Whamcloud - gitweb
fs/lustre-release.git
13 years agob=23210 don't update obd->obd_osfs if target is gone already
Johann Lombardi [Wed, 30 Jun 2010 08:53:33 +0000 (10:53 +0200)]
b=23210 don't update obd->obd_osfs if target is gone already

i=wangdi
i=landen

lov_disconnect() can clean up lov->lov_tgts while the statfs interpret
routine of an rpc in flight has not been executed yet.

13 years agob=23131 issue OBD_NOTIFY_CREATE event in lov
Johann Lombardi [Wed, 30 Jun 2010 08:28:46 +0000 (10:28 +0200)]
b=23131 issue OBD_NOTIFY_CREATE event in lov

i=nathan
i=andrew

(mds_lov.c:573:mds_lov_update_desc()) Process entered
(obd_class.h:376:obd_get_info()) Process entered
(obd_class.h:378:obd_get_info()) obd_get_info: NULL export
(obd_class.h:378:obd_get_info()) Process leaving (rc=18446744073709551597 : -19
ffffffffffffffed)

The problem is that mds_lov_connect() calls obd_notify(OBD_NOTIFY_CREATE)
*before* obd_connect(...,&mds->mds_lov_exp). Since mds_notify(OBD_NOTIFY_CREATE)
requires mds->mds_lov_exp which is not yet initialized, it fails.

13 years agob=23192 fix race when ping evictor and srv thread executes target_recovery_check_and_...
Johann Lombardi [Tue, 29 Jun 2010 11:59:03 +0000 (13:59 +0200)]
b=23192 fix race when ping evictor and srv thread executes target_recovery_check_and_stop() simultaneously

i=tappro
i=andrew

target_recovery_expired() wakes up both the ping evictor and the service thread
waiting in process_recovery_queue() which can race in
target_recovery_check_and_stop().

13 years agob=23185 check for both arches
Brian J. Murrell [Mon, 28 Jun 2010 16:33:30 +0000 (12:33 -0400)]
b=23185 check for both arches

When we build our version of the SLES kernel, we optimize it for i686
whereas the SUSE kernel is i386.  The actual arch makes a difference in
where the Module.symvers can be found, so just look in both locations
to cover both the upstream vendor kernel as well as our patched kernel.

13 years agob=17485 fix replay-single test 86 to remount client at the end
ZhangHongChao [Mon, 28 Jun 2010 16:25:42 +0000 (18:25 +0200)]
b=17485 fix replay-single test 86 to remount client at the end

i=johann

13 years agob=23196 obd reference fixes in lov_quota_check
Andrew Perepechko [Sun, 27 Jun 2010 18:27:17 +0000 (22:27 +0400)]
b=23196 obd reference fixes in lov_quota_check

i=Johann Lombardi

13 years agob=23196 quota broadcast crashes with inactive OSC
Johann Lombardi [Sun, 27 Jun 2010 05:48:51 +0000 (07:48 +0200)]
b=23196 quota broadcast crashes with inactive OSC

lov_quota_adjust_qunit() must take a lov reference
and check for lov->lov_tgts[i] != NULL when parsing
lov->lov_tgts.

13 years agob=23185 use resolve_arch() in the test too
Brian J. Murrell [Sat, 26 Jun 2010 02:57:23 +0000 (22:57 -0400)]
b=23185 use resolve_arch() in the test too

Missed a usage need for resolve_arch() in the test for the file we
extracted, using resolve_arch().

13 years agobump version to 1.8.3.56 v1_8_3_56
Johann Lombardi [Fri, 25 Jun 2010 18:18:21 +0000 (20:18 +0200)]
bump version to 1.8.3.56

13 years agob=22632 update mptsas driver to PH16-4.18.20.04
Brian J. Murrell [Fri, 25 Jun 2010 17:27:12 +0000 (13:27 -0400)]
b=22632 update mptsas driver to PH16-4.18.20.04

i=wangyb
i=yangsheng

13 years agob=17485 add regression test to verify that MDS does not allocate some objid twice
Johann Lombardi [Fri, 25 Jun 2010 16:36:54 +0000 (18:36 +0200)]
b=17485 add regression test to verify that MDS does not allocate some objid twice

o=hongchao
i=johann
i=grev

13 years agob=17485 Don't reset the lov_objid to last_id sent by OST
Johann Lombardi [Fri, 25 Jun 2010 14:06:07 +0000 (16:06 +0200)]
b=17485 Don't reset the lov_objid to last_id sent by OST

i=andrew
i=hongchao

Resetting the lov_objid values to last_id reported by
the OST in mds_lov_get_objid() unconditionally is just wrong and can cause
the same object to be allocated twice by the MDS.

The SKIP_ORPHAN feature is supposed to work as follows:
* during orphan recovery, the MDS must supply to the OST the lov_objid
  value;
* on the OST side:
  - if MDS's lov_objid > OST's last_id, the OST recreates the missing objects
    up to lov_objid and packs this value in the reply;
  - if OST's last_id > MDS's lov_objid, the OST destroys orphan objects from
    lov_objid + 1 up to last_id and does *not* change last_id. The OST packs
    in the reply the value of last_id which is from where the MDS should
    restart allocations if it does not want to reuse orphans.
* the MDS sets next_id to the value reported by the OST in the orphan recovery
  request.

If we always discard the lov_objid values as done currently in mds_lov_get_objid(),
orphan recovery is just a noop and neither clears orphans nor recreates missing
objects.

13 years agob=23185 properly support sles10 ppc64
Brian J. Murrell [Fri, 25 Jun 2010 13:49:13 +0000 (09:49 -0400)]
b=23185 properly support sles10 ppc64

We never really have properly supported ppc64 on sles10:
- RPMSMPTYPE for ppc64 on SLES10 is ppc64, not smp
- add a function to resolve an arch to their canonical value

Comment out some code to align the sles distro lbuild with rhel5.

13 years agob=21452 support for weak-modules
Brian J. Murrell [Fri, 25 Jun 2010 13:49:12 +0000 (09:49 -0400)]
b=21452 support for weak-modules

Add support to our RPM SPEC for the weak-modules script.

This requires that we install our modules under
/lib/modules/$(uname -r)/updates/kernel.  I think this is the correct
location for us in any case given that we are a kernel "addon" package.
Relax the kernel Requires: to work better with weak-modules.
Use the external dependency generator as the internal one cannot deal
with kernel modules.

i=mjmac
i=wangyb

13 years agob=21452 kABI tracking
Brian J. Murrell [Fri, 25 Jun 2010 13:49:11 +0000 (09:49 -0400)]
b=21452 kABI tracking

Changes required to build packages that properly interact with
the kABI tracking included in SLES and RHEL kernels:
- take local copies of the find_requires generators and make modif-
  ications to them so that they can find the kernel artifacts installed
  in our build tree
  + thusly, override the __find_requires macro in all of the RPM
    build operations

Fix package versioning of patchless packages to also include the
timestamp, just like patched packages do, for non-release builds.
Also, remove the _lustre from the kernel version in the lustre
package names for patchless packages.

Note that changes needed to enable the external dependency checker
and thus include kABI dependencies in the RPM are not included here
for RHEL5.  Once RH have updated their kABI whitelist to include the
symbols we need, the patches in bz attachments 30546-7 need to land
for RHEL5 weak modules to work and a bug needs to be filed with OFA
to land the change in bz attachment 30547.

i=wangyb
i=mjmac

13 years agob=23182 lustre-iokit sgpdd-survey defect
Elena Gryaznova [Fri, 25 Jun 2010 12:38:17 +0000 (16:38 +0400)]
b=23182 lustre-iokit sgpdd-survey defect

i=Vladimir.Saveliev

13 years agob=22464 add writeconf as mount option
Nathan Rutman [Thu, 24 Jun 2010 17:08:59 +0000 (10:08 -0700)]
b=22464 add writeconf as mount option

i=alexg
i=adilger

13 years agofix warning with ll_bdi_num
Johann Lombardi [Thu, 24 Jun 2010 16:54:12 +0000 (18:54 +0200)]
fix warning with ll_bdi_num

13 years agobump version to 1.8.3.55 v1_8_3_55
Johann Lombardi [Thu, 24 Jun 2010 16:02:56 +0000 (18:02 +0200)]
bump version to 1.8.3.55

13 years agob=23014 sgpdd-survey acc-sm integration
Elena Gryaznova [Thu, 24 Jun 2010 15:45:15 +0000 (19:45 +0400)]
b=23014 sgpdd-survey acc-sm integration

i=Andreas.Dilger

13 years agob=20563 libcfs_cfs_curproc_is_32bit() function
pravin [Thu, 24 Jun 2010 21:13:16 +0000 (02:43 +0530)]
b=20563 libcfs_cfs_curproc_is_32bit() function

i=johann
i=rahul

adds function cfs_curproc_is_32bit() to check compatibility.

13 years agob=21610 include fs.h for posix_acl.
yangsheng [Thu, 24 Jun 2010 13:36:34 +0000 (21:36 +0800)]
b=21610 include fs.h for posix_acl.

13 years agob=23118 Change for new backing_dev_info interface.
yangsheng [Thu, 24 Jun 2010 13:34:12 +0000 (21:34 +0800)]
b=23118 Change for new backing_dev_info interface.

i=kalpak
i=johann

13 years agob=22514 using inkernel OFED stack for rhel5 & oel5.
yangsheng [Thu, 24 Jun 2010 13:12:57 +0000 (21:12 +0800)]
b=22514 using inkernel OFED stack for rhel5 & oel5.

i=brian
i=johann

13 years agob=20563 fix LUSTRE_SEQ_MAX_WIDTH
pravin [Thu, 24 Jun 2010 09:53:23 +0000 (15:23 +0530)]
b=20563 fix LUSTRE_SEQ_MAX_WIDTH

i=rahul
i=johann

fix LUSTRE_SEQ_MAX_WIDTH for 1.8 interop client .

13 years agob=11742 OBD_FL_MMAP should only be used in conjunction with OBD_MD_FLFLAGS
Johann Lombardi [Thu, 24 Jun 2010 08:43:00 +0000 (10:43 +0200)]
b=11742 OBD_FL_MMAP should only be used in conjunction with OBD_MD_FLFLAGS

i=adilger
i=dmitry

The patch from bug 11742 uses oa.o_flags without setting OBD_MD_FLFLAGS
in oa.o_valid. This is wrong since anybody can legitimately reset o_flags
if OBD_MD_FLFLAGS is not set (e.g. the checksum code does this).

13 years agob=22514 remove jbd-slab-race-2.6-rhel5
Johann Lombardi [Wed, 23 Jun 2010 15:10:51 +0000 (17:10 +0200)]
b=22514 remove jbd-slab-race-2.6-rhel5

The problem has been fixed upstream, so we no longer need this patch.

13 years agob=23076 fix for o2iblnd reconnect to retry one more time
Maxim Patlasov [Wed, 23 Jun 2010 15:01:19 +0000 (17:01 +0200)]
b=23076 fix for o2iblnd reconnect to retry one more time

i=isaac

With peer health detection, o2iblnd makes only one attempt to reconnect
which is not enough with nodes running lustre 1.6 because of  proto version
mismatch.

13 years agob=22755 Don't consume grant twice on recoverable resend
Johann Lombardi [Wed, 23 Jun 2010 09:38:38 +0000 (11:38 +0200)]
b=22755 Don't consume grant twice on recoverable resend

i=green
i=wangdi

Original patch from Oleg.

13 years agob=22797 notify users that recovery-[random,double]-scale tests are not functional...
Elena Gryaznova [Tue, 22 Jun 2010 21:11:36 +0000 (01:11 +0400)]
b=22797 notify users that recovery-[random,double]-scale tests are not functional for SOFT failure mode

i=Brian.Murrell

13 years agob=22755 more grant fixes
Johann Lombardi [Wed, 23 Jun 2010 09:24:43 +0000 (11:24 +0200)]
b=22755 more grant fixes

i=oleg
i=wangdi

- filter_preprw_write() should check for OBD_FL_SHRINK_GRANT in o_flags instead of
  o_valid;
- The grant shrink code should not use o_flags w/o setting OBD_MD_FLFLAGS since
  o_flags can be legitimately reset if OBD_MD_FLFLAGS (e.g. it is done in the
  checksum code);
- add OBD_FL_MMAP to wiretest & wirecheck

13 years agob=14929 a small cleanup in test-framework.sh
Andrew Perepechko [Tue, 22 Jun 2010 14:45:51 +0000 (18:45 +0400)]
b=14929 a small cleanup in test-framework.sh

i=Elena Gryaznova

13 years agob=23133 interop 18<->20: sanity 57a,57b,129: error: get_param: /proc/{fs,sys}/{lnet...
Rahul Deshmukh [Tue, 22 Jun 2010 12:07:26 +0000 (17:37 +0530)]
b=23133 interop 18<->20: sanity 57a,57b,129: error: get_param: /proc/{fs,sys}/{lnet,lustre}/osd/*/mntdev: Found no match

i=adilger

Fixed interop tests (sanity 57a,b 129) after renaming osd to osd-ldiskfs in master.

13 years agob=22040 use df POSIX output format
Elena Gryaznova [Mon, 21 Jun 2010 11:04:38 +0000 (15:04 +0400)]
b=22040 use df POSIX output format

i=Oleg.Drokin

13 years agoRevert "b=21951 Init scatterlist table."
Johann Lombardi [Mon, 21 Jun 2010 22:25:02 +0000 (00:25 +0200)]
Revert "b=21951 Init scatterlist table."

This reverts commit f1480c8eec3618567288e75b5ce756b519ca9447.
This is suspected to cause bug 23123.

13 years agob=14929 a tiny fix for mkfs build
Andrew Perepechko [Mon, 21 Jun 2010 10:23:41 +0000 (14:23 +0400)]
b=14929 a tiny fix for mkfs build

i=Johann Lombardi
i=ZhiYong Tian

13 years agob=22309 get and validate lock before issuing readahead
hongchao.zhang [Sun, 13 Jun 2010 05:42:11 +0000 (13:42 +0800)]
b=22309 get and validate lock before issuing readahead

in ll_read_ahead_page, get and check the validation of the ldlm lock
to protect the readahead before issuing the actual read request

i=johann@sun.com
i=yong.fan@sun.com

13 years agob=22906 fix invalid block count with 16TB block device
Zhiqi Tao [Fri, 18 Jun 2010 21:09:09 +0000 (23:09 +0200)]
b=22906 fix invalid block count with 16TB block device

13 years agobump version to 1.8.3.54 v1_8_3_54
Johann Lombardi [Fri, 18 Jun 2010 21:02:43 +0000 (23:02 +0200)]
bump version to 1.8.3.54

13 years agob=22223 bump max # phys/hw segment in SLES11 until s/g chaining works properly
Johann Lombardi [Fri, 18 Jun 2010 20:57:57 +0000 (22:57 +0200)]
b=22223 bump max # phys/hw segment in SLES11 until s/g chaining works properly

13 years agob=22656 Prevent failover nids from registering with MGS first.
Nathan Rutman [Thu, 17 Jun 2010 18:47:58 +0000 (11:47 -0700)]
b=22656 Prevent failover nids from registering with MGS first.

Make the check in mgs_handle_target_reg()

o=Joseph Herring
i=nathan.rutman
i=andreas.dilger

13 years agob=23094 add layout lock connect flag
Nathan Rutman [Thu, 17 Jun 2010 18:31:01 +0000 (11:31 -0700)]
b=23094 add layout lock connect flag

o=jc.lafoucriere@cea.fr
i=nathan
i=rread

13 years agob=22514 fork ldiskfs patches between sles11 & rhel5
yangsheng [Fri, 18 Jun 2010 20:51:53 +0000 (22:51 +0200)]
b=22514 fork ldiskfs patches between sles11 & rhel5
i=johann

Landing for RHEL 5.5 support has broken build for SLES11.

13 years agoMerge branch 'b1_8' of git@git.lustre.org:prime/lustre into b1_8
Johann Lombardi [Thu, 17 Jun 2010 08:49:57 +0000 (10:49 +0200)]
Merge branch 'b1_8' of git@git.lustre.org:prime/lustre into b1_8

13 years agob=21835 fix mdsrate script.
Eric Mei [Thu, 17 Jun 2010 02:14:13 +0000 (20:14 -0600)]
b=21835 fix mdsrate script.

r=robert.read

13 years agob=22514 Update RHEL5.5 kernel to 2.6.18-194.3.1.
yangsheng [Wed, 16 Jun 2010 15:24:25 +0000 (23:24 +0800)]
b=22514 Update RHEL5.5 kernel to 2.6.18-194.3.1.

i=johann
i=girish,rahul
i=kalpak,zhanghongchao
i=brian,wangyb

13 years agob=22853 client should flush current sequence to make client obtain new one from serve...
Fan Yong [Wed, 16 Jun 2010 14:01:15 +0000 (22:01 +0800)]
b=22853 client should flush current sequence to make client obtain new one from server in case of disconnect/reconnect

Client should flush current sequence to make client obtain new one from server in case of disconnect/reconnect.

i=robert.read
i=johann

13 years agob=22846 debuginfo packages for lustre and ldiskfs
Brian J. Murrell [Tue, 15 Jun 2010 13:32:17 +0000 (09:32 -0400)]
b=22846 debuginfo packages for lustre and ldiskfs

With our current .spec files for lustre and ldiskfs, debuginfo packages
are not produced.  It seems that while other distros do this implicitly,
SLES needs to be told to do so explicitly.

i=mjmac
i=wangyb

13 years agob=22938 lfs find -s doesn't seem to work correctly
Dmitry Zogin [Tue, 15 Jun 2010 12:50:18 +0000 (08:50 -0400)]
b=22938 lfs find -s doesn't seem to work correctly

 Fixing find_value_cmp().

 o=yuriy.umanets
 i=johann
 i=andreas.dilger

13 years agob=22884 File to OST allocation on an empty file system varies greatly.
Dmitry Zogin [Tue, 15 Jun 2010 12:49:18 +0000 (08:49 -0400)]
b=22884 File to OST allocation on an empty file system varies greatly.

 osc_precreate() should return 0, if there are enough objects left.

 i=andreas.dilger
 i=andrew.perepechko

13 years agoUpdated service tags id.
Terry Rutledge [Tue, 15 Jun 2010 16:21:26 +0000 (09:21 -0700)]
Updated service tags id.

13 years agob=21109 mds_lov_read_objids cleanup and conf-sanity tests (1.8)
Vladimir Saveliev [Fri, 11 Jun 2010 07:09:40 +0000 (11:09 +0400)]
b=21109 mds_lov_read_objids cleanup and conf-sanity tests (1.8)

calculate mds->mds_lov_objid_lastidx and mds->mds_lov_objid_lastpage correctly
have index strings big enough to store indicies in decimal
test to check size of lov_objid and
test to check configuraion with big indicies
minor compile fix to avoid compile warnings from gcc 4.0.2

i=adilger
i=panda
i=nathan

13 years agob=22121 report the right errno in l_getgroups
Andrew Perepechko [Tue, 15 Jun 2010 10:13:52 +0000 (14:13 +0400)]
b=22121 report the right errno in l_getgroups

sometimes, errno from getpwuid could be overridden by
the calls to syslog facitilites, save it for the later
report

i=Johann Lombardi

13 years agob=22911 Don't clean EXTENTS_FL flag.
yangsheng [Tue, 15 Jun 2010 08:15:18 +0000 (16:15 +0800)]
b=22911 Don't clean EXTENTS_FL flag.

i=adilger
i=girish

13 years agouse CNETERR instead of CDEBUG(D_NETERROR ...)
Johann Lombardi [Tue, 15 Jun 2010 09:56:20 +0000 (11:56 +0200)]
use CNETERR instead of CDEBUG(D_NETERROR ...)

13 years agob=21951 Init scatterlist table.
yangsheng [Tue, 15 Jun 2010 05:23:38 +0000 (13:23 +0800)]
b=21951 Init scatterlist table.

i=liangzhen
i=rahul

13 years agob=22612 dig out Module.symvers for sles10
Brian J. Murrell [Mon, 14 Jun 2010 18:25:08 +0000 (14:25 -0400)]
b=22612 dig out Module.symvers for sles10

When we build modules in lbuild for sles10 kernels, we point to the
contents of the kernel-source RPM.  That does not contain the
Module.symvers file though and thus we get a nice big warning that
the modules built (i.e. ldiskfs) will not contain versions.
This is because we need to unpack the /usr/src/linux-*obj/Module.symvers
file from the kernel-$flavour or kernel-syms RPM into the location pointed to
by --with-linux-obj.

Fix a bug in unpack_linux_devel_rpm-$DISTRO where we were not checking
for the success of the function.

i=mjmac
i=wangyb

13 years agob=22990 Script to update copyright on recently modified source files
Robert Read [Mon, 14 Jun 2010 23:37:33 +0000 (16:37 -0700)]
b=22990 Script to update copyright on recently modified source files

Based on the original script by Andreas.

i=adilger
i=johann

13 years agob=23039 Fix the build failed of sles9.
yangsheng [Fri, 11 Jun 2010 14:32:50 +0000 (22:32 +0800)]
b=23039 Fix the build failed of sles9.

i=johann

13 years agob=15253 fix conf-sanity 30b for non-tcp networks
Nathan Rutman [Thu, 10 Jun 2010 23:16:57 +0000 (16:16 -0700)]
b=15253 fix conf-sanity 30b for non-tcp networks

13 years agob=13698 support for remote mds and oss in lfscktest.sh
Vladimir Saveliev [Wed, 9 Jun 2010 23:53:28 +0000 (03:53 +0400)]
b=13698 support for remote mds and oss in lfscktest.sh

SHARED_DIRECTORY is to be specified in case of remote servers
LFSCK_ALWAYS flag to check the filesystem from check_and_cleanup_lustre()
it is "no"  by default
SKIP_LFSCK flag to allow e2fsck part of lfscktest.sh to be run without lfsck
it is "yes" by default

i=adliger
i=grev
i=jian.yu

13 years agob=21846 fix the test_32 in sanity-quota.sh and enable it
Landen [Thu, 10 Jun 2010 08:43:23 +0000 (16:43 +0800)]
b=21846 fix the test_32 in sanity-quota.sh and enable it

i=johann

13 years agob=22890 1.8<->2.0 interop: fix conf-sanity test 24a and 33a
Jian Yu [Thu, 10 Jun 2010 01:38:50 +0000 (09:38 +0800)]
b=22890 1.8<->2.0 interop: fix conf-sanity test 24a and 33a

Fix conf-sanity test 24a and 33a to work under 1.8<->2.0 interop configuration.

i=grev
i=vladimir.saveliev

13 years agob=22237 fix issue with proc_remove
Rahul Deshmukh [Thu, 10 Jun 2010 21:28:30 +0000 (23:28 +0200)]
b=22237 fix issue with proc_remove

i=johann

13 years agob=22990 update copyright messages for Oracle
Johann Lombardi [Thu, 10 Jun 2010 13:44:26 +0000 (15:44 +0200)]
b=22990 update copyright messages for Oracle

13 years agob=15253 add failover nidlist to proc import
Nathan Rutman [Wed, 9 Jun 2010 20:44:50 +0000 (13:44 -0700)]
b=15253 add failover nidlist to proc import

i=adilger
i=rread

13 years agob=22934 fix writeconf in conf-sanity
Nathan Rutman [Wed, 9 Jun 2010 18:56:24 +0000 (11:56 -0700)]
b=22934 fix writeconf in conf-sanity

fix writeconf in conf-sanity
fix conf-sanity slow list
some better error checking in conf_sanity from 22671

13 years agob=22977 add echoclient async journal regression test to acc-sm obdfilter-survey
Elena Gryaznova [Wed, 9 Jun 2010 18:36:17 +0000 (22:36 +0400)]
b=22977 add echoclient async journal regression test to acc-sm obdfilter-survey

i=Oleg.Drokin

13 years agob=22977 re-struct obdfilter-survey.sh
Elena Gryaznova [Wed, 9 Jun 2010 18:17:22 +0000 (22:17 +0400)]
b=22977 re-struct obdfilter-survey.sh

i=Andrew.Perepechko

13 years agob=22978 fix compiler warning error
Brian J. Murrell [Wed, 9 Jun 2010 17:43:10 +0000 (13:43 -0400)]
b=22978 fix compiler warning error

Fix the warning/error:
sanity.c: In function ‘t21’:
sanity.c:708: error: suggest parentheses around operand of ‘!’ or change ‘&’ to ‘&&’ or ‘!’ to ‘~’

i=panda
i=whitebear

13 years agob=22976 fix bash syntax error
Brian J. Murrell [Wed, 9 Jun 2010 16:51:01 +0000 (12:51 -0400)]
b=22976 fix bash syntax error

When using "test" to compare strings, both sides of the comparison
should either be quoted or not quoted, not one of each.
= should be used for equality in portable tests.

i=mjmac
i=whitebear

13 years agob=22975 reduce stack usage to avoid a gcc error
Brian J. Murrell [Wed, 9 Jun 2010 16:51:00 +0000 (12:51 -0400)]
b=22975 reduce stack usage to avoid a gcc error

  CC [M]  lnet/libcfs/module.o
cc1: warnings being treated as errors
lnet/libcfs/module.c: In function ‘libcfs_ioctl’:
lnet/libcfs/module.c:337: error: the frame size of 1036 bytes is larger than 1024 bytes
make[6]: *** [lnet/libcfs/module.o] Error 1

i=johann
i=panda

13 years agob=22923 increase show_error timeout from 5 to 10
Elena Gryaznova [Wed, 9 Jun 2010 16:01:07 +0000 (20:01 +0400)]
b=22923 increase show_error timeout from 5 to 10

i=Maxim.Patlasov

13 years agob=22923 increase lst stat timeout
Elena Gryaznova [Wed, 9 Jun 2010 15:57:39 +0000 (19:57 +0400)]
b=22923 increase lst stat timeout

i=Maxim.Patlasov

13 years agob=11063 fix lock inversion
Vladimir Saveliev [Wed, 9 Jun 2010 12:55:27 +0000 (16:55 +0400)]
b=11063 fix lock inversion

unlock inode mutex before extent lock acquisition to avoid lock order inversion
few misspells are fixed

i=green
i=vitaly

13 years agob=22679 quota fixes for mds_orphan_add_link
Andrew Perepechko [Wed, 9 Jun 2010 01:32:43 +0000 (05:32 +0400)]
b=22679 quota fixes for mds_orphan_add_link

allow quota overruns when linking orphans to PENDING

i=Johann Lombardi
i=ZhiYong Tian

13 years agob=22481 Add "lfs_migrate" script from manual into lustre/scripts and RPMs
Andreas Dilger [Fri, 4 Jun 2010 17:54:07 +0000 (11:54 -0600)]
b=22481 Add "lfs_migrate" script from manual into lustre/scripts and RPMs

Add enhanced lfs_migrate script from manual into lustre/scripts directory
and in RPM packages.  It does a "poor man's" migration of files from
their current OST layout to a new OST layout as chosen by the MDS.
It is currently not safe to use for files that are being written to.
It will result in the current file becoming an open-unlinked file, and
the migrated file will be used for any subsequent opens.  Any writes to
either the old or migrated file will in clients seeing different data,
and writes to the old file will be lost when it is closed.

There is no direct control of object allocation within the script, it
depends on external control to manage space.  Use "lfs find" to
select appropriate files to migrate from full OSTs. "lctl --deactivate"
is needed on the MDS to deactivate OSTs if trying to empty an OST
completely, otherwise QOS will tend to migrate files to empty OSTs.

i=brian.murrell
i=nathan.rutman

13 years agob=22476 shrink dlm slab on servers even if __GFP_FS is not set v1_8_3_53
Johann Lombardi [Fri, 4 Jun 2010 21:35:21 +0000 (23:35 +0200)]
b=22476 shrink dlm slab on servers even if __GFP_FS is not set

i=vitaly
i=oleg

13 years agobump version to 1.8.3.53
Johann Lombardi [Fri, 4 Jun 2010 21:31:18 +0000 (23:31 +0200)]
bump version to 1.8.3.53

13 years agob=21877 protect bitfield access to rq_flags with rq_lock
ZhangHongChao [Fri, 4 Jun 2010 21:28:13 +0000 (23:28 +0200)]
b=21877 protect bitfield access to rq_flags with rq_lock

i=rread
i=fanyong
i=johann

The AT code can access some bits of rq_flags while the service thread
is updating some other bits. Everyone shoud use the rq_lock for
consistency.

13 years agob=21900 verify that the pool ost is available
Nathan Rutman [Fri, 4 Jun 2010 21:17:05 +0000 (14:17 -0700)]
b=21900 verify that the pool ost is available

i=dmitry.zoguine

13 years agostrip the .patch off of the dpatch name
Brian J. Murrell [Fri, 4 Jun 2010 21:00:36 +0000 (17:00 -0400)]
strip the .patch off of the dpatch name

When annotating the .dpatch files with the patch name, seeing as we
strip the .patch off and replace it with dpatch, do the same in the
annotation.

13 years agob=22911 don't enable extent for MDT by default
Girish Shilamkar [Fri, 4 Jun 2010 21:20:11 +0000 (23:20 +0200)]
b=22911 don't enable extent for MDT by default

i=johann
i=adilger

13 years agob=22755 Don't leak grant if shink grant comes in write RPC.
Oleg Drokin [Fri, 4 Jun 2010 15:26:01 +0000 (17:26 +0200)]
b=22755  Don't leak grant if shink grant comes in write RPC.

i=adilger
i=wangdi
i=johann

13 years agob=22480 Remove duplicated LB_DEFINE_E2FSPROGS_NAMES definition
Johann Lombardi [Fri, 4 Jun 2010 15:03:58 +0000 (17:03 +0200)]
b=22480 Remove duplicated LB_DEFINE_E2FSPROGS_NAMES  definition

Revert "Add configure option to allow repackaging of e2fsprogs as ldiskfsprogs. (by Jim Garlick)"

This reverts commit 749b473d6e38b94e9b4d88214ce8684f3c570f4e.

13 years agob=22480 Remove duplicated LB_DEFINE_E2FSPROGS_NAMES definition
Johann Lombardi [Fri, 4 Jun 2010 15:03:01 +0000 (17:03 +0200)]
b=22480 Remove duplicated LB_DEFINE_E2FSPROGS_NAMES  definition

Revert "Add configure option to allow repackaging of e2fsprogs as ldiskfsprogs. (by Jim Garlick)"

This reverts commit 8233d53e507a7dc247336c319ff80a453bf18207.

13 years agob=22889 OSS Kernel panic - Kernel BUG at fs/jbd/transaction.c:1155
Dmitry Zogin [Fri, 4 Jun 2010 14:06:24 +0000 (10:06 -0400)]
b=22889 OSS Kernel panic - Kernel BUG at fs/jbd/transaction.c:1155

  extend truncate transaction to have enough blocks for last_rcvd and VBR

 o=andreas.dilger
 i=tappro
 i=oleg.drokin

13 years agob=22747 Processes stuck waiting for inode I_LOCK bit to clear.
Dmitry Zogin [Fri, 4 Jun 2010 13:56:12 +0000 (09:56 -0400)]
b=22747 Processes stuck waiting for inode I_LOCK bit to clear.

 Use CFS_ALLOC_IO instead of CFS_ALLOC_STD in OBDO_ALLOC

 i=andreas.dilger
 i=johann

13 years agob=21556 release cpu if needed while removing extents
Johann Lombardi [Fri, 4 Jun 2010 10:19:47 +0000 (12:19 +0200)]
b=21556 release cpu if needed while removing extents

i=oleg
i=vitaly

cache_remove_extents_from_lock() can hog cpu for too long.
add a call to cond_resched().

13 years agodiscover and utilize multiple CPUs
Brian J. Murrell [Wed, 2 Jun 2010 21:56:57 +0000 (17:56 -0400)]
discover and utilize multiple CPUs

13 years agogenerate debian/patches
Brian J. Murrell [Wed, 2 Jun 2010 21:56:58 +0000 (17:56 -0400)]
generate debian/patches

If we are running from a git checkout and there are patches on the
branch, (not counting the always-present debian/changelog update
patch, which is always assumed to be at the top of the patch stack)
put them into debian/patches/ as dpatches and remove them from the
source so that what's put into the source tarball is pristine.

13 years agob=22790 fail make debs if autogen.sh is needed but not present
Brian J. Murrell [Wed, 2 Jun 2010 21:56:56 +0000 (17:56 -0400)]
b=22790 fail make debs if autogen.sh is needed but not present

If a patch is added to debian/patches/ which modifies an autoconf
file but autogen.sh is not present, fail out of the "make debs"
build command.
Also provide a more explicit message to maintainers about what is
happening in this bit of code.

13 years agob=22658 More tolerant llog processing
Andrew Perepechko [Wed, 2 Jun 2010 17:54:27 +0000 (21:54 +0400)]
b=22658 More tolerant llog processing

Do not fail OST activation when a llog is not found.

i=Johann Lombardi
i=Mike Pershin

13 years agob=21500 remove duplicate LC_MODULE_LOADING def'n
Brian J. Murrell [Tue, 1 Jun 2010 19:55:23 +0000 (15:55 -0400)]
b=21500 remove duplicate LC_MODULE_LOADING def'n

The whole LC_MODULE_LOADING definition was duplicated.  Remove one of them.

13 years agob=22040 Don't run connectathon lock tests on nfsv4
Elena Gryaznova [Tue, 1 Jun 2010 19:00:01 +0000 (23:00 +0400)]
b=22040 Don't run connectathon lock tests on nfsv4

a=Robert.Read
i=Oleg.Drokin

We won't support flock on nfsv4 until 14080 has been fixed.

13 years agob=22899 reduce OBDFILTER_SURVEY duration
Elena Gryaznova [Tue, 1 Jun 2010 18:40:15 +0000 (22:40 +0400)]
b=22899 reduce OBDFILTER_SURVEY duration

i=Andreas.Dilger

13 years agob=22360 for close, map all errors to EIO
Andrew Perepechko [Mon, 31 May 2010 23:39:47 +0000 (03:39 +0400)]
b=22360 for close, map all errors to EIO

i=Johann Lombardi
i=ZhiYong Tian

14 years agob=22244/21128 remove changelog entry
Johann Lombardi [Mon, 31 May 2010 23:22:49 +0000 (01:22 +0200)]
b=22244/21128 remove changelog entry

14 years agoRevert "b=21128 send cancel rpc in a separate thread"
Johann Lombardi [Mon, 31 May 2010 23:21:28 +0000 (01:21 +0200)]
Revert "b=21128 send cancel rpc in a separate thread"

This reverts commit 578d08cde3abba552adbe6bcc1b88c6f1552ce10.
Reverting all patches from 22244 & 21128 because of test failures.
See 22244 for more info.

14 years agoRevert "b=22244 ldlm cancel flags cleanup"
Johann Lombardi [Mon, 31 May 2010 23:20:28 +0000 (01:20 +0200)]
Revert "b=22244 ldlm cancel flags cleanup"

This reverts commit 80786e8223938e1f3379ba409dec61a15dd2fc60.
Reverting all patches from 22244 & 21128 because of test failures.
See 22244 for more info.