm.dirty_writeback_centisecs=0 # save and increase /proc/sys/vm/dirty_ratio - DIRTY_RATIO_SAVE=`cat /proc/sys/vm/dirty_ratio` - echo $MAX_DIRTY_RATIO > /proc/sys/vm/dirty_ratio + DIRTY_RATIO_SAVE=`sysctl -n vm.dirty_ratio` + sysctl -w vm.dirty_ratio=$MAX_DIRTY_RATIO # save and increase /proc/sys/vm/dirty_background_ratio - BG_DIRTY_RATIO_SAVE=`cat /proc/sys/vm/dirty_background_ratio` - echo $MAX_BG_DIRTY_RATIO > /proc/sys/vm/dirty_background_ratio + BG_DIRTY_RATIO_SAVE=`sysctl -n vm.dirty_background_ratio` + sysctl -w vm.dirty_background_ratio=$MAX_BG_DIRTY_RATIO else # if file not here, we are a 2.4 kernel kill -STOP `pidof kupdated` @@ -1875,6 +1933,7 @@ test_42d() { run_test 42d "test complete truncate of file with cached dirty data" test_43() { + mkdir -p $DIR/$tdir cp -p /bin/ls $DIR/$tdir/$tfile multiop $DIR/$tdir/$tfile Ow_c & pid=$! @@ -2004,7 +2063,7 @@ page_size() { # in a 2 stripe file (lov.sh), page 1023 maps to page 511 in its object. this # test tickles a bug where re-dirtying a page was failing to be mapped to the -# objects offset and an assert hit when an rpc was built with 1023's mapped +# objects offset and an assert hit when an rpc was built with 1023's mapped # offset 511 and 511's raw 511 offset. it also found general redirtying bugs. test_46() { f="$DIR/f46" @@ -2019,11 +2078,7 @@ test_46() { } run_test 46 "dirtying a previously written page ================" -# Check that device nodes are created and then visible correctly (#2091) -test_47() { - cmknod $DIR/test_47_node || error -} -run_test 47 "Device nodes check ================================" +# test_47 is removed "Device nodes check" is moved to test_28 test_48a() { # bug 2399 check_kernel_version 34 || return 0 @@ -2070,7 +2125,7 @@ run_test 48b "Access removed working dir (should return errors)=" test_48c() { # bug 2350 check_kernel_version 36 || return 0 - #sysctl -w lnet.debug=-1 + #lctl set_param debug=-1 #set -vx mkdir -p $DIR/d48c/dir cd $DIR/d48c/dir @@ -2093,7 +2148,7 @@ run_test 48c "Access removed working subdir (should return errors)" test_48d() { # bug 2350 check_kernel_version 36 || return 0 - #sysctl -w lnet.debug=-1 + #lctl set_param debug=-1 #set -vx mkdir -p $DIR/d48d/dir cd $DIR/d48d/dir @@ -2117,7 +2172,7 @@ run_test 48d "Access removed parent subdir (should return errors)" test_48e() { # bug 4134 check_kernel_version 41 || return 0 - #sysctl -w lnet.debug=-1 + #lctl set_param debug=-1 #set -vx mkdir -p $DIR/d48e/dir cd $DIR/d48e/dir @@ -2141,7 +2196,7 @@ test_50() { } run_test 50 "special situations: /proc symlinks ===============" -test_51() { +test_51a() { # was test_51 # bug 1516 - create an empty entry right after ".." then split dir mkdir $DIR/d51 touch $DIR/d51/foo @@ -2157,7 +2212,7 @@ test_51() { echo ls -l $DIR/d51 > /dev/null || error } -run_test 51 "special situations: split htree with empty entry ==" +run_test 51a "special situations: split htree with empty entry ==" #export NUMTEST=70000 # FIXME: I select a relatively small number to do basic test. @@ -2177,6 +2232,57 @@ test_51b() { } run_test 51b "mkdir .../t-0 --- .../t-$NUMTEST ====================" +test_51bb() { + [ $MDSCOUNT -lt 2 ] && skip "needs >= 2 MDTs" && return + + local ndirs=${TEST51BB_NDIRS:-10} + local nfiles=${TEST51BB_NFILES:-100} + + local numfree=`df -i -P $DIR | tail -n 1 | awk '{ print $4 }'` + + [ $numfree -lt $(( ndirs * nfiles)) ] && \ + nfiles=$(( numfree / ndirs - 10 )) + + local dir=$DIR/d51bb + mkdir -p $dir + local savePOLICY=$(lctl get_param -n lmv.*.placement) + lctl set_param -n lmv.*.placement=CHAR + + lfs df -i $dir + local IUSED=$(lfs df -i $dir | grep MDT | awk '{print $3}') + OLDUSED=($IUSED) + + declare -a dirs + for ((i=0; i < $ndirs; i++)); do + dirs[i]=$dir/$RANDOM + echo Creating directory ${dirs[i]} + mkdir -p ${dirs[i]} + ls $dir + echo Creating $nfiles in dir ${dirs[i]} ... + echo "createmany -o ${dirs[i]}/$tfile- $nfiles" + createmany -o ${dirs[i]}/$tfile- $nfiles + done + ls $dir + + sleep 1 + + IUSED=$(lfs df -i $dir | grep MDT | awk '{print $3}') + NEWUSED=($IUSED) + + local rc=0 + for ((i=0; i<${#NEWUSED[@]}; i++)); do + echo "mds $i: inodes count OLD ${OLDUSED[$i]} NEW ${NEWUSED[$i]}" + [ ${OLDUSED[$i]} -lt ${NEWUSED[$i]} ] || rc=$((rc + 1)) + done + + lctl set_param -n lmv.*.placement=$savePOLICY + + [ $rc -ne $MDSCOUNT ] || \ + error "Objects/inodes are not distributed over all mds servers" +} +run_test 51bb "mkdir createmany CMD $MDSCOUNT ====================" + + test_51c() { [ ! -d $DIR/d51b ] && skip "$DIR/51b missing" && \ return @@ -2203,7 +2309,7 @@ test_51d() { error "OST $N has less objects vs OST $NLAST (${OBJS[$N]} < ${OBJS[$NLAST]}" [ ${OBJS[$N]} -gt $((${OBJS[$NLAST]} + 20)) ] && \ error "OST $N has less objects vs OST $NLAST (${OBJS[$N]} < ${OBJS[$NLAST]}" - + [ ${OBJS0[$N]} -lt $((${OBJS0[$NLAST]} - 20)) ] && \ error "OST $N has less #0 objects vs OST $NLAST (${OBJS0[$N]} < ${OBJS0[$NLAST]}" [ ${OBJS0[$N]} -gt $((${OBJS0[$NLAST]} + 20)) ] && \ @@ -2252,14 +2358,22 @@ test_52b() { run_test 52b "immutable flag test (should return errors) =======" test_53() { - remote_mds && skip "remote MDS" && return + remote_mds_nodsh && skip "remote MDS with nodsh" && return + remote_ost_nodsh && skip "remote OST with nodsh" && return + + local param + local ostname + local mds_last + local ost_last + local ostnum - # only test MDT0000 - for value in `lctl get_param osc.*-osc-MDT0000.prealloc_last_id` ; do + # only test MDT0000 + for value in $(do_facet $SINGLEMDS lctl get_param osc.*-osc-MDT0000.prealloc_last_id) ; do param=`echo ${value[0]} | cut -d "=" -f1` ostname=`echo $param | cut -d "." -f2 | cut -d - -f 1-2` - ost_last=`lctl get_param -n obdfilter.$ostname.last_id` - mds_last=`lctl get_param -n $param` + mds_last=$(do_facet $SINGLEMDS lctl get_param -n $param) + ostnum=$(echo $ostname | sed "s/${FSNAME}-OST//g" | awk '{print ($1+1)}' ) + ost_last=$(do_facet ost$ostnum lctl get_param -n obdfilter.$ostname.last_id | head -n 1) echo "$ostname.last_id=$ost_last ; MDS.last_id=$mds_last" if [ $ost_last != $mds_last ]; then error "$ostname.last_id=$ost_last ; MDS.last_id=$mds_last" @@ -2281,7 +2395,7 @@ test_54b() { f="$DIR/f54b" mknod $f c 1 3 chmod 0666 $f - dd if=/dev/zero of=$f bs=`page_size` count=1 + dd if=/dev/zero of=$f bs=`page_size` count=1 } run_test 54b "char device works in lustre ======================" @@ -2303,7 +2417,7 @@ test_54c() { tdir="$DIR/d54c" loopdev="$DIR/loop54c" - find_loop_dev + find_loop_dev [ -z "$LOOPNUM" ] && echo "couldn't find empty loop device" && return mknod $loopdev b 7 $LOOPNUM echo "make a loop file system with $tfile on $loopdev ($LOOPNUM)..." @@ -2363,7 +2477,7 @@ test_55() { } run_test 55 "check iopen_connect_dentry() ======================" -test_56() { +test_56a() { # was test_56 rm -rf $DIR/d56 $SETSTRIPE -d $DIR mkdir $DIR/d56 @@ -2413,7 +2527,7 @@ test_56() { error "lfs getstripe --obd wrong: should not show file on other obd" echo "lfs getstripe --obd passed." } -run_test 56 "check lfs getstripe ====================================" +run_test 56a "check lfs getstripe ====================================" NUMFILES=3 NUMDIRS=3 @@ -2548,11 +2662,11 @@ test_56o() { setup_56 $NUMFILES $NUMDIRS TDIR=$DIR/${tdir}g - utime $TDIR/file1 > /dev/null || error - utime $TDIR/file2 > /dev/null || error - utime $TDIR/dir1 > /dev/null || error - utime $TDIR/dir2 > /dev/null || error - utime $TDIR/dir1/file1 > /dev/null || error + utime $TDIR/file1 > /dev/null || error "utime (1)" + utime $TDIR/file2 > /dev/null || error "utime (2)" + utime $TDIR/dir1 > /dev/null || error "utime (3)" + utime $TDIR/dir2 > /dev/null || error "utime (4)" + utime $TDIR/dir1/file1 > /dev/null || error "utime (5)" EXPECTED=5 NUMS=`$LFIND -mtime +1 $TDIR | wc -l` @@ -2561,14 +2675,61 @@ test_56o() { } run_test 56o "check lfs find -mtime for old files ==========================" +test_56p() { + [ $RUNAS_ID -eq $UID ] && skip "RUNAS_ID = UID = $UID -- skipping" && return + + TDIR=$DIR/${tdir}g + rm -rf $TDIR + + setup_56 $NUMFILES $NUMDIRS + + chown $RUNAS_ID $TDIR/file* || error "chown $DIR/${tdir}g/file$i failed" + EXPECTED=$NUMFILES + NUMS="`$LFIND -uid $RUNAS_ID $TDIR | wc -l`" + [ $NUMS -eq $EXPECTED ] || \ + error "lfs find -uid $TDIR wrong: found $NUMS, expected $EXPECTED" + + EXPECTED=$(( ($NUMFILES+1) * $NUMDIRS + 1)) + NUMS="`$LFIND ! -uid $RUNAS_ID $TDIR | wc -l`" + [ $NUMS -eq $EXPECTED ] || \ + error "lfs find ! -uid $TDIR wrong: found $NUMS, expected $EXPECTED" + + echo "lfs find -uid and ! -uid passed." +} +run_test 56p "check lfs find -uid and ! -uid ===============================" + +test_56q() { + [ $RUNAS_ID -eq $UID ] && skip "RUNAS_ID = UID = $UID -- skipping" && return + + TDIR=$DIR/${tdir}g + rm -rf $TDIR + + setup_56 $NUMFILES $NUMDIRS + + chgrp $RUNAS_ID $TDIR/file* || error "chown $DIR/${tdir}g/file$i failed" + EXPECTED=$NUMFILES + NUMS="`$LFIND -gid $RUNAS_ID $TDIR | wc -l`" + [ $NUMS -eq $EXPECTED ] || \ + error "lfs find -gid $TDIR wrong: found $NUMS, expected $EXPECTED" + + EXPECTED=$(( ($NUMFILES+1) * $NUMDIRS + 1)) + NUMS="`$LFIND ! -gid $RUNAS_ID $TDIR | wc -l`" + [ $NUMS -eq $EXPECTED ] || \ + error "lfs find ! -gid $TDIR wrong: found $NUMS, expected $EXPECTED" + + echo "lfs find -gid and ! -gid passed." +} +run_test 56q "check lfs find -gid and ! -gid ===============================" + test_57a() { # note test will not do anything if MDS is not local - remote_mds && skip "remote MDS" && return + remote_mds_nodsh && skip "remote MDS with nodsh" && return + local MNTDEV="osd.*MDT*.mntdev" - DEV=$(lctl get_param -n $MNTDEV) - [ -z "$DEV" ] && error "can't access $MNTDEV" - for DEV in `lctl get_param -n $MNTDEV`; do - dumpe2fs -h $DEV > $TMP/t57a.dump || error "can't access $DEV" + DEV=$(do_facet $SINGLEMDS lctl get_param -n $MNTDEV) + [ -z "$DEV" ] && error "can't access $MNTDEV" + for DEV in $(do_facet $SINGLEMDS lctl get_param -n $MNTDEV); do + do_facet $SINGLEMDS $DUMPE2FS -h $DEV > $TMP/t57a.dump || error "can't access $DEV" DEVISIZE=`awk '/Inode size:/ { print $3 }' $TMP/t57a.dump` [ "$DEVISIZE" -gt 128 ] || error "inode size $DEVISIZE" rm $TMP/t57a.dump @@ -2577,42 +2738,53 @@ test_57a() { run_test 57a "verify MDS filesystem created with large inodes ==" test_57b() { - FILECOUNT=100 - FILE1=$DIR/d57b/f1 - FILEN=$DIR/d57b/f$FILECOUNT - rm -rf $DIR/d57b || error "removing $DIR/d57b" - mkdir -p $DIR/d57b || error "creating $DIR/d57b" + local dir=$DIR/d57b + + local FILECOUNT=100 + local FILE1=$dir/f1 + local FILEN=$dir/f$FILECOUNT + + rm -rf $dir || error "removing $dir" + mkdir -p $dir || error "creating $dir" + local num=$(get_mds_dir $dir) + local mymds=mds$num + echo "mcreating $FILECOUNT files" - createmany -m $DIR/d57b/f 1 $FILECOUNT || \ - error "creating files in $DIR/d57b" + createmany -m $dir/f 1 $FILECOUNT || \ + error "creating files in $dir" # verify that files do not have EAs yet $GETSTRIPE $FILE1 2>&1 | grep -q "no stripe" || error "$FILE1 has an EA" $GETSTRIPE $FILEN 2>&1 | grep -q "no stripe" || error "$FILEN has an EA" - MDSFREE="`lctl get_param -n osd.*MDT*.kbytesfree 2> /dev/null`" - MDCFREE="`lctl get_param -n mdc.*.kbytesfree | head -n 1`" + sync + sleep 1 + df $dir #make sure we get new statfs data + local MDSFREE=$(do_facet $mymds lctl get_param -n osd.*MDT000$((num -1)).kbytesfree) + local MDCFREE=$(lctl get_param -n mdc.*MDT000$((num -1))-mdc-*.kbytesfree) echo "opening files to create objects/EAs" - for FILE in `seq -f $DIR/d57b/f%g 1 $FILECOUNT`; do - $OPENFILE -f O_RDWR $FILE > /dev/null || error "opening $FILE" + local FILE + for FILE in `seq -f $dir/f%g 1 $FILECOUNT`; do + $OPENFILE -f O_RDWR $FILE > /dev/null 2>&1 || error "opening $FILE" done # verify that files have EAs now $GETSTRIPE $FILE1 | grep -q "obdidx" || error "$FILE1 missing EA" $GETSTRIPE $FILEN | grep -q "obdidx" || error "$FILEN missing EA" - sleep 1 # make sure we get new statfs data -# MDSFREE2="`lctl get_param -n mds.*.kbytesfree`" -# MDCFREE2="`lctl get_param -n mdc.*.kbytesfree`" -# if [ "$MDCFREE2" -lt "$((MDCFREE - 8))" ]; then -# if [ "$MDSFREE" != "$MDSFREE2" ]; then -# error "MDC before $MDCFREE != after $MDCFREE2" -# else -# echo "MDC before $MDCFREE != after $MDCFREE2" -# echo "unable to confirm if MDS has large inodes" -# fi -# fi - rm -rf $DIR/d57b + sleep 1 #make sure we get new statfs data + df $dir + local MDSFREE2=$(do_facet $mymds lctl get_param -n osd.*MDT000$((num -1)).kbytesfree) + local MDCFREE2=$(lctl get_param -n mdc.*MDT000$((num -1))-mdc-*.kbytesfree) + if [ "$MDCFREE2" -lt "$((MDCFREE - 8))" ]; then + if [ "$MDSFREE" != "$MDSFREE2" ]; then + error "MDC before $MDCFREE != after $MDCFREE2" + else + echo "MDC before $MDCFREE != after $MDCFREE2" + echo "unable to confirm if MDS has large inodes" + fi + fi + rm -rf $dir } run_test 57b "default LOV EAs are stored inside large inodes ===" @@ -2637,7 +2809,7 @@ TEST60_HEAD="test_60 run $RANDOM" test_60a() { [ ! -f run-llog.sh ] && skip "missing subtest run-llog.sh" && return log "$TEST60_HEAD - from kernel mode" -# sh run-llog.sh + sh run-llog.sh } run_test 60a "llog sanity tests run from kernel module ==========" @@ -2658,34 +2830,33 @@ test_60b() { # bug 6411 [ $LLOG_COUNT -gt 50 ] && error "CDEBUG_LIMIT not limiting messages ($LLOG_COUNT)"|| true } run_test 60b "limit repeated messages from CERROR/CWARN ========" - + test_60c() { - echo "create 5000 files" + echo "create 5000 files" createmany -o $DIR/f60c- 5000 -#define OBD_FAIL_MDS_LLOG_CREATE_FAILED 0x13c - sysctl -w lustre.fail_loc=0x8000013c +#define OBD_FAIL_MDS_LLOG_CREATE_FAILED 0x137 + lctl set_param fail_loc=0x80000137 unlinkmany $DIR/f60c- 5000 - sysctl -w lustre.fail_loc=0 + lctl set_param fail_loc=0 } run_test 60c "unlink file when mds full" test_60d() { - SAVEPRINTK=$(sysctl -n lnet.printk) + SAVEPRINTK=$(lctl get_param -n printk) # verify "lctl mark" is even working" MESSAGE="test message ID $RANDOM $$" $LCTL mark "$MESSAGE" || error "$LCTL mark failed" dmesg | grep -q "$MESSAGE" || error "didn't find debug marker in log" - sysctl -w lnet.printk=0 || error "set lnet.printk failed" - sysctl -n lnet.printk | grep emerg || error "lnet.printk dropped emerg" - + lctl set_param printk=0 || error "set lnet.printk failed" + lctl get_param -n printk | grep emerg || error "lnet.printk dropped emerg" MESSAGE="new test message ID $RANDOM $$" # Assume here that libcfs_debug_mark_buffer() uses D_WARNING $LCTL mark "$MESSAGE" || error "$LCTL mark failed" dmesg | grep -q "$MESSAGE" && error "D_WARNING wasn't masked" || true - sysctl -w lnet.printk="$SAVEPRINTK" + lctl set_param -n printk="$SAVEPRINTK" } run_test 60d "test printk console message masking" @@ -2703,14 +2874,16 @@ test_62() { f="$DIR/f62" echo foo > $f cancel_lru_locks osc - sysctl -w lustre.fail_loc=0x405 + lctl set_param fail_loc=0x405 cat $f && error "cat succeeded, expect -EIO" - sysctl -w lustre.fail_loc=0 + lctl set_param fail_loc=0 } -run_test 62 "verify obd_match failure doesn't LBUG (should -EIO)" +# This test is now irrelevant (as of bug 10718 inclusion), we no longer +# match every page all of the time. +#run_test 62 "verify obd_match failure doesn't LBUG (should -EIO)" # bug 2319 - oig_wait() interrupted causes crash because of invalid waitq. -test_63() { +test_63a() { # was test_63 MAX_DIRTY_MB=`lctl get_param -n osc.*.max_dirty_mb | head -n 1` lctl set_param -n osc.*.max_dirty_mb 0 for i in `seq 10` ; do @@ -2723,20 +2896,20 @@ test_63() { lctl set_param -n osc.*.max_dirty_mb $MAX_DIRTY_MB rm -f $DIR/f63 || true } -run_test 63 "Verify oig_wait interruption does not crash =======" +run_test 63a "Verify oig_wait interruption does not crash =======" # bug 2248 - async write errors didn't return to application on sync # bug 3677 - async write errors left page locked test_63b() { debugsave - sysctl -w lnet.debug=-1 + lctl set_param debug=-1 # ensure we have a grant to do async writes dd if=/dev/zero of=$DIR/$tfile bs=4k count=1 rm $DIR/$tfile #define OBD_FAIL_OSC_BRW_PREP_REQ 0x406 - sysctl -w lustre.fail_loc=0x80000406 + lctl set_param fail_loc=0x80000406 multiop $DIR/$tfile Owy && \ error "sync didn't return ENOMEM" sync; sleep 2; sync # do a real sync this time to flush page @@ -2809,7 +2982,7 @@ test_65e() { touch $DIR/d65/f6 $LVERIFY $DIR/d65 $DIR/d65/f6 || error "lverify failed" } -run_test 65e "directory setstripe 0 -1 0 =======================" +run_test 65e "directory setstripe defaults =======================" test_65f() { mkdir -p $DIR/d65f @@ -2834,12 +3007,27 @@ test_65h() { "`$GETSTRIPE -v $DIR/d65/dd1 | grep "^count"`" ] || error "stripe info inherit failed" } run_test 65h "directory stripe info inherit ====================" - + test_65i() { # bug6367 - $SETSTRIPE $MOUNT -s 65536 -c -1 + $SETSTRIPE $MOUNT -s 65536 -c -1 } run_test 65i "set non-default striping on root directory (bug 6367)=" +test_65ia() { # bug12836 + $LFS getstripe $MOUNT || error "getstripe $MOUNT failed" +} +run_test 65ia "getstripe on -1 default directory striping" + +test_65ib() { # bug12836 + $LFS getstripe -v $MOUNT || error "getstripe -v $MOUNT failed" +} +run_test 65ib "getstripe -v on -1 default directory striping" + +test_65ic() { # bug12836 + $LFS find -mtime -1 $MOUNT || error "find $MOUNT failed" +} +run_test 65ic "new find on -1 default directory striping" + test_65j() { # bug6367 sync; sleep 1 # if we aren't already remounting for each test, do so for this test @@ -2853,22 +3041,24 @@ run_test 65j "set default striping on root directory (bug 6367)=" test_65k() { # bug11679 [ "$OSTCOUNT" -lt 2 ] && skip "too few OSTs" && return - remote_mds_nodsh && skip "remote MDS" && return + remote_mds_nodsh && skip "remote MDS with nodsh" && return echo "Check OST status: " - MDS_OSCS=`do_facet mds lctl dl | awk '/[oO][sS][cC].*md[ts]/ { print $4 }'` + MDS_OSCS=`do_facet $SINGLEMDS lctl dl | awk '/[oO][sS][cC].*md[ts]/ { print $4 }'` for OSC in $MDS_OSCS; do echo $OSC "is activate" - do_facet mds lctl --device %$OSC activate + do_facet $SINGLEMDS lctl --device %$OSC activate done do_facet client mkdir -p $DIR/$tdir for INACTIVE_OSC in $MDS_OSCS; do echo $INACTIVE_OSC "is Deactivate:" - do_facet mds lctl --device %$INACTIVE_OSC deactivate + do_facet $SINGLEMDS lctl --device %$INACTIVE_OSC deactivate for STRIPE_OSC in $MDS_OSCS; do STRIPE_OST=`osc_to_ost $STRIPE_OSC` - STRIPE_INDEX=`do_facet mds lctl get_param -n lov.*md*.target_obd | - grep $STRIPE_OST | awk -F: '{print $1}'` + STRIPE_INDEX=`do_facet $SINGLEMDS lctl get_param -n lov.*md*.target_obd | + grep $STRIPE_OST | awk -F: '{print $1}' | head -n 1` + + [ -f $DIR/$tdir/${STRIPE_INDEX} ] && continue echo "$SETSTRIPE $DIR/$tdir/${STRIPE_INDEX} -i ${STRIPE_INDEX} -c 1" do_facet client $SETSTRIPE $DIR/$tdir/${STRIPE_INDEX} -i ${STRIPE_INDEX} -c 1 RC=$? @@ -2876,7 +3066,7 @@ test_65k() { # bug11679 done do_facet client rm -f $DIR/$tdir/* echo $INACTIVE_OSC "is Activate." - do_facet mds lctl --device %$INACTIVE_OSC activate + do_facet $SINGLEMDS lctl --device %$INACTIVE_OSC activate done } run_test 65k "validate manual striping works properly with deactivated OSCs" @@ -2902,12 +3092,15 @@ LLOOP= cleanup_68() { trap 0 if [ ! -z "$LLOOP" ]; then - swapoff $LLOOP || error "swapoff failed" + if swapon -s | grep -q $LLOOP; then + swapoff $LLOOP || error "swapoff failed" + fi + $LCTL blockdev_detach $LLOOP || error "detach failed" rm -f $LLOOP unset LLOOP fi - rm -f $DIR/f68 + rm -f $DIR/f68* } meminfo() { @@ -2918,10 +3111,29 @@ swap_used() { swapon -s | awk '($1 == "'$1'") { print $4 }' } +# test case for lloop driver, basic function +test_68a() { + [ "$UID" != 0 ] && skip "must run as root" && return + + grep -q llite_lloop /proc/modules + [ $? -ne 0 ] && skip "can't find module llite_lloop" && return + + LLOOP=$TMP/lloop.`date +%s`.`date +%N` + dd if=/dev/zero of=$DIR/f68a bs=4k count=1024 + $LCTL blockdev_attach $DIR/f68a $LLOOP || error "attach failed" + + trap cleanup_68 EXIT + + directio rdwr $LLOOP 0 1024 4096 || error "direct write failed" + directio rdwr $LLOOP 0 1025 4096 && error "direct write should fail" + + cleanup_68 +} +run_test 68a "lloop driver - basic test ========================" # excercise swapping to lustre by adding a high priority swapfile entry # and then consuming memory until it is used. -test_68() { +test_68b() { # was test_68 [ "$UID" != 0 ] && skip "must run as root" && return lctl get_param -n devices | grep -q obdfilter && \ skip "local OST" && return @@ -2937,10 +3149,10 @@ test_68() { [[ $NR_BLOCKS -le 2048 ]] && NR_BLOCKS=2048 LLOOP=$TMP/lloop.`date +%s`.`date +%N` - dd if=/dev/zero of=$DIR/f68 bs=64k seek=$NR_BLOCKS count=1 - mkswap $DIR/f68 + dd if=/dev/zero of=$DIR/f68b bs=64k seek=$NR_BLOCKS count=1 + mkswap $DIR/f68b - $LCTL blockdev_attach $DIR/f68 $LLOOP || error "attach failed" + $LCTL blockdev_attach $DIR/f68b $LLOOP || error "attach failed" trap cleanup_68 EXIT @@ -2955,68 +3167,48 @@ test_68() { [ $SWAPUSED -eq 0 ] && echo "no swap used???" || true } -run_test 68 "support swapping to Lustre ========================" +run_test 68b "support swapping to Lustre ========================" # bug5265, obdfilter oa2dentry return -ENOENT # #define OBD_FAIL_OST_ENOENT 0x217 test_69() { - [ $(lctl get_param -n devices | grep -c obdfilter) -eq 0 ] && \ - skip "skipping test for remote OST" && return + remote_ost_nodsh && skip "remote OST with nodsh" && return f="$DIR/$tfile" - touch $f + $SETSTRIPE $f -c 1 -i 0 $DIRECTIO write ${f}.2 0 1 || error "directio write error" - sysctl -w lustre.fail_loc=0x217 + do_facet ost1 lctl set_param fail_loc=0x217 truncate $f 1 # vmtruncate() will ignore truncate() error. $DIRECTIO write $f 0 2 && error "write succeeded, expect -ENOENT" - sysctl -w lustre.fail_loc=0 + do_facet ost1 lctl set_param fail_loc=0 $DIRECTIO write $f 0 2 || error "write error" cancel_lru_locks osc $DIRECTIO read $f 0 1 || error "read error" - sysctl -w lustre.fail_loc=0x217 + do_facet ost1 lctl set_param fail_loc=0x217 $DIRECTIO read $f 1 1 && error "read succeeded, expect -ENOENT" - sysctl -w lustre.fail_loc=0 + do_facet ost1 lctl set_param fail_loc=0 rm -f $f } run_test 69 "verify oa2dentry return -ENOENT doesn't LBUG ======" test_71() { - which dbench > /dev/null 2>&1 || { skip "dbench not installed, skip this test" && return 0; } - DBENCH_LIB=${DBENCH_LIB:-/usr/lib/dbench} - PATH=${DBENCH_LIB}:${PATH} - cp `which dbench` $DIR - - TGT=$DIR/client.txt - SRC=${SRC:-$DBENCH_LIB/client.txt} - [ ! -e $TGT -a -e $SRC ] && echo "copying $SRC to $TGT" && cp $SRC $TGT - SRC=$DBENCH_LIB/client_plain.txt - [ ! -e $TGT -a -e $SRC ] && echo "copying $SRC to $TGT" && cp $SRC $TGT - echo "copying necessary libs to $DIR" - LIBS71=$(ldd $DIR/dbench|sed -e 's/\t*//' -e 's/.*=> //' -e 's/ .*//' -e 's/^\///') - (cd / && tar chf - $LIBS71) | (cd $DIR && tar xvf -) - [ $? = 0 ] || error "can't copy libs $LIBS71 to $DIR" - echo "chroot $DIR /dbench -c client.txt 2" - chroot $DIR /dbench -c client.txt 2 - RC=$? - - rm -rf $DIR/dbench $DIR/lib $DIR/lib64 - - return $RC + mkdir -p $DIR/$tdir + sh rundbench -C -D $DIR/$tdir 2 || error "dbench failed!" } run_test 71 "Running dbench on lustre (don't segment fault) ====" test_72() { # bug 5695 - Test that on 2.6 remove_suid works properly check_kernel_version 43 || return 0 [ "$RUNAS_ID" = "$UID" ] && skip "RUNAS_ID = UID = $UID -- skipping" && return - + # Check that testing environment is properly set up. Skip if not - FAIL_ON_ERROR=false check_runas_id_ret $RUNAS_ID $RUNAS || { + FAIL_ON_ERROR=false check_runas_id_ret $RUNAS_ID $RUNAS_ID $RUNAS || { skip "User $RUNAS_ID does not exist - skipping" return 0 } @@ -3032,20 +3224,21 @@ test_72() { # bug 5695 - Test that on 2.6 remove_suid works properly cancel_lru_locks mdc test -u $DIR/f72 -o -g $DIR/f72 && error "S/gid is not dropped on MDS" true + rm -f $DIR/f72 } run_test 72 "Test that remove suid works properly (bug5695) ====" # bug 3462 - multiple simultaneous MDC requests test_73() { - mkdir $DIR/d73-1 + mkdir $DIR/d73-1 mkdir $DIR/d73-2 multiop_bg_pause $DIR/d73-1/f73-1 O_c || return 1 pid1=$! - echo 0x80000129 > /proc/sys/lustre/fail_loc + lctl set_param fail_loc=0x80000129 multiop $DIR/d73-1/f73-2 Oc & sleep 1 - echo 0 > /proc/sys/lustre/fail_loc + lctl set_param fail_loc=0 multiop $DIR/d73-2/f73-3 Oc & pid3=$! @@ -3056,8 +3249,8 @@ test_73() { sleep 25 $CHECKSTAT -t file $DIR/d73-1/f73-1 || return 4 - $CHECKSTAT -t file $DIR/d73-1/f73-2 || return 5 - $CHECKSTAT -t file $DIR/d73-2/f73-3 || return 6 + $CHECKSTAT -t file $DIR/d73-1/f73-2 || return 5 + $CHECKSTAT -t file $DIR/d73-2/f73-3 || return 6 rm -rf $DIR/d73-* } @@ -3069,11 +3262,12 @@ test_74a() { # bug 6149, 6184 # very important to OR with OBD_FAIL_ONCE (0x80000000) -- otherwise it # will spin in a tight reconnection loop touch $DIR/f74a - sysctl -w lustre.fail_loc=0x8000030e + lctl set_param fail_loc=0x8000030e # get any lock that won't be difficult - lookup works. ls $DIR/f74a - sysctl -w lustre.fail_loc=0 + lctl set_param fail_loc=0 true + rm -f $DIR/f74a } run_test 74a "ldlm_enqueue freed-export error path, ls (shouldn't LBUG)" @@ -3082,11 +3276,12 @@ test_74b() { # bug 13310 # # very important to OR with OBD_FAIL_ONCE (0x80000000) -- otherwise it # will spin in a tight reconnection loop - sysctl -w lustre.fail_loc=0x8000030e + lctl set_param fail_loc=0x8000030e # get a "difficult" lock touch $DIR/f74b - sysctl -w lustre.fail_loc=0 + lctl set_param fail_loc=0 true + rm -f $DIR/f74b } run_test 74b "ldlm_enqueue freed-export error path, touch (shouldn't LBUG)" @@ -3099,25 +3294,22 @@ export T75_PREP=no test75_prep() { [ $T75_PREP = "yes" ] && return echo "using F75=$F75, F128k=$F128k, FHEAD=$FHEAD, FTAIL=$FTAIL" - + dd if=/dev/urandom of=${F75}_128k bs=128k count=1 || error "dd failed" log "finished dd" chmod 777 ${F128k} T75_PREP=yes } - + test_75a() { -# skipped temporarily: we do not have join file currently -# please remove this when ready - huanghua - return test75_prep - + cp -p ${F128k} ${FHEAD} log "finished cp to $FHEAD" cp -p ${F128k} ${FTAIL} log "finished cp to $FTAIL" cat ${F128k} ${F128k} > ${F75}_sim_sim - + $JOIN ${FHEAD} ${FTAIL} || error "join ${FHEAD} ${FTAIL} error" log "finished join $FHEAD to ${F75}_sim_sim" cmp ${FHEAD} ${F75}_sim_sim || error "${FHEAD} ${F75}_sim_sim differ" @@ -3125,13 +3317,10 @@ test_75a() { $CHECKSTAT -a ${FTAIL} || error "tail ${FTAIL} still exist after join" } run_test 75a "TEST join file ====================================" - + test_75b() { -# skipped temporarily: we do not have join file currently -# please remove this when ready - huanghua - return test75_prep - + cp -p ${F128k} ${FTAIL} cat ${F75}_sim_sim >> ${F75}_join_sim cat ${F128k} >> ${F75}_join_sim @@ -3141,13 +3330,10 @@ test_75b() { $CHECKSTAT -a ${FTAIL} || error "tail ${FTAIL} exist after join" } run_test 75b "TEST join file 2 ==================================" - + test_75c() { -# skipped temporarily: we do not have join file currently -# please remove this when ready - huanghua - return test75_prep - + cp -p ${F128k} ${FTAIL} cat ${F128k} >> ${F75}_sim_join cat ${F75}_join_sim >> ${F75}_sim_join @@ -3157,13 +3343,10 @@ test_75c() { $CHECKSTAT -a ${FHEAD} || error "tail ${FHEAD} exist after join" } run_test 75c "TEST join file 3 ==================================" - + test_75d() { -# skipped temporarily: we do not have join file currently -# please remove this when ready - huanghua - return test75_prep - + cp -p ${F128k} ${FHEAD} cp -p ${F128k} ${FHEAD}_tmp cat ${F75}_sim_sim >> ${F75}_join_join @@ -3174,23 +3357,17 @@ test_75d() { $CHECKSTAT -a ${FTAIL} || error "tail ${FTAIL} exist after join (2)" } run_test 75d "TEST join file 4 ==================================" - + test_75e() { -# skipped temporarily: we do not have join file currently -# please remove this when ready - huanghua - return test75_prep - + rm -rf ${FHEAD} || "delete join file error" } run_test 75e "TEST join file 5 (remove joined file) =============" - + test_75f() { -# skipped temporarily: we do not have join file currently -# please remove this when ready - huanghua - return test75_prep - + cp -p ${F128k} ${F75}_join_10_compare cp -p ${F128k} ${F75}_join_10 for ((i = 0; i < 10; i++)); do @@ -3204,16 +3381,13 @@ test_75f() { error "files ${F75}_join_10 ${F75}_join_10_compare differ" } run_test 75f "TEST join file 6 (join 10 files) ==================" - + test_75g() { -# skipped temporarily: we do not have join file currently -# please remove this when ready - huanghua - return [ ! -f ${F75}_join_10 ] && echo "${F75}_join_10 missing" && return $LFS getstripe ${F75}_join_10 - + $OPENUNLINK ${F75}_join_10 ${F75}_join_10 || error "files unlink open" - + ls -l $F75* } run_test 75g "TEST join file 7 (open unlink) ====================" @@ -3263,7 +3437,7 @@ set_checksum_type() [ "$ORIG_CSUM_TYPE" ] || \ ORIG_CSUM_TYPE=`lctl get_param -n osc/*osc-[^mM]*/checksum_type | sed 's/.*\[\(.*\)\].*/\1/g' | head -n1` - lctl set_param -n osc.*osc-*.checksum_type $1 + lctl set_param -n osc.*osc-[^mM]*.checksum_type $1 log "set checksum type to $1" return 0 } @@ -3275,35 +3449,40 @@ setup_f77() { } test_77a() { # bug 10889 + $GSS && skip "could not run with gss" && return [ ! -f $F77_TMP ] && setup_f77 set_checksums 1 dd if=$F77_TMP of=$DIR/$tfile bs=1M count=$F77SZ || error "dd error" set_checksums 0 + rm -f $DIR/$tfile } run_test 77a "normal checksum read/write operation =============" test_77b() { # bug 10889 + $GSS && skip "could not run with gss" && return [ ! -f $F77_TMP ] && setup_f77 #define OBD_FAIL_OSC_CHECKSUM_SEND 0x409 - sysctl -w lustre.fail_loc=0x80000409 + lctl set_param fail_loc=0x80000409 set_checksums 1 dd if=$F77_TMP of=$DIR/f77b bs=1M count=$F77SZ conv=sync || \ error "dd error: $?" - sysctl -w lustre.fail_loc=0 + lctl set_param fail_loc=0 set_checksums 0 + rm -f $DIR/f77b } run_test 77b "checksum error on client write ====================" test_77c() { # bug 10889 - [ ! -f $DIR/f77b ] && skip "requires 77b - skipping" && return + $GSS && skip "could not run with gss" && return + [ ! -f $DIR/f77b ] && skip "requires 77b - skipping" && return set_checksums 1 for algo in $CKSUM_TYPES; do cancel_lru_locks osc set_checksum_type $algo #define OBD_FAIL_OSC_CHECKSUM_RECEIVE 0x408 - sysctl -w lustre.fail_loc=0x80000408 + lctl set_param fail_loc=0x80000408 cmp $F77_TMP $DIR/f77b || error "file compare failed" - sysctl -w lustre.fail_loc=0 + lctl set_param fail_loc=0 done set_checksums 0 set_checksum_type $ORIG_CSUM_TYPE @@ -3311,39 +3490,42 @@ test_77c() { # bug 10889 run_test 77c "checksum error on client read ===================" test_77d() { # bug 10889 + $GSS && skip "could not run with gss" && return #define OBD_FAIL_OSC_CHECKSUM_SEND 0x409 - sysctl -w lustre.fail_loc=0x80000409 + lctl set_param fail_loc=0x80000409 set_checksums 1 directio write $DIR/f77 0 $F77SZ $((1024 * 1024)) || \ error "direct write: rc=$?" - sysctl -w lustre.fail_loc=0 + lctl set_param fail_loc=0 set_checksums 0 } run_test 77d "checksum error on OST direct write ===============" test_77e() { # bug 10889 - [ ! -f $DIR/f77 ] && skip "requires 77d - skipping" && return + $GSS && skip "could not run with gss" && return + [ ! -f $DIR/f77 ] && skip "requires 77d - skipping" && return #define OBD_FAIL_OSC_CHECKSUM_RECEIVE 0x408 - sysctl -w lustre.fail_loc=0x80000408 + lctl set_param fail_loc=0x80000408 set_checksums 1 cancel_lru_locks osc directio read $DIR/f77 0 $F77SZ $((1024 * 1024)) || \ error "direct read: rc=$?" - sysctl -w lustre.fail_loc=0 + lctl set_param fail_loc=0 set_checksums 0 } run_test 77e "checksum error on OST direct read ================" test_77f() { # bug 10889 + $GSS && skip "could not run with gss" && return set_checksums 1 for algo in $CKSUM_TYPES; do cancel_lru_locks osc set_checksum_type $algo #define OBD_FAIL_OSC_CHECKSUM_SEND 0x409 - sysctl -w lustre.fail_loc=0x409 + lctl set_param fail_loc=0x409 directio write $DIR/f77 0 $F77SZ $((1024 * 1024)) && \ error "direct write succeeded" - sysctl -w lustre.fail_loc=0 + lctl set_param fail_loc=0 done set_checksum_type $ORIG_CSUM_TYPE set_checksums 0 @@ -3351,38 +3533,43 @@ test_77f() { # bug 10889 run_test 77f "repeat checksum error on write (expect error) ====" test_77g() { # bug 10889 - [ $(lctl get_param -n devices | grep -c obdfilter) -eq 0 ] && \ - skip "remote OST" && return + $GSS && skip "could not run with gss" && return + remote_ost_nodsh && skip "remote OST with nodsh" && return + [ ! -f $F77_TMP ] && setup_f77 + + $SETSTRIPE $DIR/f77g -c 1 -i 0 #define OBD_FAIL_OST_CHECKSUM_RECEIVE 0x21a - sysctl -w lustre.fail_loc=0x8000021a + do_facet ost1 lctl set_param fail_loc=0x8000021a set_checksums 1 - dd if=$F77_TMP of=$DIR/f77 bs=1M count=$F77SZ || \ + dd if=$F77_TMP of=$DIR/f77g bs=1M count=$F77SZ || \ error "write error: rc=$?" - sysctl -w lustre.fail_loc=0 + do_facet ost1 lctl set_param fail_loc=0 set_checksums 0 } run_test 77g "checksum error on OST write ======================" test_77h() { # bug 10889 - [ $(lctl get_param -n devices | grep -c obdfilter) -eq 0 ] && \ - skip "remote OST" && return - [ ! -f $DIR/f77 ] && skip "requires 77g - skipping" && return + $GSS && skip "could not run with gss" && return + remote_ost_nodsh && skip "remote OST with nodsh" && return + + [ ! -f $DIR/f77g ] && skip "requires 77g - skipping" && return cancel_lru_locks osc #define OBD_FAIL_OST_CHECKSUM_SEND 0x21b - sysctl -w lustre.fail_loc=0x8000021b + do_facet ost1 lctl set_param fail_loc=0x8000021b set_checksums 1 - cmp $F77_TMP $DIR/f77 || error "file compare failed" - sysctl -w lustre.fail_loc=0 + cmp $F77_TMP $DIR/f77g || error "file compare failed" + do_facet ost1 lctl set_param fail_loc=0 set_checksums 0 } run_test 77h "checksum error on OST read =======================" test_77i() { # bug 13805 + $GSS && skip "could not run with gss" && return #define OBD_FAIL_OSC_CONNECT_CKSUM 0x40b - sysctl -w lustre.fail_loc=0x40b + lctl set_param fail_loc=0x40b remount_client $MOUNT - sysctl -w lustre.fail_loc=0 + lctl set_param fail_loc=0 for VALUE in `lctl get_param osc.*osc-[^mM]*.checksum_type`; do PARAM=`echo ${VALUE[0]} | cut -d "=" -f1` algo=`lctl get_param -n $PARAM | sed 's/.*\[\(.*\)\].*/\1/g'` @@ -3393,10 +3580,12 @@ test_77i() { # bug 13805 run_test 77i "client not supporting OSD_CONNECT_CKSUM ==========" test_77j() { # bug 13805 + $GSS && skip "could not run with gss" && return #define OBD_FAIL_OSC_CKSUM_ADLER_ONLY 0x40c - sysctl -w lustre.fail_loc=0x40c + lctl set_param fail_loc=0x40c remount_client $MOUNT - sysctl -w lustre.fail_loc=0 + lctl set_param fail_loc=0 + sleep 2 # wait async osc connect to finish for VALUE in `lctl get_param osc.*osc-[^mM]*.checksum_type`; do PARAM=`echo ${VALUE[0]} | cut -d "=" -f1` algo=`lctl get_param -n $PARAM | sed 's/.*\[\(.*\)\].*/\1/g'` @@ -3418,7 +3607,12 @@ test_78() { # bug 10901 echo "MemTotal: $MEMTOTAL" # reserve 256MB of memory for the kernel and other running processes, # and then take 1/2 of the remaining memory for the read/write buffers. - MEMTOTAL=$(((MEMTOTAL - 256 ) / 2)) + if [ $MEMTOTAL -gt 512 ] ;then + MEMTOTAL=$(((MEMTOTAL - 256 ) / 2)) + else + # for those poor memory-starved high-end clusters... + MEMTOTAL=$((MEMTOTAL / 2)) + fi echo "Mem to use for directio: $MEMTOTAL" [ $F78SIZE -gt $MEMTOTAL ] && F78SIZE=$MEMTOTAL [ $F78SIZE -gt 512 ] && F78SIZE=512 @@ -3428,11 +3622,12 @@ test_78() { # bug 10901 [ $SMALLESTOST -lt 10240 ] && \ skip "too small OSTSIZE, useless to run large O_DIRECT test" && return 0 - [ $F78SIZE -gt $((SMALLESTOST * $OSTCOUNT / 1024 - 5)) ] && \ - F78SIZE=$((SMALLESTOST * $OSTCOUNT / 1024 - 5)) + [ $F78SIZE -gt $((SMALLESTOST * $OSTCOUNT / 1024 - 80)) ] && \ + F78SIZE=$((SMALLESTOST * $OSTCOUNT / 1024 - 80)) + [ "$SLOW" = "no" ] && NSEQ=1 && [ $F78SIZE -gt 32 ] && F78SIZE=32 echo "File size: $F78SIZE" - $SETSTRIPE $DIR/$tfile -c -1 || error "setstripe failed" + $SETSTRIPE $DIR/$tfile -c $OSTCOUNT || error "setstripe failed" for i in `seq 1 $NSEQ` do FSIZE=$(($F78SIZE / ($NSEQ - $i + 1))) @@ -3445,14 +3640,12 @@ test_78() { # bug 10901 run_test 78 "handle large O_DIRECT writes correctly ============" test_79() { # bug 12743 - [ $(lctl get_param -n devices | grep -c obdfilter) -eq 0 ] && - skip "skipping test for remote OST" && return - wait_delete_completed - BKTOTAL=`lctl get_param -n obdfilter.*.kbytestotal | awk 'BEGIN{total=0}; {total+=$1}; END{print total}'` - BKFREE=`lctl get_param -n obdfilter.*.kbytesfree | awk 'BEGIN{free=0}; {free+=$1}; END{print free}'` - BKAVAIL=`lctl get_param -n obdfilter.*.kbytesavail | awk 'BEGIN{avail=0}; {avail+=$1}; END{print avail}'` + BKTOTAL=$(calc_osc_kbytes kbytestotal) + BKFREE=$(calc_osc_kbytes kbytesfree) + BKAVAIL=$(calc_osc_kbytes kbytesavail) + STRING=`df -P $MOUNT | tail -n 1 | awk '{print $2","$3","$4}'` DFTOTAL=`echo $STRING | cut -d, -f1` DFUSED=`echo $STRING | cut -d, -f2` @@ -3461,40 +3654,51 @@ test_79() { # bug 12743 ALLOWANCE=$((64 * $OSTCOUNT)) - if [ $DFTOTAL -lt $(($BKTOTAL - $ALLOWANCE)) ] || + if [ $DFTOTAL -lt $(($BKTOTAL - $ALLOWANCE)) ] || [ $DFTOTAL -gt $(($BKTOTAL + $ALLOWANCE)) ] ; then error "df total($DFTOTAL) mismatch OST total($BKTOTAL)" fi - if [ $DFFREE -lt $(($BKFREE - $ALLOWANCE)) ] || + if [ $DFFREE -lt $(($BKFREE - $ALLOWANCE)) ] || [ $DFFREE -gt $(($BKFREE + $ALLOWANCE)) ] ; then error "df free($DFFREE) mismatch OST free($BKFREE)" fi - if [ $DFAVAIL -lt $(($BKAVAIL - $ALLOWANCE)) ] || + if [ $DFAVAIL -lt $(($BKAVAIL - $ALLOWANCE)) ] || [ $DFAVAIL -gt $(($BKAVAIL + $ALLOWANCE)) ] ; then error "df avail($DFAVAIL) mismatch OST avail($BKAVAIL)" fi } run_test 79 "df report consistency check =======================" -# on the LLNL clusters, runas will still pick up root's $TMP settings, -# which will not be writable for the runas user, and then you get a CVS -# error message with a corrupt path string (CVS bug) and panic. -# We're not using much space, so just stick it in /tmp, which is safe. -OLDTMPDIR=$TMPDIR -OLDTMP=$TMP -TMPDIR=/tmp -TMP=/tmp -OLDHOME=$HOME -[ $RUNAS_ID -ne $UID ] && HOME=/tmp +test_80() { # bug 10718 + dd if=/dev/zero of=$DIR/$tfile bs=1M count=1 seek=1M + sync; sleep 1; sync + local BEFORE=`date +%s` + cancel_lru_locks osc + local AFTER=`date +%s` + local DIFF=$((AFTER-BEFORE)) + if [ $DIFF -gt 1 ] ; then + error "elapsed for 1M@1T = $DIFF" + fi + true + rm -f $DIR/$tfile +} +run_test 80 "Page eviction is equally fast at high offsets too ====" test_99a() { + [ -z "$(which cvs 2>/dev/null)" ] && skip "could not find cvs" && \ + return mkdir -p $DIR/d99cvsroot chown $RUNAS_ID $DIR/d99cvsroot + local oldPWD=$PWD # bug 13584, use $TMP as working dir + cd $TMP + $RUNAS cvs -d $DIR/d99cvsroot init || error + cd $oldPWD } run_test 99a "cvs init =========================================" test_99b() { + [ -z "$(which cvs 2>/dev/null)" ] && skip "could not find cvs" && return [ ! -d $DIR/d99cvsroot ] && test_99a cd /etc/init.d # some versions of cvs import exit(1) when asked to import links or @@ -3507,6 +3711,7 @@ test_99b() { run_test 99b "cvs import =======================================" test_99c() { + [ -z "$(which cvs 2>/dev/null)" ] && skip "could not find cvs" && return [ ! -d $DIR/d99cvsroot ] && test_99b cd $DIR mkdir -p $DIR/d99reposname @@ -3516,6 +3721,7 @@ test_99c() { run_test 99c "cvs checkout =====================================" test_99d() { + [ -z "$(which cvs 2>/dev/null)" ] && skip "could not find cvs" && return [ ! -d $DIR/d99cvsroot ] && test_99c cd $DIR/d99reposname $RUNAS touch foo99 @@ -3524,6 +3730,7 @@ test_99d() { run_test 99d "cvs add ==========================================" test_99e() { + [ -z "$(which cvs 2>/dev/null)" ] && skip "could not find cvs" && return [ ! -d $DIR/d99cvsroot ] && test_99c cd $DIR/d99reposname $RUNAS cvs update @@ -3531,17 +3738,30 @@ test_99e() { run_test 99e "cvs update =======================================" test_99f() { + [ -z "$(which cvs 2>/dev/null)" ] && skip "could not find cvs" && return [ ! -d $DIR/d99cvsroot ] && test_99d cd $DIR/d99reposname $RUNAS cvs commit -m 'nomsg' foo99 + rm -fr $DIR/d99cvsroot } run_test 99f "cvs commit =======================================" test_100() { - netstat -tna | while read PROT SND RCV LOCAL REMOTE STAT; do + [ "$NETTYPE" = tcp ] || \ + { skip "TCP secure port test, not useful for NETTYPE=$NETTYPE" && \ + return ; } + + remote_ost_nodsh && skip "remote OST with nodsh" && return + remote_mds_nodsh && skip "remote MDS with nodsh" && return + remote_servers || \ + { skip "useless for local single node setup" && return; } + + netstat -tna | ( rc=1; while read PROT SND RCV LOCAL REMOTE STAT; do [ "$PROT" != "tcp" ] && continue - RPORT=`echo $REMOTE | cut -d: -f2` + RPORT=$(echo $REMOTE | cut -d: -f2) [ "$RPORT" != "$ACCEPTOR_PORT" ] && continue + + rc=0 LPORT=`echo $LOCAL | cut -d: -f2` if [ $LPORT -ge 1024 ]; then echo "bad: $PROT $SND $RCV $LOCAL $REMOTE $STAT" @@ -3549,7 +3769,7 @@ test_100() { error "local: $LPORT > 1024, remote: $RPORT" fi done - true + [ "$rc" = 0 ] || error "privileged port not found" ) } run_test 100 "check local port using privileged port ===========" @@ -3591,58 +3811,134 @@ test_101() { # randomly read 10000 of 64K chunks from file 3x 32MB in size # echo "nreads: $nreads file size: $((cache_limit * 3))MB" - $RANDOM_READS -f $DIR/$tfile -s$((cache_limit * 3192 * 1024)) -b65536 -C -n$nreads -t 180 + $READS -f $DIR/$tfile -s$((cache_limit * 3192 * 1024)) -b65536 -C -n$nreads -t 180 discard=0 - for s in `lctl get_param -n llite.*.read_ahead_stats | get_named_value 'read but discarded'`; do - discard=$(($discard + $s)) + for s in `lctl get_param -n llite.*.read_ahead_stats | \ + get_named_value 'read but discarded' | cut -d" " -f1`; do + discard=$(($discard + $s)) done cleanup_101 if [ $(($discard * 10)) -gt $nreads ] ;then lctl get_param osc.*-osc*.rpc_stats lctl get_param llite.*.read_ahead_stats - error "too many ($discard) discarded pages" + error "too many ($discard) discarded pages" fi rm -f $DIR/$tfile || true } run_test 101 "check read-ahead for random reads ================" +export SETUP_TEST101=no +setup_test101() { + [ "$SETUP_TEST101" = "yes" ] && return + mkdir -p $DIR/$tdir + STRIPE_SIZE=1048576 + STRIPE_COUNT=$OSTCOUNT + STRIPE_OFFSET=0 + + trap cleanup_test101 EXIT + # prepare the read-ahead file + $SETSTRIPE $DIR/$tfile -s $STRIPE_SIZE -i $STRIPE_OFFSET -c $OSTCOUNT + + dd if=/dev/zero of=$DIR/$tfile bs=1024k count=100 2> /dev/null + SETUP_TEST101=yes +} + +cleanup_test101() { + [ "$SETUP_TEST101" = "yes" ] || return + trap 0 + rm -rf $DIR/$tdir + rm -f $DIR/$tfile + SETUP_TEST101=no +} + +calc_total() { + awk 'BEGIN{total=0}; {total+=$1}; END{print total}' +} + +ra_check_101() { + local READ_SIZE=$1 + local STRIPE_SIZE=1048576 + local RA_INC=1048576 + local STRIDE_LENGTH=$((STRIPE_SIZE/READ_SIZE)) + local FILE_LENGTH=$((64*100)) + local discard_limit=$((((STRIDE_LENGTH - 1)*3/(STRIDE_LENGTH*OSTCOUNT))* \ + (STRIDE_LENGTH*OSTCOUNT - STRIDE_LENGTH))) + DISCARD=`$LCTL get_param -n llite.*.read_ahead_stats | \ + get_named_value 'read but discarded' | \ + cut -d" " -f1 | calc_total` + + if [ $DISCARD -gt $discard_limit ]; then + lctl get_param llite.*.read_ahead_stats + error "Too many ($DISCARD) discarded pages with size (${READ_SIZE})" + else + echo "Read-ahead success for size ${READ_SIZE}" + fi +} + +test_101b() { + [ "$OSTCOUNT" -lt "2" ] && skip "skipping stride IO stride-ahead test" && return + local STRIPE_SIZE=1048576 + local STRIDE_SIZE=$((STRIPE_SIZE*OSTCOUNT)) + local FILE_LENGTH=$((STRIPE_SIZE*100)) + local ITERATION=$((FILE_LENGTH/STRIDE_SIZE)) + # prepare the read-ahead file + setup_test101 + cancel_lru_locks osc + for BIDX in 2 4 8 16 32 64 128 256 + do + local BSIZE=$((BIDX*4096)) + local READ_COUNT=$((STRIPE_SIZE/BSIZE)) + local STRIDE_LENGTH=$((STRIDE_SIZE/BSIZE)) + local OFFSET=$((STRIPE_SIZE/BSIZE*(OSTCOUNT - 1))) + $LCTL set_param -n llite.*.read_ahead_stats 0 + $READS -f $DIR/$tfile -l $STRIDE_LENGTH -o $OFFSET \ + -s $FILE_LENGTH -b $STRIPE_SIZE -a $READ_COUNT -n $ITERATION + cancel_lru_locks osc + ra_check_101 $BSIZE + done + cleanup_test101 + true +} +run_test 101b "check stride-io mode read-ahead =================" + export SETUP_TEST102=no setup_test102() { [ "$SETUP_TEST102" = "yes" ] && return mkdir -p $DIR/$tdir + chown $RUNAS_ID $DIR/$tdir STRIPE_SIZE=65536 - STRIPE_COUNT=4 - STRIPE_OFFSET=2 + STRIPE_OFFSET=1 + STRIPE_COUNT=$OSTCOUNT + [ $OSTCOUNT -gt 4 ] && STRIPE_COUNT=4 trap cleanup_test102 EXIT cd $DIR - $SETSTRIPE $tdir -s $STRIPE_SIZE -i $STRIPE_OFFSET -c $STRIPE_COUNT - cd $DIR/$tdir + $1 $SETSTRIPE $tdir -s $STRIPE_SIZE -i $STRIPE_OFFSET -c $STRIPE_COUNT + cd $DIR/$tdir for num in 1 2 3 4 do - for count in 1 2 3 4 + for count in `seq 1 $STRIPE_COUNT` do - for offset in 0 1 2 3 + for offset in `seq 0 $[$STRIPE_COUNT - 1]` do local stripe_size=`expr $STRIPE_SIZE \* $num` local file=file"$num-$offset-$count" - $SETSTRIPE $file -s $stripe_size -i $offset -c $count + $1 $SETSTRIPE $file -s $stripe_size -i $offset -c $count done done done cd $DIR - star -c f=$TMP/f102.tar $tdir + $1 $TAR cf $TMP/f102.tar $tdir --xattrs SETUP_TEST102=yes } cleanup_test102() { - [ "$SETUP_TEST102" = "yes" ] || return trap 0 + [ "$SETUP_TEST102" = "yes" ] || return 0 rm -f $TMP/f102.tar - rm -rf $DIR/$tdir SETUP_TEST102=no } @@ -3661,7 +3957,7 @@ test_102a() { setfattr -n trusted.name1 -v value1 $testfile || error [ "`getfattr -n trusted.name1 $testfile 2> /dev/null | \ grep "trusted.name1"`" == "trusted.name1=\"value1\"" ] || error - + setfattr -n user.author1 -v author1 $testfile || error [ "`getfattr -n user.author1 $testfile 2> /dev/null | \ grep "user.author1"`" == "user.author1=\"author1\"" ] || error @@ -3672,7 +3968,7 @@ test_102a() { [ `getfattr -d -m "^trusted" $testfile 2> /dev/null | \ grep "trusted.name" | wc -l` -eq 3 ] || error - + setfattr -n user.author2 -v author2 $testfile || error setfattr -n user.author3 -v author3 $testfile || error [ `getfattr -d -m "^user" $testfile 2> /dev/null | \ @@ -3716,6 +4012,7 @@ test_102b() { local stripe_count=`grep "count" $tmp_file| awk '{print $2}'` [ "$stripe_size" -eq 65536 ] || error "stripe size $stripe_size != 65536" [ "$stripe_count" -eq 2 ] || error "stripe count $stripe_count != 2" + rm -f $DIR/$tfile } run_test 102b "getfattr/setfattr for trusted.lov EAs ============" @@ -3745,42 +4042,25 @@ test_102c() { } run_test 102c "non-root getfattr/setfattr for lustre.lov EAs ===========" -get_stripe_info() { - stripe_size=0 - stripe_count=0 - stripe_offset=0 - local lines=`sed -n '/obdidx/=' $1` - stripe_size=`awk '{if($1~/size/) print $2}' $1` - stripe_count=`awk '{if($1~/count/) print $2}' $1` - lines=`expr $lines + 1` - stripe_offset=`sed -n ${lines}p $1 |awk '{print $1}'` -} - compare_stripe_info1() { for num in 1 2 3 4 do - for count in 1 2 3 4 + for count in `seq 1 $STRIPE_COUNT` do - for offset in 0 1 2 3 + for offset in `seq 0 $[$STRIPE_COUNT - 1]` do local size=`expr $STRIPE_SIZE \* $num` local file=file"$num-$offset-$count" - local tmp_file=out - $GETSTRIPE -v $file > $tmp_file - get_stripe_info $tmp_file - if test $stripe_size -ne $size - then - error "$file: different stripe size" && return + get_stripe_info client $PWD/$file "$1" + if [ $stripe_size -ne $size ]; then + error "$file: different stripe size $stripe_size, expected $size" && return fi - if test $stripe_count -ne $count - then - error "$file: different stripe count" && return + if [ $stripe_count -ne $count ]; then + error "$file: different stripe count $stripe_count, expected $count" && return fi - if test $stripe_offset -ne 0 - then - error "$file: different stripe offset" && return + if [ $stripe_index -ne 0 ]; then + error "$file: different stripe offset $stripe_index, expected 0" && return fi - rm -f $tmp_file done done done @@ -3789,91 +4069,57 @@ compare_stripe_info1() { compare_stripe_info2() { for num in 1 2 3 4 do - for count in 1 2 3 4 + for count in `seq 1 $STRIPE_COUNT` do - for offset in 0 1 2 3 + for offset in `seq 0 $[$STRIPE_COUNT - 1]` do local size=`expr $STRIPE_SIZE \* $num` local file=file"$num-$offset-$count" - local tmp_file=out - $GETSTRIPE -v $file > $tmp_file - get_stripe_info $tmp_file - if test $stripe_size -ne $size - then - error "$file: different stripe size" && return + get_stripe_info client $PWD/$file + if [ $stripe_size -ne $size ]; then + error "$file: different stripe size $stripe_size, expected $size" && return fi - if test $stripe_count -ne $count - then - error "$file: different stripe count" && return + if [ $stripe_count -ne $count ]; then + error "$file: different stripe count $stripe_count, expected $count" && return fi - if test $stripe_offset -ne $offset - then - error "$file: different stripe offset" && return + if [ $stripe_index -ne $offset ]; then + error "$file: different stripe offset $stripe_index, expected $offset" && return fi - rm -f $tmp_file done done done } +find_lustre_tar() { + [ -n "$(which tar 2>/dev/null)" ] && strings $(which tar) | grep -q lustre && echo tar +} + test_102d() { - # b10930: star test for trusted.lov xattr - star --xhelp 2>&1 | grep -q nolustre - if [ $? -ne 0 ] - then - skip "being skipped because a lustre-aware star is not installed." && return - fi - [ "$OSTCOUNT" -lt "4" ] && skip "skipping 4-stripe test" && return + # b10930: tar test for trusted.lov xattr + TAR=$(find_lustre_tar) + [ -z "$TAR" ] && skip "lustre-aware tar is not installed" && return + [ "$OSTCOUNT" -lt "2" ] && skip "skipping N-stripe test" && return setup_test102 mkdir -p $DIR/d102d - star -x f=$TMP/f102.tar -C $DIR/d102d + $TAR xf $TMP/f102.tar -C $DIR/d102d --xattrs cd $DIR/d102d/$tdir compare_stripe_info1 - -} -run_test 102d "star restore stripe info from tarfile,not keep osts ===========" - -test_102e() { - # b10930: star test for trusted.lov xattr - star --xhelp 2>&1 | grep -q nolustre - [ $? -ne 0 ] && skip "lustre-aware star is not installed" && return - [ "$OSTCOUNT" -lt "4" ] && skip "skipping 4-stripe test" && return - setup_test102 - mkdir -p $DIR/d102e - star -x -preserve-osts f=$TMP/f102.tar -C $DIR/d102e - cd $DIR/d102e/$tdir - compare_stripe_info2 } -run_test 102e "star restore stripe info from tarfile, keep osts ===========" +run_test 102d "tar restore stripe info from tarfile,not keep osts ===========" test_102f() { - # b10930: star test for trusted.lov xattr - star --xhelp 2>&1 | grep -q nolustre - [ $? -ne 0 ] && skip "lustre-aware star is not installed" && return - [ "$OSTCOUNT" -lt "4" ] && skip "skipping 4-stripe test" && return + # b10930: tar test for trusted.lov xattr + TAR=$(find_lustre_tar) + [ -z "$TAR" ] && skip "lustre-aware tar is not installed" && return + [ "$OSTCOUNT" -lt "2" ] && skip "skipping N-stripe test" && return setup_test102 mkdir -p $DIR/d102f cd $DIR - star -copy $tdir $DIR/d102f + $TAR cf - --xattrs $tdir | $TAR xf - --xattrs -C $DIR/d102f cd $DIR/d102f/$tdir compare_stripe_info1 } -run_test 102f "star copy files, not keep osts ===========" - -test_102g() { - # b10930: star test for trusted.lov xattr - star --xhelp 2>&1 | grep -q nolustre - [ $? -ne 0 ] && skip "lustre-aware star is not installed" && return - [ "$OSTCOUNT" -lt "4" ] && skip "skipping 4-stripe test" && return - setup_test102 - mkdir -p $DIR/d102g - cd $DIR - star -copy -preserve-osts $tdir $DIR/d102g - cd $DIR/d102g/$tdir - compare_stripe_info2 - cleanup_test102 -} -run_test 102g "star copy files, keep osts ===========" +run_test 102f "tar copy files, not keep osts ===========" test_102h() { # bug 15777 [ -z $(lctl get_param -n mdc.*.connect_flags | grep xattr) ] && @@ -3916,9 +4162,34 @@ test_102h() { # bug 15777 error "$XBIG different after growing $XSML" fi log "$XBIG still valid after growing $XSML" + rm -f $file } run_test 102h "grow xattr from inside inode to external block" +test_102i() { # bug 17038 + touch $DIR/$tfile + ln -s $DIR/$tfile $DIR/${tfile}link + getfattr -n trusted.lov $DIR/$tfile || error "lgetxattr on $DIR/$tfile failed" + getfattr -h -n trusted.lov $DIR/${tfile}link 2>&1 | grep -i "no such attr" || error "error for lgetxattr on $DIR/${tfile}link is not ENODATA" + rm -f $DIR/$tfile $DIR/${tfile}link +} +run_test 102i "lgetxattr test on symbolic link ============" + +test_102j() { + TAR=$(find_lustre_tar) + [ -z "$TAR" ] && skip "lustre-aware tar is not installed" && return + [ "$OSTCOUNT" -lt "2" ] && skip "skipping N-stripe test" && return + setup_test102 "$RUNAS" + mkdir -p $DIR/d102j + chown $RUNAS_ID $DIR/d102j + $RUNAS $TAR xf $TMP/f102.tar -C $DIR/d102j --xattrs + cd $DIR/d102j/$tdir + compare_stripe_info1 "$RUNAS" +} +run_test 102j "non-root tar restore stripe info from tarfile, not keep osts ===" + +cleanup_test102 + run_acl_subtest() { $LUSTRE/tests/acl/run $LUSTRE/tests/acl/$1.test @@ -3984,6 +4255,7 @@ test_104() { lfs df || error "lfs df with deactivated OSC failed" lctl --device %$OSC recover lfs df || error "lfs df with reactivated OSC failed" + rm -f $DIR/$tfile } run_test 104 "lfs df [-ih] [path] test =========================" @@ -3992,10 +4264,11 @@ test_105a() { touch $DIR/$tfile if [ -n "`mount | grep \"$DIR.*flock\" | grep -v noflock`" ]; then - flocks_test on -f $DIR/$tfile || error "fail flock on" + flocks_test 1 on -f $DIR/$tfile || error "fail flock on" else - flocks_test off -f $DIR/$tfile || error "fail flock off" + flocks_test 1 off -f $DIR/$tfile || error "fail flock off" fi + rm -f $DIR/$tfile } run_test 105a "flock when mounted without -o flock test ========" @@ -4003,10 +4276,11 @@ test_105b() { touch $DIR/$tfile if [ -n "`mount | grep \"$DIR.*flock\" | grep -v noflock`" ]; then - flocks_test on -c $DIR/$tfile || error "fail flock on" + flocks_test 1 on -c $DIR/$tfile || error "fail flock on" else - flocks_test off -c $DIR/$tfile || error "fail flock off" + flocks_test 1 off -c $DIR/$tfile || error "fail flock off" fi + rm -f $DIR/$tfile } run_test 105b "fcntl when mounted without -o flock test ========" @@ -4014,14 +4288,26 @@ test_105c() { touch $DIR/$tfile if [ -n "`mount | grep \"$DIR.*flock\" | grep -v noflock`" ]; then - flocks_test on -l $DIR/$tfile || error "fail flock on" + flocks_test 1 on -l $DIR/$tfile || error "fail flock on" else - flocks_test off -l $DIR/$tfile || error "fail flock off" + flocks_test 1 off -l $DIR/$tfile || error "fail flock off" fi + rm -f $DIR/$tfile } run_test 105c "lockf when mounted without -o flock test ========" +test_105d() { # bug 15924 + mkdir -p $DIR/$tdir + [ -z "`mount | grep \"$DIR.*flock\" | grep -v noflock`" ] && \ + skip "mount w/o flock enabled" && return + #define OBD_FAIL_LDLM_CP_CB_WAIT 0x315 + $LCTL set_param fail_loc=0x80000315 + flocks_test 2 $DIR/$tdir +} +run_test 105d "flock race (should not freeze) ========" + test_106() { #bug 10921 + mkdir -p $DIR/$tdir $DIR/$tdir && error "exec $DIR/$tdir succeeded" chmod 777 $DIR/$tdir || error "chmod $DIR/$tdir failed" } @@ -4030,14 +4316,19 @@ run_test 106 "attempt exec of dir followed by chown of that dir" test_107() { CDIR=`pwd` cd $DIR + + local file=core + rm -f $file + + local save_pattern=$(sysctl -n kernel.core_pattern) + local save_uses_pid=$(sysctl -n kernel.core_uses_pid) + sysctl -w kernel.core_pattern=$file + sysctl -w kernel.core_uses_pid=0 + ulimit -c unlimited sleep 60 & SLEEPPID=$! - file=`cat /proc/sys/kernel/core_pattern` - core_pid=`cat /proc/sys/kernel/core_uses_pid` - [ $core_pid -eq 1 ] && file=$file.$SLEEPPID - rm -f $file sleep 1 kill -s 11 $SLEEPPID @@ -4049,6 +4340,8 @@ test_107() { error "Fail to create core file $file" fi rm -f $file + sysctl -w kernel.core_pattern=$save_pattern + sysctl -w kernel.core_uses_pid=$save_uses_pid cd $CDIR } run_test 107 "Coredump on SIG" @@ -4061,6 +4354,7 @@ test_110() { touch $DIR/d110/yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy && error ""create with 256 char should fail, but not ls -l $DIR/d110 + rm -fr $DIR/d110 } run_test 110 "filename length checking" @@ -4084,7 +4378,7 @@ test_115() { # don't return an error [ $OSTIO_post -eq $OSTIO_pre ] && echo \ - "FAIL: No addition ll_ost_io threads were created ($OSTIO_pre)" &&\ + "WARNING: No new ll_ost_io threads were created ($OSTIO_pre)" &&\ echo "This may be fine, depending on what ran before this test" &&\ echo "and how fast this system is." && return @@ -4107,13 +4401,12 @@ free_min_max () { MINV=${AVAIL[i]}; MINI=$i fi done - echo Min free space: OST $MINI: $MINV - echo Max free space: OST $MAXI: $MAXV + echo Min free space: OST $MINI: $MINV + echo Max free space: OST $MAXI: $MAXV } test_116() { [ "$OSTCOUNT" -lt "2" ] && skip "$OSTCOUNT < 2 OSTs" && return - remote_mds && skip "remote MDS" && return echo -n "Free space priority " lctl get_param -n lov.*-clilov-*.qos_prio_free @@ -4148,7 +4441,7 @@ test_116() { echo "ok" else echo "failed - QOS mode won't be used" - error "QOS imbalance criteria not met" + error_ignore "QOS imbalance criteria not met" return fi @@ -4173,14 +4466,14 @@ test_116() { free_min_max DIFF2=$(($MAXV - $MINV)) echo "free space delta: orig $DIFF final $DIFF2" - [ $DIFF2 -gt $DIFF ] && echo "delta got worse!" + [ $DIFF2 -gt $DIFF ] && echo "delta got worse!" DIFF=$(($MINV1 - ${AVAIL[$MINI1]})) echo "Wrote $DIFF to smaller OST $MINI1" DIFF2=$(($MAXV1 - ${AVAIL[$MAXI1]})) echo "Wrote $DIFF2 to larger OST $MAXI1" [ $DIFF -gt 0 ] && echo "Wrote $(($DIFF2 * 100 / $DIFF - 100))% more data to larger OST $MAXI1" - # Figure out which files were written where + # Figure out which files were written where UUID=$(lctl get_param -n lov.${FSNAME}-clilov-*.target_obd | awk '/'$MINI1': / {print $2; exit}') echo $UUID @@ -4192,7 +4485,9 @@ test_116() { MAXC=$($GETSTRIPE --obd $UUID $DIR/$tdir | wc -l) echo "$MAXC files created on larger OST $MAXI1" [ $MINC -gt 0 ] && echo "Wrote $(($MAXC * 100 / $MINC - 100))% more files to larger OST $MAXI1" - [ $MAXC -gt $MINC ] || error "stripe QOS didn't balance free space" + [ $MAXC -gt $MINC ] || error_ignore "stripe QOS didn't balance free space" + + rm -rf $DIR/$tdir/OST${MINI} } run_test 116 "stripe QOS: free space balance ===================" @@ -4200,10 +4495,11 @@ test_117() # bug 10891 { dd if=/dev/zero of=$DIR/$tfile bs=1M count=1 #define OBD_FAIL_OST_SETATTR_CREDITS 0x21e - sysctl -w lustre.fail_loc=0x21e + lctl set_param fail_loc=0x21e > $DIR/$tfile || error "truncate failed" - sysctl -w lustre.fail_loc=0 + lctl set_param fail_loc=0 echo "Truncate succeeded." + rm -f $DIR/$tfile } run_test 117 "verify fsfilt_extend ==========" @@ -4222,7 +4518,7 @@ reset_async() { FILE=$DIR/reset_async # Ensure all OSCs are cleared - $LSTRIPE $FILE 0 -1 -1 + $LSTRIPE -c -1 $FILE dd if=/dev/zero of=$FILE bs=64k count=$OSTCOUNT sync rm $FILE @@ -4240,12 +4536,13 @@ test_118a() #bug 11710 error "Dirty pages not flushed to disk, dirty=$DIRTY, writeback=$WRITEBACK" return 1; fi + rm -f $DIR/$tfile } run_test 118a "verify O_SYNC works ==========" test_118b() { - remote_ost_nodsh && skip "remote OST" && return + remote_ost_nodsh && skip "remote OST with nodsh" && return reset_async @@ -4281,7 +4578,7 @@ run_test 118b "Reclaim dirty pages on fatal error ==========" test_118c() { - remote_ost_nodsh && skip "remote OST" && return + remote_ost_nodsh && skip "remote OST with nodsh" && return reset_async @@ -4325,7 +4622,7 @@ run_test 118c "Fsync blocks on EROFS until dirty pages are flushed ==========" test_118d() { - remote_ost_nodsh && skip "remote OST" && return + remote_ost_nodsh && skip "remote OST with nodsh" && return reset_async @@ -4366,7 +4663,7 @@ test_118f() { reset_async #define OBD_FAIL_OSC_BRW_PREP_REQ2 0x40a - sysctl -w lustre.fail_loc=0x8000040a + lctl set_param fail_loc=0x8000040a # Should simulate EINVAL error which is fatal multiop $DIR/$tfile oO_CREAT:O_RDWR:O_SYNC:w4096c @@ -4375,8 +4672,8 @@ test_118f() { error "Must return error due to dropped pages, rc=$RC" fi - sysctl -w lustre.fail_loc=0x0 - + lctl set_param fail_loc=0x0 + LOCKED=$(lctl get_param -n llite.*.dump_page_cache | grep -c locked) DIRTY=$(lctl get_param -n llite.*.dump_page_cache | grep -c dirty) WRITEBACK=$(lctl get_param -n llite.*.dump_page_cache | @@ -4398,24 +4695,24 @@ test_118f() { run_test 118f "Simulate unrecoverable OSC side error ==========" test_118g() { - reset_async + reset_async #define OBD_FAIL_OSC_BRW_PREP_REQ 0x406 - sysctl -w lustre.fail_loc=0x406 + lctl set_param fail_loc=0x406 # simulate local -ENOMEM - multiop $DIR/$tfile oO_CREAT:O_RDWR:O_SYNC:w4096c - RC=$? + multiop $DIR/$tfile oO_CREAT:O_RDWR:O_SYNC:w4096c + RC=$? - sysctl -w lustre.fail_loc=0 + lctl set_param fail_loc=0 if [[ $RC -eq 0 ]]; then error "Must return error due to dropped pages, rc=$RC" fi - LOCKED=$(lctl get_param -n llite.*.dump_page_cache | grep -c locked) - DIRTY=$(lctl get_param -n llite.*.dump_page_cache | grep -c dirty) - WRITEBACK=$(lctl get_param -n llite.*.dump_page_cache | - grep -c writeback) + LOCKED=$(lctl get_param -n llite.*.dump_page_cache | grep -c locked) + DIRTY=$(lctl get_param -n llite.*.dump_page_cache | grep -c dirty) + WRITEBACK=$(lctl get_param -n llite.*.dump_page_cache | + grep -c writeback) if [[ $LOCKED -ne 0 ]]; then error "Locked pages remain in cache, locked=$LOCKED" fi @@ -4427,13 +4724,13 @@ test_118g() { rm -f $DIR/$tfile echo "No pages locked after fsync" - reset_async + reset_async return 0 } run_test 118g "Don't stay in wait if we got local -ENOMEM ==========" test_118h() { - remote_ost_nodsh && skip "remote OST" && return + remote_ost_nodsh && skip "remote OST with nodsh" && return reset_async @@ -4468,7 +4765,7 @@ test_118h() { run_test 118h "Verify timeout in handling recoverables errors ==========" test_118i() { - remote_ost_nodsh && skip "remote OST" && return + remote_ost_nodsh && skip "remote OST with nodsh" && return reset_async @@ -4506,7 +4803,7 @@ test_118i() { run_test 118i "Fix error before timeout in recoverable error ==========" test_118j() { - remote_ost_nodsh && skip "remote OST" && return + remote_ost_nodsh && skip "remote OST with nodsh" && return reset_async @@ -4542,6 +4839,8 @@ run_test 118j "Simulate unrecoverable OST side error ==========" test_118k() { + remote_ost_nodsh && skip "remote OSTs with nodsh" && return + #define OBD_FAIL_OST_BRW_WRITE_BULK 0x20e set_nodes_failloc "$(osts_nodes)" 0x20e mkdir -p $DIR/$tdir @@ -4599,7 +4898,33 @@ test_119c() # bug 13099 } run_test 119c "Testing for direct read hitting hole" -test_120a() { +test_119d() # bug 15950 +{ + MAX_RPCS_IN_FLIGHT=`$LCTL get_param -n osc.*OST0000-osc-[^mM]*.max_rpcs_in_flight` + $LCTL set_param -n osc.*OST0000-osc-[^mM]*.max_rpcs_in_flight 1 + BSIZE=1048576 + $SETSTRIPE $DIR/$tfile -i 0 -c 1 || error "setstripe failed" + $DIRECTIO write $DIR/$tfile 0 1 $BSIZE || error "first directio failed" + #define OBD_FAIL_OSC_DIO_PAUSE 0x40d + lctl set_param fail_loc=0x40d + $DIRECTIO write $DIR/$tfile 1 4 $BSIZE & + pid_dio=$! + sleep 1 + cat $DIR/$tfile > /dev/null & + lctl set_param fail_loc=0 + pid_reads=$! + wait $pid_dio + log "the DIO writes have completed, now wait for the reads (should not block very long)" + sleep 2 + [ -n "`ps h -p $pid_reads -o comm`" ] && \ + error "the read rpcs have not completed in 2s" + rm -f $DIR/$tfile + $LCTL set_param -n osc.*OST0000-osc-[^mM]*.max_rpcs_in_flight $MAX_RPCS_IN_FLIGHT +} +run_test 119d "The DIO path should try to send a new rpc once one is completed" + +test_120a() { + mkdir -p $DIR/$tdir [ -z "`lctl get_param -n mdc.*.connect_flags | grep early_lock_cancel`" ] && \ skip "no early lock cancel on server" && return 0 lru_resize_disable mdc @@ -4619,6 +4944,7 @@ test_120a() { run_test 120a "Early Lock Cancel: mkdir test" test_120b() { + mkdir -p $DIR/$tdir [ -z "`lctl get_param -n mdc.*.connect_flags | grep early_lock_cancel`" ] && \ skip "no early lock cancel on server" && return 0 lru_resize_disable mdc @@ -4638,6 +4964,7 @@ test_120b() { run_test 120b "Early Lock Cancel: create test" test_120c() { + mkdir -p $DIR/$tdir [ -z "`lctl get_param -n mdc.*.connect_flags | grep early_lock_cancel`" ] && \ skip "no early lock cancel on server" && return 0 lru_resize_disable mdc @@ -4659,6 +4986,7 @@ test_120c() { run_test 120c "Early Lock Cancel: link test" test_120d() { + mkdir -p $DIR/$tdir [ -z "`lctl get_param -n mdc.*.connect_flags | grep early_lock_cancel`" ] && \ skip "no early lock cancel on server" && return 0 lru_resize_disable mdc @@ -4679,6 +5007,7 @@ test_120d() { run_test 120d "Early Lock Cancel: setattr test" test_120e() { + mkdir -p $DIR/$tdir [ -z "`lctl get_param -n mdc.*.connect_flags | grep early_lock_cancel`" ] && \ skip "no early lock cancel on server" && return 0 lru_resize_disable mdc @@ -4707,6 +5036,7 @@ run_test 120e "Early Lock Cancel: unlink test" test_120f() { [ -z "`lctl get_param -n mdc.*.connect_flags | grep early_lock_cancel`" ] && \ skip "no early lock cancel on server" && return 0 + mkdir -p $DIR/$tdir lru_resize_disable mdc lru_resize_disable osc mkdir -p $DIR/$tdir/d1 $DIR/$tdir/d2 @@ -4778,53 +5108,92 @@ test_121() { #bug #10589 rm -rf $DIR/$tfile writes=$(LANG=C dd if=/dev/zero of=$DIR/$tfile count=1 2>&1 | awk -F '+' '/out/ {print $1}') #define OBD_FAIL_LDLM_CANCEL_RACE 0x310 - sysctl -w lustre.fail_loc=0x310 + lctl set_param fail_loc=0x310 cancel_lru_locks osc > /dev/null reads=$(LANG=C dd if=$DIR/$tfile of=/dev/null 2>&1 | awk -F '+' '/in/ {print $1}') - sysctl -w lustre.fail_loc=0 + lctl set_param fail_loc=0 [ "$reads" -eq "$writes" ] || error "read" $reads "blocks, must be" $writes } run_test 121 "read cancel race =========" test_123a() { # was test 123, statahead(bug 11401) + SLOWOK=0 if [ -z "$(grep "processor.*: 1" /proc/cpuinfo)" ]; then log "testing on UP system. Performance may be not as good as expected." + SLOWOK=1 fi - remount_client $MOUNT + rm -rf $DIR/$tdir mkdir -p $DIR/$tdir - error=0 NUMFREE=`df -i -P $DIR | tail -n 1 | awk '{ print $4 }'` [ $NUMFREE -gt 100000 ] && NUMFREE=100000 || NUMFREE=$((NUMFREE-1000)) MULT=10 - for ((i=1, j=0; i<=$NUMFREE; j=$i, i=$((i * MULT)) )); do + for ((i=100, j=0; i<=$NUMFREE; j=$i, i=$((i * MULT)) )); do createmany -o $DIR/$tdir/$tfile $j $((i - j)) - lctl get_param -n llite.*.statahead_max | grep '[0-9]' - cancel_lru_locks mdc - cancel_lru_locks osc - stime=`date +%s` - ls -l $DIR/$tdir > /dev/null - etime=`date +%s` - delta_sa=$((etime - stime)) - log "ls $i files with statahead: $delta_sa sec" - lctl get_param -n llite.*.statahead_stats - max=`lctl get_param -n llite.*.statahead_max | head -n 1` lctl set_param -n llite.*.statahead_max 0 lctl get_param llite.*.statahead_max cancel_lru_locks mdc cancel_lru_locks osc stime=`date +%s` - ls -l $DIR/$tdir > /dev/null + time ls -l $DIR/$tdir | wc -l etime=`date +%s` delta=$((etime - stime)) log "ls $i files without statahead: $delta sec" - lctl set_param llite.*.statahead_max=$max - if [ $delta_sa -gt $(($delta + 2)) ]; then - log "ls $i files is slower with statahead!" - error=1 + + swrong=`lctl get_param -n llite.*.statahead_stats | grep "statahead wrong:" | awk '{print $3}'` + lctl get_param -n llite.*.statahead_max | grep '[0-9]' + cancel_lru_locks mdc + cancel_lru_locks osc + stime=`date +%s` + time ls -l $DIR/$tdir | wc -l + etime=`date +%s` + delta_sa=$((etime - stime)) + log "ls $i files with statahead: $delta_sa sec" + lctl get_param -n llite.*.statahead_stats + ewrong=`lctl get_param -n llite.*.statahead_stats | grep "statahead wrong:" | awk '{print $3}'` + + if [ $swrong -lt $ewrong ]; then + log "statahead was stopped, maybe too many locks held!" + fi + + [ $delta -eq 0 ] && continue + + if [ $((delta_sa * 100)) -gt $((delta * 105)) ]; then + if [ $SLOWOK -eq 0 ]; then + error "ls $i files is slower with statahead!" + + max=`lctl get_param -n llite.*.statahead_max | head -n 1` + lctl set_param -n llite.*.statahead_max 0 + lctl get_param llite.*.statahead_max + cancel_lru_locks mdc + cancel_lru_locks osc + $LCTL dk > /dev/null + stime=`date +%s` + time ls -l $DIR/$tdir | wc -l + etime=`date +%s` + $LCTL dk > $TMP/sanity_test_123a_${i}_disable_${etime}.log + delta=$((etime - stime)) + log "ls $i files without statahead: $delta sec, dump to $TMP/sanity_test_123a_${i}_disable_${etime}.log" + lctl set_param llite.*.statahead_max=$max + + lctl get_param -n llite.*.statahead_max | grep '[0-9]' + cancel_lru_locks mdc + cancel_lru_locks osc + $LCTL dk > /dev/null + stime=`date +%s` + time ls -l $DIR/$tdir | wc -l + etime=`date +%s` + $LCTL dk > $TMP/sanity_test_123a_${i}_enable_${etime}.log + delta_sa=$((etime - stime)) + log "ls $i files with statahead: $delta_sa sec, dump to $TMP/sanity_test_123a_${i}_enable_${etime}.log" + lctl get_param -n llite.*.statahead_stats + else + log "ls $i files is slower with statahead!" + fi + break fi [ $delta -gt 20 ] && break @@ -4841,10 +5210,6 @@ test_123a() { # was test 123, statahead(bug 11401) log "rm -r $DIR/$tdir/: $delta seconds" log "rm done" lctl get_param -n llite.*.statahead_stats - # wait for commitment of removal - sleep 2 - [ $error -ne 0 ] && error "statahead is slow!" - return 0 } run_test 123a "verify statahead work" @@ -4856,10 +5221,10 @@ test_123b () { # statahead(bug 15027) cancel_lru_locks osc #define OBD_FAIL_MDC_GETATTR_ENQUEUE 0x803 - sysctl -w lustre.fail_loc=0x80000803 + lctl set_param fail_loc=0x80000803 ls -lR $DIR/$tdir > /dev/null log "ls done" - sysctl -w lustre.fail_loc=0x0 + lctl set_param fail_loc=0x0 lctl get_param -n llite.*.statahead_stats rm -r $DIR/$tdir sync @@ -4870,25 +5235,24 @@ run_test 123b "not panic with network error in statahead enqueue (bug 15027)" test_124a() { [ -z "`lctl get_param -n mdc.*.connect_flags | grep lru_resize`" ] && \ skip "no lru resize on server" && return 0 - NR=2000 + local NR=2000 mkdir -p $DIR/$tdir || error "failed to create $DIR/$tdir" - # use touch to produce $NR new locks log "create $NR files at $DIR/$tdir" - createmany -o $DIR/$tdir/f $NR || - error "failed to create $NR files in $DIR/$tdir" - + createmany -o $DIR/$tdir/f $NR || + error "failed to create $NR files in $DIR/$tdir" + cancel_lru_locks mdc ls -l $DIR/$tdir > /dev/null - NSDIR="" - LRU_SIZE=0 + local NSDIR="" + local LRU_SIZE=0 for VALUE in `lctl get_param ldlm.namespaces.*mdc-*.lru_size`; do - PARAM=`echo ${VALUE[0]} | cut -d "=" -f1` + local PARAM=`echo ${VALUE[0]} | cut -d "=" -f1` LRU_SIZE=$(lctl get_param -n $PARAM) if [ $LRU_SIZE -gt $(default_lru_size) ]; then NSDIR=$(echo $PARAM | cut -d "." -f1-3) - log "using $(basename $NSDIR) namespace" + log "NS=$(basename $NSDIR)" break fi done @@ -4897,59 +5261,83 @@ test_124a() { skip "Not enough cached locks created!" return 0 fi - log "created $LRU_SIZE lock(s)" - - # we want to sleep 30s to not make test too long - SLEEP=30 - SLEEP_ADD=2 - - # we know that lru resize allows one client to hold $LIMIT locks for 10h - MAX_HRS=10 - - # get the pool limit - LIMIT=`lctl get_param -n $NSDIR.pool.limit` - - # calculate lock volume factor taking into account data set size and the - # rule that number of locks will be getting smaller durring sleep interval - # and we need to additionally enforce LVF to take this into account. - # Use $LRU_SIZE_B here to take into account real number of locks created - # in the case of CMD, LRU_SIZE_B != $NR in most of cases - LVF=$(($MAX_HRS * 60 * 60 * $LIMIT / $SLEEP)) - LRU_SIZE_B=$LRU_SIZE - log "make client drop locks $LVF times faster so that ${SLEEP}s is enough to cancel $LRU_SIZE lock(s)" - OLD_LVF=`lctl get_param -n $NSDIR.pool.lock_volume_factor` + log "LRU=$LRU_SIZE" + + local SLEEP=30 + + # We know that lru resize allows one client to hold $LIMIT locks + # for 10h. After that locks begin to be killed by client. + local MAX_HRS=10 + local LIMIT=`lctl get_param -n $NSDIR.pool.limit` + + # Make LVF so higher that sleeping for $SLEEP is enough to _start_ + # killing locks. Some time was spent for creating locks. This means + # that up to the moment of sleep finish we must have killed some of + # them (10-100 locks). This depends on how fast ther were created. + # Many of them were touched in almost the same moment and thus will + # be killed in groups. + local LVF=$(($MAX_HRS * 60 * 60 / $SLEEP * $LIMIT / $LRU_SIZE)) + + # Use $LRU_SIZE_B here to take into account real number of locks + # created in the case of CMD, LRU_SIZE_B != $NR in most of cases + local LRU_SIZE_B=$LRU_SIZE + log "LVF=$LVF" + local OLD_LVF=`lctl get_param -n $NSDIR.pool.lock_volume_factor` lctl set_param -n $NSDIR.pool.lock_volume_factor $LVF - log "sleep for $((SLEEP+SLEEP_ADD))s" - sleep $((SLEEP+SLEEP_ADD)) + + # Let's make sure that we really have some margin. Client checks + # cached locks every 10 sec. + SLEEP=$((SLEEP+20)) + log "Sleep ${SLEEP} sec" + local SEC=0 + while ((SEC<$SLEEP)); do + echo -n "..." + sleep 5 + SEC=$((SEC+5)) + LRU_SIZE=`lctl get_param -n $NSDIR/lru_size` + echo -n "$LRU_SIZE" + done + echo "" lctl set_param -n $NSDIR.pool.lock_volume_factor $OLD_LVF - LRU_SIZE_A=`lctl get_param -n $NSDIR.lru_size` + local LRU_SIZE_A=`lctl get_param -n $NSDIR/lru_size` [ $LRU_SIZE_B -gt $LRU_SIZE_A ] || { - error "No locks dropped in "$((SLEEP+SLEEP_ADD))"s. LRU size: $LRU_SIZE_A" + error "No locks dropped in ${SLEEP}s. LRU size: $LRU_SIZE_A" unlinkmany $DIR/$tdir/f $NR return } - log "Dropped "$((LRU_SIZE_B-LRU_SIZE_A))" locks in "$((SLEEP+SLEEP_ADD))"s" + log "Dropped "$((LRU_SIZE_B-LRU_SIZE_A))" locks in ${SLEEP}s" log "unlink $NR files at $DIR/$tdir" unlinkmany $DIR/$tdir/f $NR } run_test 124a "lru resize =======================================" +get_max_pool_limit() +{ + local limit=`lctl get_param -n ldlm.namespaces.*-MDT0000-mdc-*.pool.limit` + local max=0 + for l in $limit; do + if test $l -gt $max; then + max=$l + fi + done + echo $max +} + test_124b() { [ -z "`lctl get_param -n mdc.*.connect_flags | grep lru_resize`" ] && \ skip "no lru resize on server" && return 0 - # even for cmd no matter what metadata namespace to use for getting - # the limit, we use appropriate. - LIMIT=`lctl get_param -n ldlm.namespaces.*mdc*.pool.limit` + LIMIT=`get_max_pool_limit` NR=$(($(default_lru_size)*20)) if [ $NR -gt $LIMIT ]; then + log "Limit lock number by $LIMIT locks" NR=$LIMIT fi lru_resize_disable mdc - mkdir -p $DIR/$tdir/disable_lru_resize || + mkdir -p $DIR/$tdir/disable_lru_resize || error "failed to create $DIR/$tdir/disable_lru_resize" createmany -o $DIR/$tdir/disable_lru_resize/f $NR @@ -4973,7 +5361,7 @@ test_124b() { unlinkmany $DIR/$tdir/disable_lru_resize/f $NR lru_resize_enable mdc - mkdir -p $DIR/$tdir/enable_lru_resize || + mkdir -p $DIR/$tdir/enable_lru_resize || error "failed to create $DIR/$tdir/enable_lru_resize" createmany -o $DIR/$tdir/enable_lru_resize/f $NR @@ -5019,6 +5407,8 @@ run_test 125 "don't return EPROTO when a dir has a non-default striping and ACLs test_126() { # bug 12829/13455 [ -z "$(lctl get_param -n llite.*.client_type | grep local)" ] && skip "must run as local client" && return [ "$UID" != 0 ] && echo "skipping $TESTNAME (must run as root)" && return + $GSS && skip "must run as gss disabled" && return + $RUNAS -u 0 -g 1 touch $DIR/$tfile || error "touch failed" gid=`ls -n $DIR/$tfile | awk '{print $4}'` rm -f $DIR/$tfile @@ -5036,19 +5426,20 @@ test_127() { # bug 15521 $LCTL get_param osc.*0000-osc-*.stats | grep samples > $DIR/${tfile}.tmp while read NAME COUNT SAMP UNIT MIN MAX SUM SUMSQ; do - eval $NAME=$COUNT echo "got $COUNT $NAME" - + [ ! $MIN ] && error "Missing min value for $NAME proc entry" + eval $NAME=$COUNT || error "Wrong proc format" + case $NAME in - ost_read|ost_write) + read_bytes|write_bytes) [ $MIN -lt 4096 ] && error "min is too small: $MIN" [ $MIN -gt $FSIZE ] && error "min is too big: $MIN" [ $MAX -lt 4096 ] && error "max is too small: $MAX" [ $MAX -gt $FSIZE ] && error "max is too big: $MAX" [ $SUM -ne $FSIZE ] && error "sum is wrong: $SUM" - [ $SUMSQ -lt $(((FSIZE /4096) * (4096 * 4096))) ] && + [ $SUMSQ -lt $(((FSIZE /4096) * (4096 * 4096))) ] && error "sumsquare is too small: $SUMSQ" - [ $SUMSQ -gt $((FSIZE * FSIZE)) ] && + [ $SUMSQ -gt $((FSIZE * FSIZE)) ] && error "sumsquare is too big: $SUMSQ" ;; *) ;; @@ -5056,8 +5447,10 @@ test_127() { # bug 15521 done < $DIR/${tfile}.tmp #check that we actually got some stats - [ "$ost_read" ] || error "no read done" - [ "$ost_write" ] || error "no write done" + [ "$read_bytes" ] || error "Missing read_bytes stats" + [ "$write_bytes" ] || error "Missing write_bytes stats" + [ "$read_bytes" != 0 ] || error "no read done" + [ "$write_bytes" != 0 ] || error "no write done" } run_test 127 "verify the client stats are sane" @@ -5074,51 +5467,857 @@ test_128() { # bug 15212 } run_test 128 "interactive lfs for 2 consecutive find's" +set_dir_limits () { + local mntdev + local node + + local LDPROC=/proc/fs/ldiskfs + + for node in $(mdts_nodes); do + devs=$(do_node $node "lctl get_param -n devices" | awk '($3 ~ "mdt" && $4 ~ "MDT") { print $4 }') + for dev in $devs; do + mntdev=$(do_node $node "lctl get_param -n osd.$dev.mntdev") + do_node $node "echo $1 >$LDPROC/\\\$(basename $mntdev)/max_dir_size" + done + done +} test_129() { - [ "$FSTYPE" = "ldiskfs" ] || return 0 - mkdir $DIR/$tdir - - EFBIG=27 - ldiskfs_prefix=/proc/fs/ldiskfs - proc_file=max_dir_size - max_bytes=16385 - - echo $max_bytes >$ldiskfs_prefix/$proc_file - - I=0 - J=0 - while [ $I -lt $max_bytes ]; do - touch $DIR/$tdir/$J - J=$((J+1)) - I=$(stat -c%s "$DIR/$tdir") + [ "$FSTYPE" != "ldiskfs" ] && skip "not needed for FSTYPE=$FSTYPE" && return 0 + remote_mds_nodsh && skip "remote MDS with nodsh" && return + + EFBIG=27 + MAX=16384 + + set_dir_limits $MAX + + mkdir -p $DIR/$tdir + + I=0 + J=0 + while [ ! $I -gt $((MAX * MDSCOUNT)) ]; do + multiop $DIR/$tdir/$J Oc + rc=$? + if [ $rc -eq $EFBIG ]; then + set_dir_limits 0 + echo "return code $rc received as expected" + return 0 + elif [ $rc -ne 0 ]; then + set_dir_limits 0 + error_exit "return code $rc received instead of expected $EFBIG" + fi + J=$((J+1)) + I=$(stat -c%s "$DIR/$tdir") + done + + error "exceeded dir size limit $MAX x $MDSCOUNT $((MAX * MDSCOUNT)) : $I bytes" + do_facet $SINGLEMDS "echo 0 >$LDPROC" +} +run_test 129 "test directory size limit ========================" + +OLDIFS="$IFS" +cleanup_130() { + trap 0 + IFS="$OLDIFS" +} + +test_130a() { + filefrag_op=$(filefrag -e 2>&1 | grep "invalid option") + [ -n "$filefrag_op" ] && skip "filefrag does not support FIEMAP" && return + + trap cleanup_130 EXIT RETURN + + local fm_file=$DIR/$tfile + lfs setstripe -s 65536 -c 1 $fm_file || error "setstripe failed on $fm_file" + dd if=/dev/zero of=$fm_file bs=65536 count=1 || error "dd failed for $fm_file" + + filefrag -ves $fm_file || error "filefrag $fm_file failed" + filefrag_op=`filefrag -ve $fm_file | grep -A 100 "ext:" | grep -v "ext:" | grep -v "found"` + + lun=`$GETSTRIPE $fm_file | grep -A 10 obdidx | awk '{print $1}' | grep -v "obdidx"` + + start_blk=`echo $filefrag_op | cut -d: -f2 | cut -d. -f1` + IFS=$'\n' + tot_len=0 + for line in $filefrag_op + do + frag_lun=`echo $line | cut -d: -f5` + ext_len=`echo $line | cut -d: -f4` + if (( $frag_lun != $lun )); then + cleanup_130 + error "FIEMAP on 1-stripe file($fm_file) failed" + return + fi + (( tot_len += ext_len )) done - # One more file and we should be over the limit - multiop $DIR/$tdir/$J Oc - rc=$? - if [ $rc -eq 0 ]; then - rm -rf $DIR/$tdir - error "exceeded dir size limit: $I bytes" - elif [ $rc -ne $EFBIG ]; then - rm -rf $DIR/$tdir - error "return code $rc received instead of expected $EFBIG" - else - echo "return code $rc received as expected" + if (( lun != frag_lun || start_blk != 0 || tot_len != 64 )); then + cleanup_130 + error "FIEMAP on 1-stripe file($fm_file) failed;" + return fi - echo 0 >$ldiskfs_prefix/$proc_file - rm -rf $DIR/$tdir + cleanup_130 + + echo "FIEMAP on single striped file succeeded" } -run_test 129 "test directory size limit ========================" +run_test 130a "FIEMAP (1-stripe file)" + +test_130b() { + [ "$OSTCOUNT" -lt "2" ] && skip "skipping FIEMAP on 2-stripe file test" && return + + filefrag_op=$(filefrag -e 2>&1 | grep "invalid option") + [ -n "$filefrag_op" ] && skip "filefrag does not support FIEMAP" && return + + trap cleanup_130 EXIT RETURN + + local fm_file=$DIR/$tfile + lfs setstripe -s 65536 -c 2 $fm_file || error "setstripe failed on $fm_file" + dd if=/dev/zero of=$fm_file bs=1M count=2 || error "dd failed on $fm_file" + + filefrag -ves $fm_file || error "filefrag $fm_file failed" + filefrag_op=`filefrag -ve $fm_file | grep -A 100 "ext:" | grep -v "ext:" | grep -v "found"` + + last_lun=`echo $filefrag_op | cut -d: -f5` + + IFS=$'\n' + tot_len=0 + num_luns=1 + for line in $filefrag_op + do + frag_lun=`echo $line | cut -d: -f5` + ext_len=`echo $line | cut -d: -f4` + if (( $frag_lun != $last_lun )); then + if (( tot_len != 1024 )); then + cleanup_130 + error "FIEMAP on $fm_file failed; returned len $tot_len for OST $last_lun instead of 256" + return + else + (( num_luns += 1 )) + tot_len=0 + fi + fi + (( tot_len += ext_len )) + last_lun=$frag_lun + done + if (( num_luns != 2 || tot_len != 1024 )); then + cleanup_130 + error "FIEMAP on $fm_file failed; returned wrong number of luns or wrong len for OST $last_lun" + return + fi + + cleanup_130 + + echo "FIEMAP on 2-stripe file succeeded" +} +run_test 130b "FIEMAP (2-stripe file)" + +test_130c() { + [ "$OSTCOUNT" -lt "2" ] && skip "skipping FIEMAP on 2-stripe file with hole test" && return + + filefrag_op=$(filefrag -e 2>&1 | grep "invalid option") + [ -n "$filefrag_op" ] && skip "filefrag does not support FIEMAP" && return + + trap cleanup_130 EXIT RETURN + + local fm_file=$DIR/$tfile + lfs setstripe -s 65536 -c 2 $fm_file || error "setstripe failed on $fm_file" + dd if=/dev/zero of=$fm_file seek=1 bs=1M count=1 || error "dd failed on $fm_file" + + filefrag -ves $fm_file || error "filefrag $fm_file failed" + filefrag_op=`filefrag -ve $fm_file | grep -A 100 "ext:" | grep -v "ext:" | grep -v "found"` + + last_lun=`echo $filefrag_op | cut -d: -f5` + + IFS=$'\n' + tot_len=0 + num_luns=1 + for line in $filefrag_op + do + frag_lun=`echo $line | cut -d: -f5` + ext_len=`echo $line | cut -d: -f4` + if (( $frag_lun != $last_lun )); then + logical=`echo $line | cut -d: -f2 | cut -d. -f1` + if (( logical != 512 )); then + cleanup_130 + error "FIEMAP on $fm_file failed; returned logical start for lun $logical instead of 512" + return + fi + if (( tot_len != 512 )); then + cleanup_130 + error "FIEMAP on $fm_file failed; returned len $tot_len for OST $last_lun instead of 1024" + return + else + (( num_luns += 1 )) + tot_len=0 + fi + fi + (( tot_len += ext_len )) + last_lun=$frag_lun + done + if (( num_luns != 2 || tot_len != 512 )); then + cleanup_130 + error "FIEMAP on $fm_file failed; returned wrong number of luns or wrong len for OST $last_lun" + return + fi + + cleanup_130 + + echo "FIEMAP on 2-stripe file with hole succeeded" +} +run_test 130c "FIEMAP (2-stripe file with hole)" + +test_130d() { + [ "$OSTCOUNT" -lt "3" ] && skip "skipping FIEMAP on N-stripe file test" && return + + filefrag_op=$(filefrag -e 2>&1 | grep "invalid option") + [ -n "$filefrag_op" ] && skip "filefrag does not support FIEMAP" && return + + trap cleanup_130 EXIT RETURN + + local fm_file=$DIR/$tfile + lfs setstripe -s 65536 -c $OSTCOUNT $fm_file || error "setstripe failed on $fm_file" + dd if=/dev/zero of=$fm_file bs=1M count=$OSTCOUNT || error "dd failed on $fm_file" + + filefrag -ves $fm_file || error "filefrag $fm_file failed" + filefrag_op=`filefrag -ve $fm_file | grep -A 100 "ext:" | grep -v "ext:" | grep -v "found"` + + last_lun=`echo $filefrag_op | cut -d: -f5` + + IFS=$'\n' + tot_len=0 + num_luns=1 + for line in $filefrag_op + do + frag_lun=`echo $line | cut -d: -f5` + ext_len=`echo $line | cut -d: -f4` + if (( $frag_lun != $last_lun )); then + if (( tot_len != 1024 )); then + cleanup_130 + error "FIEMAP on $fm_file failed; returned len $tot_len for OST $last_lun instead of 1024" + return + else + (( num_luns += 1 )) + tot_len=0 + fi + fi + (( tot_len += ext_len )) + last_lun=$frag_lun + done + if (( num_luns != OSTCOUNT || tot_len != 1024 )); then + cleanup_130 + error "FIEMAP on $fm_file failed; returned wrong number of luns or wrong len for OST $last_lun" + return + fi + + cleanup_130 + + echo "FIEMAP on N-stripe file succeeded" +} +run_test 130d "FIEMAP (N-stripe file)" + +test_130e() { + [ "$OSTCOUNT" -lt "2" ] && skip "skipping continuation FIEMAP test" && return + + filefrag_op=$(filefrag -e 2>&1 | grep "invalid option") + [ -n "$filefrag_op" ] && skip "filefrag does not support FIEMAP" && return + + trap cleanup_130 EXIT RETURN + + local fm_file=$DIR/$tfile + lfs setstripe -s 65536 -c 2 $fm_file || error "setstripe failed on $fm_file" + NUM_BLKS=512 + EXPECTED_LEN=$(( (NUM_BLKS / 2) * 4 )) + for ((i = 0; i < $NUM_BLKS; i++)) + do + dd if=/dev/zero of=$fm_file count=1 bs=4096 seek=$((2*$i)) conv=notrunc > /dev/null 2>&1 + done + + filefrag -ves $fm_file || error "filefrag $fm_file failed" + filefrag_op=`filefrag -ve $fm_file | grep -A 750 "ext:" | grep -v "ext:" | grep -v "found"` + + last_lun=`echo $filefrag_op | cut -d: -f5` + + IFS=$'\n' + tot_len=0 + num_luns=1 + for line in $filefrag_op + do + frag_lun=`echo $line | cut -d: -f5` + ext_len=`echo $line | cut -d: -f4` + if (( $frag_lun != $last_lun )); then + if (( tot_len != $EXPECTED_LEN )); then + cleanup_130 + error "FIEMAP on $fm_file failed; returned len $tot_len for OST $last_lun instead of $EXPECTED_LEN" + return + else + (( num_luns += 1 )) + tot_len=0 + fi + fi + (( tot_len += ext_len )) + last_lun=$frag_lun + done + if (( num_luns != 2 || tot_len != $EXPECTED_LEN )); then + cleanup_130 + error "FIEMAP on $fm_file failed; returned wrong number of luns or wrong len for OST $last_lun" + return + fi + + cleanup_130 + + echo "FIEMAP with continuation calls succeeded" +} +run_test 130e "FIEMAP (test continuation FIEMAP calls)" + +# Test for writev/readv +test_131a() { + rwv -f $DIR/$tfile -w -n 3 524288 1048576 1572864 || \ + error "writev test failed" + rwv -f $DIR/$tfile -r -v -n 2 1572864 1048576 || \ + error "readv failed" + rm -f $DIR/$tfile +} +run_test 131a "test iov's crossing stripe boundary for writev/readv" + +test_131b() { + rwv -f $DIR/$tfile -w -a -n 3 524288 1048576 1572864 || \ + error "append writev test failed" + rwv -f $DIR/$tfile -w -a -n 2 1572864 1048576 || \ + error "append writev test failed" + rm -f $DIR/$tfile +} +run_test 131b "test append writev" + +test_131c() { + rwv -f $DIR/$tfile -w -d -n 1 1048576 || return 0 + error "NOT PASS" +} +run_test 131c "test read/write on file w/o objects" + +test_131d() { + rwv -f $DIR/$tfile -w -n 1 1572864 + NOB=`rwv -f $DIR/$tfile -r -n 3 524288 524288 1048576 | awk '/error/ {print $6}'` + if [ "$NOB" != 1572864 ]; then + error "Short read filed: read $NOB bytes instead of 1572864" + fi + rm -f $DIR/$tfile +} +run_test 131d "test short read" + +test_131e() { + rwv -f $DIR/$tfile -w -s 1048576 -n 1 1048576 + rwv -f $DIR/$tfile -r -z -s 0 -n 1 524288 || \ + error "read hitting hole failed" + rm -f $DIR/$tfile +} +run_test 131e "test read hitting hole" + +test_140() { #bug-17379 + mkdir -p $DIR/$tdir || error "Creating dir $DIR/$tdir" + cd $DIR/$tdir || error "Changing to $DIR/$tdir" + cp /usr/bin/stat . || error "Copying stat to $DIR/$tdir" + + # VFS limits max symlink depth to 5(4KSTACK) or 8 + local i=0 + while i=`expr $i + 1`; do + mkdir -p $i || error "Creating dir $i" + cd $i || error "Changing to $i" + ln -s ../stat stat || error "Creating stat symlink" + # Read the symlink until ELOOP present, + # not LBUGing the system is considered success, + # we didn't overrun the stack. + $OPENFILE -f O_RDONLY stat >/dev/null 2>&1; ret=$? + [ $ret -ne 0 ] && { + if [ $ret -eq 40 ]; then + break # -ELOOP + else + error "Open stat symlink" + return + fi + } + done + i=`expr $i - 1` + echo "The symlink depth = $i" + [ $i -eq 4 -o $i -eq 8 ] || error "Invalid symlink depth" +} +run_test 140 "Check reasonable stack depth (shouldn't LBUG) ====" + +test_150() { + local TF="$TMP/$tfile" + + dd if=/dev/urandom of=$TF bs=6096 count=1 || error "dd failed" + cp $TF $DIR/$tfile + cancel_lru_locks osc + cmp $TF $DIR/$tfile || error "$TMP/$tfile $DIR/$tfile differ" + remount_client $MOUNT + cmp $TF $DIR/$tfile || error "$TF $DIR/$tfile differ (remount)" + + $TRUNCATE $TF 6000 + $TRUNCATE $DIR/$tfile 6000 + cancel_lru_locks osc + cmp $TF $DIR/$tfile || error "$TF $DIR/$tfile differ (truncate1)" + + echo "12345" >>$TF + echo "12345" >>$DIR/$tfile + cancel_lru_locks osc + cmp $TF $DIR/$tfile || error "$TF $DIR/$tfile differ (append1)" + + echo "12345" >>$TF + echo "12345" >>$DIR/$tfile + cancel_lru_locks osc + cmp $TF $DIR/$tfile || error "$TF $DIR/$tfile differ (append2)" + + rm -f $TF + true +} +run_test 150 "truncate/append tests" + +function roc_access() { + ACCNUM=`$LCTL get_param -n obdfilter.*.stats | \ + grep 'cache_access'| awk '{print $2}' | \ + awk '{sum=sum+$3} END{print sum}'` + echo $ACCNUM +} + +function roc_hit() { + ACCNUM=`$LCTL get_param -n obdfilter.*.stats | \ + grep 'cache_hit'|awk '{print $2}' | \ + awk '{sum=sum+$1} END{print sum}'` + echo $ACCNUM +} + +test_151() { + local CPAGES=3 + + # check whether obdfilter is cache capable at all + if ! $LCTL get_param -n obdfilter.*.read_cache_enable; then + echo "not cache-capable obdfilter" + return 0 + fi + + # check cache is enabled on all obdfilters + if $LCTL get_param -n obdfilter.*.read_cache_enable | grep 0 >&/dev/null; then + echo "oss cache is disabled" + return 0 + fi + + $LCTL set_param -n obdfilter.*.writethrough_cache_enable 1 + + # pages should be in the case right after write + dd if=/dev/urandom of=$DIR/$tfile bs=4k count=$CPAGES || error "dd failed" + local BEFORE=`roc_hit` + cancel_lru_locks osc + cat $DIR/$tfile >/dev/null + local AFTER=`roc_hit` + if ! let "AFTER - BEFORE == CPAGES"; then + error "NOT IN CACHE: before: $BEFORE, after: $AFTER" + fi + + # the following read invalidates the cache + cancel_lru_locks osc + $LCTL set_param -n obdfilter.*.read_cache_enable 0 + cat $DIR/$tfile >/dev/null + + # now data shouldn't be found in the cache + BEFORE=`roc_hit` + cancel_lru_locks osc + cat $DIR/$tfile >/dev/null + AFTER=`roc_hit` + if let "AFTER - BEFORE != 0"; then + error "IN CACHE: before: $BEFORE, after: $AFTER" + fi + + $LCTL set_param -n obdfilter.*.read_cache_enable 1 + rm -f $DIR/$tfile +} +run_test 151 "test cache on oss and controls ===============================" + +test_152() { + local TF="$TMP/$tfile" + + # simulate ENOMEM during write +#define OBD_FAIL_OST_NOMEM 0x226 + lctl set_param fail_loc=0x80000226 + dd if=/dev/urandom of=$TF bs=6096 count=1 || error "dd failed" + cp $TF $DIR/$tfile + sync || error "sync failed" + lctl set_param fail_loc=0 + + # discard client's cache + cancel_lru_locks osc + + # simulate ENOMEM during read + lctl set_param fail_loc=0x80000226 + cmp $TF $DIR/$tfile || error "cmp failed" + lctl set_param fail_loc=0 -TMPDIR=$OLDTMPDIR -TMP=$OLDTMP -HOME=$OLDHOME + rm -f $TF +} +run_test 152 "test read/write with enomem ============================" + +test_153() { + multiop $DIR/$tfile Ow4096Ycu || error "multiop failed" +} +run_test 153 "test if fdatasync does not crash =======================" + +err17935 () { + if [ $MDSCOUNT -gt 1 ]; then + error_ignore 17935 $* + else + error $* + fi +} + +test_154() { + cp /etc/hosts $DIR/$tfile + + fid=`$LFS path2fid $DIR/$tfile` + rc=$? + [ $rc -ne 0 ] && error "error: could not get fid for $DIR/$tfile." + + diff $DIR/$tfile $DIR/.lustre/fid/$fid || error "open by fid failed: did not find expected data in file." + + echo "Opening a file by FID succeeded" +} +run_test 154 "Opening a file by FID" + +#Changelogs +test_160() { + remote_mds && skip "remote MDS" && return + lctl set_param -n mdd.*.changelog on + $LFS changelog_clear $FSNAME 0 + + # change something + mkdir -p $DIR/$tdir/pics/2008/zachy + touch $DIR/$tdir/pics/2008/zachy/timestamp + cp /etc/hosts $DIR/$tdir/pics/2008/zachy/pic1.jpg + mv $DIR/$tdir/pics/2008/zachy $DIR/$tdir/pics/zach + ln $DIR/$tdir/pics/zach/pic1.jpg $DIR/$tdir/pics/2008/portland.jpg + ln -s $DIR/$tdir/pics/2008/portland.jpg $DIR/$tdir/pics/desktop.jpg + rm $DIR/$tdir/pics/desktop.jpg + + # verify contents + $LFS changelog $FSNAME + # check target fid + fidc=$($LFS changelog $FSNAME | grep timestamp | grep "CREAT" | tail -1 | \ + awk '{print $5}') + fidf=$($LFS path2fid $DIR/$tdir/pics/zach/timestamp) + [ "$fidc" == "t=$fidf" ] || \ + err17935 "fid in changelog $fidc != file fid $fidf" + # check parent fid + fidc=$($LFS changelog $FSNAME | grep timestamp | grep "CREAT" | tail -1 | \ + awk '{print $6}') + fidf=$($LFS path2fid $DIR/$tdir/pics/zach) + [ "$fidc" == "p=$fidf" ] || \ + err17935 "pfid in changelog $fidc != dir fid $fidf" + + # verify purge + FIRST_REC=$($LFS changelog $FSNAME | head -1 | awk '{print $1}') + $LFS changelog_clear $FSNAME $(($FIRST_REC + 5)) + PURGE_REC=$($LFS changelog $FSNAME | head -1 | awk '{print $1}') + [ $PURGE_REC == $(($FIRST_REC + 6)) ] || \ + err17935 "first rec after purge should be $(($FIRST_REC + 6)); is $PURGE_REC" + # purge all + $LFS changelog_clear $FSNAME 0 + lctl set_param -n mdd.*.changelog off +} +run_test 160 "changelog sanity" + +test_161() { + # need local MDT for fid2path + remote_mds && skip "remote MDS" && return + + mkdir -p $DIR/$tdir + cp /etc/hosts $DIR/$tdir/$tfile + mkdir $DIR/$tdir/foo1 + mkdir $DIR/$tdir/foo2 + ln $DIR/$tdir/$tfile $DIR/$tdir/foo1/sofia + ln $DIR/$tdir/$tfile $DIR/$tdir/foo2/zachary + ln $DIR/$tdir/$tfile $DIR/$tdir/foo1/luna + ln $DIR/$tdir/$tfile $DIR/$tdir/foo2/thor + local FID=$($LFS path2fid $DIR/$tdir/$tfile) + if [ "$($LFS fid2path ${mds1_svc} $FID | wc -l)" != "5" ]; then + $LFS fid2path ${mds1_svc} $FID + err17935 "bad link ea" + fi + # middle + rm $DIR/$tdir/foo2/zachary + # last + rm $DIR/$tdir/foo2/thor + # first + rm $DIR/$tdir/$tfile + # rename + mv $DIR/$tdir/foo1/sofia $DIR/$tdir/foo2/maggie + if [ "$($LFS fid2path ${mds1_svc} --link 1 $FID)" != "/$tdir/foo2/maggie" ] + then + $LFS fid2path ${mds1_svc} $FID + err17935 "bad link rename" + fi + rm $DIR/$tdir/foo2/maggie + + # overflow the EA + local longname=filename_avg_len_is_thirty_two_ + createmany -l$DIR/$tdir/foo1/luna $DIR/$tdir/foo2/$longname 1000 || \ + error "failed to hardlink many files" + links=$($LFS fid2path ${mds1_svc} $FID | wc -l) + echo -n "${links}/1000 links in link EA" + [ ${links} -gt 60 ] || err17935 "expected at least 60 links in link EA" + unlinkmany $DIR/$tdir/foo2/$longname 1000 || \ + error "failed to unlink many hardlinks" +} +run_test 161 "link ea sanity" + +check_path() { + local expected=$1 + shift + local fid=$2 + + local path=$(${LFS} fid2path $*) + RC=$? + + if [ $RC -ne 0 ]; then + err17935 "path looked up of $expected failed. Error $RC" + return $RC + elif [ "${path}" != "${expected}" ]; then + err17935 "path looked up \"${path}\" instead of \"${expected}\"" + return 2 + fi + echo "fid $fid resolves to path $path" +} + +test_162() { + # need local MDT for fid2path + remote_mds && skip "remote MDS" && return + + # Make changes to filesystem + mkdir -p $DIR/$tdir/d2 + touch $DIR/$tdir/d2/$tfile + touch $DIR/$tdir/d2/x1 + touch $DIR/$tdir/d2/x2 + mkdir -p $DIR/$tdir/d2/a/b/c + mkdir -p $DIR/$tdir/d2/p/q/r + fid=$($LFS path2fid $DIR/$tdir/d2/$tfile) + check_path "/$tdir/d2/$tfile" ${mds1_svc} $fid --link 0 + ln $DIR/$tdir/d2/$tfile $DIR/$tdir/d2/p/q/r/hlink + mv $DIR/$tdir/d2/$tfile $DIR/$tdir/d2/a/b/c/new_file + fid=$($LFS path2fid $DIR/$tdir/d2/a/b/c/new_file) + check_path "/$tdir/d2/a/b/c/new_file" ${mds1_svc} $fid --link 1 + check_path "/$tdir/d2/p/q/r/hlink" ${mds1_svc} $fid --link 0 + # check that there are 2 links, and that --rec doesnt break anything + ${LFS} fid2path ${mds1_svc} $fid --rec 20 | wc -l | grep -q 2 || \ + err17935 "expected 2 links" + + rm $DIR/$tdir/d2/p/q/r/hlink + check_path "/$tdir/d2/a/b/c/new_file" ${mds1_svc} $fid --link 0 + # Doesnt work with CMD yet: 17935 + return 0 +} +run_test 162 "path lookup sanity" + +test_170() { + $LCTL debug_daemon start $TMP/${tfile}_log_good + touch $DIR/$tfile + $LCTL debug_daemon stop + sed -e "s/^...../a/g" $TMP/${tfile}_log_good > $TMP/${tfile}_log_bad || + error "sed failed to read log_good" + + $LCTL debug_daemon start $TMP/${tfile}_log_good + rm -rf $DIR/$tfile + $LCTL debug_daemon stop + + $LCTL df $TMP/${tfile}_log_bad 2&> $TMP/${tfile}_log_bad.out || + error "lctl df log_bad failed" + + local bad_line=$(tail -n 1 $TMP/${tfile}_log_bad.out | awk '{print $9}') + local good_line1=$(tail -n 1 $TMP/${tfile}_log_bad.out | awk '{print $5}') + + $LCTL df $TMP/${tfile}_log_good 2&>$TMP/${tfile}_log_good.out + local good_line2=$(tail -n 1 $TMP/${tfile}_log_good.out | awk '{print $5}') + + [ "$bad_line" ] && [ "$good_line1" ] && [ "$good_line2" ] || + error "bad_line good_line1 good_line2 are empty" + + cat $TMP/${tfile}_log_good >> $TMP/${tfile}_logs_corrupt + cat $TMP/${tfile}_log_bad >> $TMP/${tfile}_logs_corrupt + cat $TMP/${tfile}_log_good >> $TMP/${tfile}_logs_corrupt + + $LCTL df $TMP/${tfile}_logs_corrupt 2&> $TMP/${tfile}_log_bad.out + local bad_line_new=$(tail -n 1 $TMP/${tfile}_log_bad.out | awk '{print $9}') + local good_line_new=$(tail -n 1 $TMP/${tfile}_log_bad.out | awk '{print $5}') + + [ "$bad_line_new" ] && [ "$good_line_new" ] || + error "bad_line_new good_line_new are empty" + + local expected_good=$((good_line1 + good_line2*2)) + + rm -rf $TMP/${tfile}* + if [ $bad_line -ne $bad_line_new ]; then + error "expected $bad_line bad lines, but got $bad_line_new" + return 1 + fi + + if [ $expected_good -ne $good_line_new ]; then + error "expected $expected_good good lines, but got $good_line_new" + return 2 + fi + true +} +run_test 170 "test lctl df to handle corrupted log =====================" + +# OST pools tests +POOL=${POOL:-cea1} +TGT_COUNT=$OSTCOUNT +TGTPOOL_FIRST=1 +TGTPOOL_MAX=$(($TGT_COUNT - 1)) +TGTPOOL_STEP=2 +TGTPOOL_LIST=`seq $TGTPOOL_FIRST $TGTPOOL_STEP $TGTPOOL_MAX` +POOL_ROOT=${POOL_ROOT:-$DIR/d200.pools} +POOL_DIR=$POOL_ROOT/dir_tst +POOL_FILE=$POOL_ROOT/file_tst + +check_file_in_pool() +{ + file=$1 + res=$($GETSTRIPE $file | grep 0x | cut -f2) + for i in $res + do + found=$(echo :$TGTPOOL_LIST: | tr " " ":" | grep :$i:) + if [[ "$found" == "" ]] + then + echo "pool list: $TGTPOOL_LIST" + echo "striping: $res" + error "$file not allocated in $POOL" + return 1 + fi + done + return 0 +} + +test_200a() { + remote_mgs_nodsh && skip "remote MGS with nodsh" && return + do_facet mgs $LCTL pool_new $FSNAME.$POOL + # get param should return err until pool is created + wait_update $HOSTNAME "lctl get_param -n lov.$FSNAME-*.pools.$POOL 2>/dev/null || echo foo" "" || error "Pool creation of $POOL failed" +} +run_test 200a "Create new pool ==========================================" + +test_200b() { + remote_mgs_nodsh && skip "remote MGS with nodsh" && return + TGT=$(for i in `seq $TGTPOOL_FIRST $TGTPOOL_STEP $TGTPOOL_MAX`; do printf "$FSNAME-OST%04x_UUID " $i; done) + do_facet mgs $LCTL pool_add $FSNAME.$POOL \ + $FSNAME-OST[$TGTPOOL_FIRST-$TGTPOOL_MAX/$TGTPOOL_STEP] + wait_update $HOSTNAME "lctl get_param -n lov.$FSNAME-*.pools.$POOL | sort -u | tr '\n' ' ' " "$TGT" || error "Add to pool failed" +} +run_test 200b "Add targets to a pool ====================================" + +test_200c() { + remote_mgs_nodsh && skip "remote MGS with nodsh" && return + mkdir -p $POOL_DIR + $SETSTRIPE -c 2 -p $POOL $POOL_DIR + [ $? = 0 ] || error "Cannot set pool $POOL to $POOL_DIR" +} +run_test 200c "Set pool on a directory =================================" + +test_200d() { + remote_mgs_nodsh && skip "remote MGS with nodsh" && return + res=$($GETSTRIPE $POOL_DIR | grep pool: | cut -f8 -d " ") + [ "$res" = $POOL ] || error "Pool on $POOL_DIR is not $POOL" +} +run_test 200d "Check pool on a directory ===============================" + +test_200e() { + remote_mgs_nodsh && skip "remote MGS with nodsh" && return + failed=0 + for i in $(seq -w 1 $(($TGT_COUNT * 3))) + do + file=$POOL_DIR/file-$i + touch $file + check_file_in_pool $file + if [[ $? != 0 ]] + then + failed=$(($failed + 1)) + fi + done + [ "$failed" = 0 ] || error "$failed files not allocated in $POOL" +} +run_test 200e "Check files allocation from directory pool ==============" + +test_200f() { + remote_mgs_nodsh && skip "remote MGS with nodsh" && return + mkdir -p $POOL_FILE + failed=0 + for i in $(seq -w 1 $(($TGT_COUNT * 3))) + do + file=$POOL_FILE/spoo-$i + $SETSTRIPE -p $POOL $file + check_file_in_pool $file + if [[ $? != 0 ]] + then + failed=$(($failed + 1)) + fi + done + [ "$failed" = 0 ] || error "$failed files not allocated in $POOL" +} +run_test 200f "Create files in a pool ===================================" + +test_200g() { + remote_mgs_nodsh && skip "remote MGS with nodsh" && return + TGT=$($LCTL get_param -n lov.$FSNAME-*.pools.$POOL | head -1) + do_facet mgs $LCTL pool_remove $FSNAME.$POOL $TGT + wait_update $HOSTNAME "lctl get_param -n lov.$FSNAME-*.pools.$POOL | grep $TGT" "" || error "$TGT not removed from $FSNAME.$POOL" +} +run_test 200g "Remove a target from a pool =============================" + +test_200h() { + remote_mgs_nodsh && skip "remote MGS with nodsh" && return + for TGT in $($LCTL get_param -n lov.$FSNAME-*.pools.$POOL | sort -u) + do + do_facet mgs $LCTL pool_remove $FSNAME.$POOL $TGT + done + wait_update $HOSTNAME "lctl get_param -n lov.$FSNAME-*.pools.$POOL" ""\ + || error "Pool $FSNAME.$POOL cannot be drained" +} +run_test 200h "Remove all targets from a pool ==========================" + +test_200i() { + remote_mgs_nodsh && skip "remote MGS with nodsh" && return + do_facet mgs $LCTL pool_destroy $FSNAME.$POOL + # get param should return err once pool is gone + wait_update $HOSTNAME "lctl get_param -n lov.$FSNAME-*.pools.$POOL 2>/dev/null || echo foo" "foo" && return 0 + error "Pool $FSNAME.$POOL is not destroyed" +} +run_test 200i "Remove a pool ============================================" + +test_212() { + size=`date +%s` + size=$((size % 8192 + 1)) + dd if=/dev/urandom of=$DIR/f212 bs=1k count=$size + sendfile $DIR/f212 $DIR/f212.xyz || error "sendfile wrong" + rm -f $DIR/f212 $DIR/f212.xyz +} +run_test 212 "Sendfile test ============================================" + +# +# tests that do cleanup/setup should be run at the end +# + +test_900() { + local ls + #define OBD_FAIL_MGC_PAUSE_PROCESS_LOG 0x903 + $LCTL set_param fail_loc=0x903 + # cancel_lru_locks mgc - does not work due to lctl set_param syntax + for ls in /proc/fs/lustre/ldlm/namespaces/MGC*/lru_size; do + echo "clear" > $ls + done + FAIL_ON_ERROR=true cleanup + FAIL_ON_ERROR=true setup +} +run_test 900 "umount should not race with any mgc requeue thread" log "cleanup: ======================================================" check_and_cleanup_lustre if [ "$I_MOUNTED" != "yes" ]; then - sysctl -w lnet.debug="$OLDDEBUG" 2> /dev/null || true + lctl set_param debug="$OLDDEBUG" 2> /dev/null || true fi echo '=========================== finished ==============================='