as fail_loc is shared among client and server on a local setup,
it can hit just before rm -f $TMP/$tfile*. given it's marked
with FAIL_ONCE, the log won't be dumped again.
to fix this do cleanup rm earlier.
Test-Parameters: trivial
Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Change-Id: Ifd70335c44bfe2e1970e1441837dcb3545c40ab0
Reviewed-on: https://review.whamcloud.com/43308
Tested-by: jenkins <devops@whamcloud.com>
Reviewed-by: Bobi Jam <bobijam@hotmail.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
stack_trap "$LCTL set_param debug_path=$old_path"
stack_trap "rm -f $TMP/$tfile*"
+ rm -f $TMP/$tfile* 2> /dev/null
#define OBD_FAIL_PTLRPC_DUMP_LOG 0x50e
$LCTL set_param debug_path=$TMP/$tfile fail_loc=0x8000050e
test_mkdir $DIR/$tdir
- rm -f $TMP/$tfile* 2> /dev/null
# retry in case the open is cached and not released
for (( i = 0; i < 100 && $(ls $TMP/$tfile* | wc -l) == 0; i++ )); do
echo $i > $DIR/$tdir/$tfile.$i && cat $DIR/$tdir/$tfile.$i