Cluster Project branch, RHEL47, created. gfs-kernel_2_6_9_76-44-gc9f85c1
cfeist@sourceware.org
cfeist@sourceware.org
Tue Apr 15 20:01:00 GMT 2008
This is an automated email from the git hooks/post-receive script. It was
generated because a ref change was pushed to the repository containing
the project "Cluster Project".
http://sources.redhat.com/git/gitweb.cgi?p=cluster.git;a=commitdiff;h=c9f85c1b3cccdd529005d08fb8e4618f752acaab
The branch, RHEL47 has been created
at c9f85c1b3cccdd529005d08fb8e4618f752acaab (commit)
- Log -----------------------------------------------------------------
commit c9f85c1b3cccdd529005d08fb8e4618f752acaab
Author: Lon Hohberger <lhh@redhat.com>
Date: Tue Apr 15 15:20:32 2008 -0400
[rgmanager] Set +x on SAPInstance/SAPDatabase
commit 8dc51a64cf9a0bc1b467353a4161220a7cec0de0
Author: Christine Caulfield <ccaulfie@redhat.com>
Date: Tue Apr 15 16:34:48 2008 +0100
[DLM] Add tcp_nodelay option to DLM comms
Under some very odd loads, the TCP nagle algorithm can cause
annoying delays in lock traffic. This option allows it to
be switched off with
echo "1" > /proc/cluster/config/dlm/tcp_nodelay
This should NOT be a default, in most cases it will cause
a huge performance drop. It is global to all DLM communications
and cannot be changed per lockspace.
It is included specifically for bz#244708.
Signed-off-by: Christine Caulfield <ccaulfie@redhat.com>
commit 69a3d82a661d0c67dadfc0c02a6839c94ab0cdfb
Author: Lon Hohberger <lhh@redhat.com>
Date: Tue Apr 15 11:02:33 2008 -0400
[rgmanager] Fix several bugzillas (see below)
- 245381 - restart counter/thresholds before switching to
relocation during service recovery
- 247772 - one service following another (event scripting
ability). This isn't fixed in the usual way, as
this feature has limited use. Instead, we allow
an alternative method of operation for rgmanager
in which administrators can define event triggers
for whatever they would like. (Called "RIND")
- 247945 - rgmanager restarting services when noncritical
parameters are changed
- 247980 - strong / weak service dependencies. These are
trivial. Weak dependencies are only allowed with
the event scripting as noted for 247772
- 250101 - RG event API (internals change). This was
required for RIND to operate as well as 247980
- 439948 - clushutdown reference in clusvcadm man page
- 440006 - rgmanager stuck on stop in some cases
- 440645 - quotaoff causes hanges in some circumstances
- 441577 - Symlinks in mount point path cause erroneous failures
commit 24945860f4de6118437688aa3e4b7bb3e3edbbf0
Author: Lon Hohberger <lhh@redhat.com>
Date: Tue Apr 8 00:32:04 2008 -0400
[rgmanager] Fix minor crash bug
commit 070de9cc3db1a66d427a2aaaef0dbf71d4d404b9
Author: Lon Hohberger <lhh@redhat.com>
Date: Fri Apr 4 13:54:28 2008 -0400
[rgmanager] Fixes for RIND
commit 179da28103b843845cad5eaec61a48855b5cf807
Author: Lon Hohberger <lhh@redhat.com>
Date: Fri Apr 4 11:20:09 2008 -0400
[rgmanager] More RIND merges.
commit 7c3f3ad887d460c2bd92a7a398048be58f1b4a07
Author: Lon Hohberger <lhh@redhat.com>
Date: Tue Apr 1 16:25:08 2008 -0400
[rgmanager] Fix clusvcadm build
commit d35287124adf15a981703c3e0785fa7c644223c1
Author: Lon Hohberger <lhh@redhat.com>
Date: Tue Apr 1 16:23:17 2008 -0400
[rgmanager] Fix build
commit fed6e7804f9b092b09a1d243abd03eb87747e566
Author: Lon Hohberger <lhh@redhat.com>
Date: Tue Apr 1 16:22:15 2008 -0400
[rgmanager] Remove extraneous copy of event_script.txt
commit 71a51fb8ed4c91c5963665940afca5fc44f42559
Author: Lon Hohberger <lhh@redhat.com>
Date: Tue Apr 1 16:17:55 2008 -0400
[rgmanager] Misc fixes. Update auto testing
commit 652fe516bfa1efd5ceff9e1551e8dabc573e294d
Author: Lon Hohberger <lhh@redhat.com>
Date: Tue Apr 1 12:54:15 2008 -0400
[rgmanager] Make RIND build on RHEL4 branch
commit 0d3a7a758217efbfec3a662269856a30d68fdd63
Author: Lon Hohberger <lhh@redhat.com>
Date: Mon Mar 10 09:58:08 2008 -0400
Merge, part 2
commit 00a395dd63da4728c30b61a04a71224a4f05b82d
Author: Lon Hohberger <lhh@redhat.com>
Date: Mon Mar 10 09:56:40 2008 -0400
Commit phase 1 of update from rhel5 branch
commit 5f03c06d964894819c182b5112af75e5ee44a256
Author: Bob Peterson <rpeterso@redhat.com>
Date: Sun Apr 13 11:23:49 2008 -0500
bz440896/440897 GFS: gfs_fsck should repair gfs_grow corruption
(see bug #436383)
commit 625364c99277445b97bbb3120736c358b802fb0f
Author: Jonathan Brassow <jbrassow@redhat.com>
Date: Mon Apr 7 10:48:24 2008 -0500
- dm-cmirror.ko: Fix infinite election loop (bug 217895)
It was possible for an election to be run just before
a machine that was next in line to be server suspened.
This would cause the remaining nodes to think a suspended
node was the active server - worse, the suspended node
would allow itself to become server and then say in was
not in charge of the log. This would simply keep going
round and round.
Fix is to increment 'suspended' before exiting postsuspend,
and not allowing a node to be elected server if 'suspended'
was set.
commit aac75cba36115a59a489c0b1a52be5259f54fb87
Author: Jonathan Brassow <jbrassow@redhat.com>
Date: Tue Apr 8 11:30:38 2008 -0500
dm-cmirror.ko: Part of the fix for bug 193067 (cluster pvmove)
Allow mirror log to be created without the 'block_on_error'
flag.
commit 9133a6446a47c56bd735d01d761f0c91f8491fe0
Author: Ryan O'Hara <rohara@redhat.com>
Date: Mon Apr 7 11:13:40 2008 -0500
BZ 409391 : fence_scsi and scsi_reserve script no longer use state file for
tracking registered devices (/var/run/scsi_reserve).
commit 481965dc9ddb3cbbe06e82ecd507a2b77621206a
Author: Lon Hohberger <lhh@redhat.com>
Date: Mon Apr 7 09:47:01 2008 -0400
[cman] Resolve 441266 - uninitialized variable causing qdiskd to fail
commit 6626fda2200aa1abed902646371eee265be7fc04
Author: Ryan O'Hara <rohara@redhat.com>
Date: Fri Apr 4 17:52:41 2008 -0500
BZ 409391 : fence_scsi and scsi_reserve script no longer use state file for
tracking registered devices (/var/run/scsi_reserve).
commit c128baa7cc5397f1fcfc72d35cdb39c089c56043
Author: Jonathan Brassow <jbrassow@redhat.com>
Date: Thu Apr 3 10:48:46 2008 -0500
rgmanager/lvm.sh: Fix bug 431705
- Add another check for improper HA LVM setup. Specifically,
check for fail-over VG name in 'volume_list', which would
allow a machine to activate a VG regardless of tagging.
- Clearer error messages
- Better comments
commit 0536f85e62db2ac86b0f47c6e6089090261eceae
Author: Bob Peterson <rpeterso@redhat.com>
Date: Wed Apr 2 16:07:53 2008 -0500
Resolves: bz 436383: GFS filesystem size inconsistent
commit 36e10fa435afc59e5a0e658c1b9b7ee2c34f184a
Author: Jonathan Brassow <jbrassow@redhat.com>
Date: Wed Apr 2 14:25:22 2008 -0500
rgmanager/lvm.sh: Make lvm.sh call into lvm_by_*.sh (RE: bz427377)
lvm_by_*.sh scripts are updated. Now call them. This makes
lvm.sh a wrapper around lvm_by_*.sh scripts
commit a0e32d4eea87f15b2e07c8f61ecd53bd2462e73f
Author: Jonathan Brassow <jbrassow@redhat.com>
Date: Wed Apr 2 14:19:02 2008 -0500
rgmanager/lvm.sh: Updated fix for bz242798
When fixing bug Bug 427377 (allow multiple LVs / VG),
I failed to commit the changes to lvm.sh - which
effectively just became a shell that called functions
in other files.
Now, I need to update the other files with recent bug
fixes and then make the change that was necessary for
bug 427377 - removing the functional contents of lvm.sh.
commit d0aecbf552f27bb2bf9d3c66015dd4d8df559dd8
Author: Jonathan Brassow <jbrassow@redhat.com>
Date: Tue Apr 1 15:24:50 2008 -0500
rgmanager/lvm.sh: Fix bug bz242798
Allow a machine to fence itself in the event that it
cannot deactivate logical volumes. (The user must
explicitly enable this option.) This is useful in cases
where one machine in the cluster looses connectivity to
its resources, but the others don't. The machine fences
itself and the service moves to another machine.
commit 2114f9e99321008beef4cc18bcbe4aa8712015f9
Author: Jonathan Brassow <jbrassow@redhat.com>
Date: Wed Mar 26 09:02:53 2008 -0500
rgmanager/lvm.sh: Fix for bug 438997
Missing VG name parameter to 'vgchange --deltag' call
caused tag to be remove from all VGs, not just the
intended one.
Thanks to Simone Gotti for the patch.
commit 22b442ed10069372d306e7f0fd162cf5387c5211
Author: Bob Peterson <rpeterso@redhat.com>
Date: Tue Mar 25 13:33:48 2008 -0500
Resolves: bz 435469: gfs_tool: Cannot allocate memory
commit 9d2d37ee3db933f01c8e5021d6d6a5897d5fe017
Merge: 8c6094c0c5d457c9572ff386c5b53f075ce54f9c c118d0ce03910523ebd6cd533a410283640caf37
Author: Jonathan Brassow <jbrassow@redhat.com>
Date: Thu Mar 20 10:55:12 2008 -0500
Merge branch 'RHEL4' of ssh://jbrassow@sources.redhat.com/git/cluster into rhel4
commit 8c6094c0c5d457c9572ff386c5b53f075ce54f9c
Author: Jonathan Brassow <jbrassow@redhat.com>
Date: Thu Mar 20 10:37:50 2008 -0500
dm-cmirror.ko: change a blocking I/O to async I/O
Sometimes, the mirror subdevices get suspended before
the log server suspends. This can lead to the server issuing
I/O requests on devices that are suspended. If it waits for the
I/O to complete, it could wait indefinitly. Now there is a
timeout in place for the I/O, which should prevent the server
from getting stuck.
commit c118d0ce03910523ebd6cd533a410283640caf37
Author: Bob Peterson <rpeterso@redhat.com>
Date: Fri Mar 14 11:17:46 2008 -0500
Resolves: bz 421761: 'gfs_tool lockdump' wrongly says 'unknown
mountpoint' re HP cciss RAID array
commit a623bf5bbecedd41e592d584c790f68fd218b569
Author: Bob Peterson <rpeterso@redhat.com>
Date: Fri Mar 14 11:09:09 2008 -0500
Resolves: bz 431945: GFS: gfs-kernel should use device major:minor
commit b2fb1017061f86f02f3cd864e10b44e9fa347441
Author: Lon Hohberger <lhh@redhat.com>
Date: Tue Mar 11 16:14:01 2008 -0400
Revert "Commit phase 1 of update from rhel5 branch"
This reverts commit df101a4507641607e5869f2f15c05cb1be6e0639.
This wasn't finished; I committed from the wrong branch.
commit edf5a1e10e95e0b02afbbc1e0a260141b128695e
Author: Lon Hohberger <lhh@redhat.com>
Date: Tue Mar 11 16:13:32 2008 -0400
Revert "Merge, part 2"
This reverts commit 43571d7af75bdb6a0b19d40048da1491b6499efb.
This wasn't finished. I pushed from the wrong branch :(
commit db6a74a9900257848aa4c07256b77b2d10a7db13
Author: Lon Hohberger <lhh@redhat.com>
Date: Tue Mar 4 17:31:32 2008 -0500
[CMAN] make qdisk master abdication work in all cases
Red Hat bugzilla 430264
commit 5aeaf6345cd4672e0ef52a753b4f6cf56a64d5b5
Author: Lon Hohberger <lhh@redhat.com>
Date: Tue Mar 11 14:54:49 2008 -0400
[rgmanager] Make ip.sh check link states of non-ethernet devices
Resolves: Red Hat Bugzilla #331661
commit 43571d7af75bdb6a0b19d40048da1491b6499efb
Author: Lon Hohberger <lhh@redhat.com>
Date: Mon Mar 10 09:58:08 2008 -0400
Merge, part 2
commit df101a4507641607e5869f2f15c05cb1be6e0639
Author: Lon Hohberger <lhh@redhat.com>
Date: Mon Mar 10 09:56:40 2008 -0400
Commit phase 1 of update from rhel5 branch
commit 1befab3b0e83584f08b06b816adbc0f8b9dc11ab
Merge: 2a08798bc9b438789e315e638f1911223863ed4f 595e3cad6ac837bd37a7bd6b9280d1ca0c1e868c
Author: Chris Feist <cfeist@redhat.com>
Date: Tue Mar 4 18:08:41 2008 -0600
Merge branch 'RHEL4' of ssh://sources.redhat.com/git/cluster into r4
commit 2a08798bc9b438789e315e638f1911223863ed4f
Author: Chris Feist <cfeist@redhat.com>
Date: Tue Mar 4 17:49:04 2008 -0600
Fixed makefiles to build.
commit 595e3cad6ac837bd37a7bd6b9280d1ca0c1e868c
Author: Lon Hohberger <lhh@redhat.com>
Date: Tue Mar 4 17:58:50 2008 -0500
[rgmanager] Add Oracle 10g database failover agent to RHEL4 branch
commit b95f5dd435df474b5b5c55927ae86b52f21c0b28
Author: David Teigland <teigland@redhat.com>
Date: Thu Feb 28 09:49:26 2008 -0600
lock_dlm: skip busy plocks in unlock
bz 432707
Another problem with multiple threads manipulating the same
plocks (also see bz 383391). In this case thread1 was waiting
on a plock that was in the dlm, when thread2 came and tried to
unlock it (due to the process exiting and closing all files).
dlm returns EBUSY for the unlock, which causes lock_dlm to assert
(another work around could be to just not assert on EBUSY.)
The "busy" flag was recently added to plocks to work around
another similar problem with multiple threads; use that to
just skip the unlock on any plock that's busy.
commit 9e8a8cd7262ad0228c1d1311d2f6ea45f01034fa
Author: Lon Hohberger <lhh@redhat.com>
Date: Tue Feb 26 16:48:23 2008 -0500
Resolve bug 434886
commit 29dfb2374ede46fe2087f2159c91721e035af16c
Author: Lon Hohberger <lhh@ayanami.boston.devel.redhat.com>
Date: Mon Feb 25 16:57:30 2008 -0500
Correct incorrect netmask handling in ip.sh
commit 7cd1be561cdb08155964672d0afc2cbe32a5af37
Author: Lon Hohberger <lhh@redhat.com>
Date: Fri Feb 8 18:38:41 2008 +0000
Fix unused variable error
commit 515aa3da10e9d73a238cc017ef9a5e4dc9556266
Author: Ryan O'Hara <rohara@redhat.com>
Date: Wed Feb 6 17:39:17 2008 +0000
BZ 431002 - Support for LVM stripes, mirrors.
commit 44d293d8ff4b39f31a993fad22bb7b353621dfe5
Author: Benjamin Marzinski <bmarzins@redhat.com>
Date: Tue Jan 29 22:21:45 2008 +0000
Fix for bz #419391. gfs_glock_dq was traversing the gl_holders list without
holding the gl_spin spinlock, this was causing a problem when the list item
it was currently looking at got removed from the list. The solution is to
not traverse the list, because it is unncessary. Unfortunately, there is also
a bug in this section of code, where you can't guarantee that you will not
cache a glock held with GL_NOCACHE. Fixing this issue requires significantly
more work.
commit efad9110a2e0d467f4b548a6a24b77d45b4148ee
Author: Lon Hohberger <lhh@redhat.com>
Date: Mon Jan 28 19:42:55 2008 +0000
Fix #292861 - nfsclient doesn't unexport correctly
commit 0844bfc6e6d6a4ffb16752ad6520abc655ecb848
Author: Lon Hohberger <lhh@redhat.com>
Date: Mon Jan 28 19:34:32 2008 +0000
Fix #289521 - clulog not respecting rgmanager log level
commit cbae6db7de4834347b964953692b55c35bfc99fa
Author: Lon Hohberger <lhh@redhat.com>
Date: Mon Jan 28 19:27:20 2008 +0000
Fix #430538 - unbounded status/clustat thread counts cause timeout errors
commit 8755d2e9cbb11871c6c01e2b4f73e20808eac039
Author: Lon Hohberger <lhh@redhat.com>
Date: Thu Jan 24 14:58:38 2008 +0000
Unblock signals after fork() so heuristics using signals don't hang
commit 05506c5b8f9744f878a4b6790e49ab1d7047f320
Author: Ryan McCabe <rmccabe@redhat.com>
Date: Tue Jan 15 16:21:17 2008 +0000
fix bz428243: fence_ilo can not recognize option="off" but fence_wti can
commit 6242d7d010107098e320b866cdae550c31ef5a45
Author: David Teigland <teigland@redhat.com>
Date: Mon Jan 14 16:00:20 2008 +0000
bz 352731
I've looked at other users of kthread_stop()/kthread_should_stop(),
and most seem to incorporate a call to kthread_should_stop() within
the test that adds the thread to a wait_queue. I'm guessing that
that's what we should be doing also (and at the same time using
wait_event_interruptible() instead of an open-coded equivalent.)
commit c8d815e711e20c54f38c381df40cf5a6ca75884b
Author: David Teigland <teigland@redhat.com>
Date: Mon Jan 14 15:57:46 2008 +0000
bz 351321
add_to_requestqueue() can add a new message to the requestqueue
just after process_requestqueue() checks it and determines it's
empty. This means dlm_recvd will spin forever in wait_requestqueue()
waiting for the message to be removed.
The same problem was found and fixed in the RHEL5 code (and then
subsequently changed again). This patch is the RHEL4 equivalent of the
original RHEL5 fix.
commit 54f7f27654b5691e1057dcab09370f523162da4d
Author: David Teigland <teigland@redhat.com>
Date: Mon Jan 14 15:35:30 2008 +0000
bz 324881
It's easy to tell if you've hit this bug, because a message like this will
always appear in /var/log/messages:
SM: 02000378 ignoring service callback id=2000144 event=1324
If you look at /proc/cluster/lock_dlm/debug on this node at this point,
you'll see something like this at the end, which shows what the problem
is:
others_may_mount start_done 1322 b
The event_id that others_may_mount uses when calling kcl_start_done()
is incorrect; it's using 1322 when it should be 1324.
I believe the fix is for others_may_mount() to read the event_id
after taking the umount_lock semaphore which serializes
others_may_mount() with a start callback from the lock_dlm thread.
In this case, I believe the start callback is changing the event_id
after others_may_mount reads it, and before othres_may_mount gets
the umount_lock semaphore.
commit 1801d5ee98bc29996dd6f37c3875e949f4a8c22c
Author: David Teigland <teigland@redhat.com>
Date: Mon Jan 14 15:29:42 2008 +0000
bz 353311
Disassembling the module, the oops appears to be when dereferencing
dlm->mg_nodes. I see one place where the mg_nodes list is modified
without holding mg_nodes_lock, and that's in release_mg_nodes().
commit 2e2c73cfc58c632170a3af839426a33f04b9ff1e
Author: Jonathan Brassow <jbrassow@redhat.com>
Date: Fri Jan 11 21:33:31 2008 +0000
- Bug #428475
- HA LVM service fails to relocate when I/O is running
Was failing to add new tag when relocating.
commit 56a8258560ccb6c909a9857c63f046a31105bae1
Author: Patrick Caulfield <pcaulfie@redhat.com>
Date: Fri Jan 11 14:22:11 2008 +0000
Always use the nodename in cluster.conf for cluster name.
bz#428219
commit 903b90674afb18669f232c0b61e9a6d720dbe95c
Author: Lon Hohberger <lhh@redhat.com>
Date: Wed Jan 9 18:55:12 2008 +0000
Fix #294491 - make magma_sm.so retry in dlm EINPROG condition
commit e7d713b2b9782ddf9ca85a42152d890cb38761ff
Author: Lon Hohberger <lhh@redhat.com>
Date: Wed Jan 9 18:54:46 2008 +0000
Fix #294491 - make magma_sm.so retry in dlm EINPROG condition
commit e4519b6ebf713dbce2e97adf2029c8213c9326c5
Author: Jonathan Brassow <jbrassow@redhat.com>
Date: Fri Jan 4 20:01:50 2008 +0000
- a regression... When tagging at the LV-level, the script should complain
if there is more than one LV / VG.
commit a6a2d56a114428d184739353d09eb14e1e54c173
Author: David Teigland <teigland@redhat.com>
Date: Fri Jan 4 16:12:05 2008 +0000
Some message gets out of place, but there's no need to panic
the machine; just ignore it. bz 427531
commit 2b9bf3c8e87ca307b5c026db89db5c9fdd14a590
Author: Jonathan Brassow <jbrassow@redhat.com>
Date: Thu Jan 3 20:35:39 2008 +0000
BUG 427377
HA LVM now allows multiple LVs/VG as long as they move together
Package builder, please note the addition of 3 new files.
commit 59b49a3ba34412f6e614721695eec1989979024f
Author: Wendy Cheng <wcheng@redhat.com>
Date: Wed Jan 2 19:17:09 2008 +0000
Bugzilla 227892:
1. Warn people about the RG corruption and request a gfs_fsck
2. Upon error detection, perform a minimum error data collection.
commit f4973c85c022808bd07dc7ae9d2e95bf6cebd841
Author: Ryan O'Hara <rohara@redhat.com>
Date: Tue Dec 11 20:56:45 2007 +0000
BZ 418961
Fix issue with endian conversion that caused problems for mixed architecture
nodes on same subnet. Need to correct swap byte ordering of comm_header_t
structure before copying into buffer and when dereferencing.
commit 4d91801fac0677798e84600643c3313611b32f64
Author: Lon Hohberger <lhh@redhat.com>
Date: Tue Dec 4 21:06:58 2007 +0000
Remove artificial label size limit
commit eaf3713948ba886aa9b7743da3860f6795c35a22
Author: Lon Hohberger <lhh@redhat.com>
Date: Tue Dec 4 20:40:54 2007 +0000
Make qdiskd work with sector sizes other than 512 bytes. Import patch from Fabio M. Di Nitto to make qdiskd use (node_count - 1) for votes if there's none specified in cluster.conf
commit 7169d5b23a8ddd3f656a2fb9aa1cfe1c3d873a04
Author: David Teigland <teigland@redhat.com>
Date: Mon Dec 3 16:40:41 2007 +0000
Posix locks don't work between threads, but it seems some programs do
involve threads sharing plocks incidentally (with meaningless results).
Given the plock state we keep in lock_dlm, multiple threads accessing the
locks would not only be meaningless, but could corrupt the state, leaving
the threads permanently hung. This patch tries to keep threads from
stomping on each other, hopefully preventing most hangs.
bz 383391
commit 92e2bf31caf42feb5f92663a399d0b2e188a3b41
Author: Patrick Caulfield <pcaulfie@redhat.com>
Date: Tue Nov 20 11:02:46 2007 +0000
A fix to the last patch. The last_ackneeded_seq_recv variable needed to
be cleared when the node went down, otherwise we end up comparing
received sequence numbers against old ones and end up throwing all
new messages away!
bz#387081 shows this happening. The "inconsistent" message is slightly
misleading in this context
commit c5cd62b6d2e7d8d6a5592f3dff0c9eafd7deed90
Author: Lon Hohberger <lhh@redhat.com>
Date: Wed Nov 14 19:03:37 2007 +0000
Fix #360401 - hang forever during shutdown due to previous service boot problem
commit b4f45f8d65c4f07aa0d996c43526e3cd7ca517a5
Author: Lon Hohberger <lhh@redhat.com>
Date: Wed Nov 14 18:49:51 2007 +0000
Apply patch from Marco Ceci to fix #358161
commit 3bd7e7d0931a9d6d7df04ffe7ffa3125a99f7be6
Author: Lon Hohberger <lhh@redhat.com>
Date: Wed Nov 14 17:14:21 2007 +0000
Clarify qdisk man page some
commit 7848ceda86f5aa9b32910d7ebf4d283e43880a5e
Author: Patrick Caulfield <pcaulfie@redhat.com>
Date: Mon Nov 12 14:35:20 2007 +0000
Message test program for flooding cman with messages.
commit 03f840a836e24132931d00c0f46773bf06f42d4f
Author: Ryan O'Hara <rohara@redhat.com>
Date: Thu Nov 8 17:00:44 2007 +0000
BZ 248715
- Use cluster ID and node ID for key rather than IP address.
commit ef671a3afd9591d3eb23a611c9d715684b1b48b1
Author: Lon Hohberger <lhh@redhat.com>
Date: Wed Nov 7 15:24:19 2007 +0000
Revert broken patch & apply correct patch from 338511
commit c92628dcc39e03a4e9eccc4fa76257c871e5ba00
Author: David Teigland <teigland@redhat.com>
Date: Wed Nov 7 15:22:31 2007 +0000
bz 349001
For the entire life of the dlm, there's been an annoying issue that we've
worked around and not "fixed" directly. It's the source of all these
messages:
process_lockqueue_reply id 2c0224 state 0
The problem that a lock master sends an async "granted" message for a
convert request *before* actually sending the reply for the original
convert. The work-around is that the requesting node just takes the
granted message as an implicit reply to the conversion and ignores the
convert reply when it arrives later (the message above is printed when
it gets the out-of-order reply for its convert). Apart from the annoying
messages, it's never been a problem.
Now we've found a case where it's a real problem:
1. nodeA: send convert PR->CW to nodeB
nodeB: send granted message to nodeA
nodeB: send convert reply to nodeA
2. nodeA: receive granted message for conversion
complete request, sending ast to gfs
3. nodeA: send convert CW->EX to nodeB
4. nodeA: receive reply for convert in step 1, which we ordinarily
ignore, but since another convert has been sent, we mistake this
message as the reply for the convert in step 3, and complete
the convert request which is *not* really completed yet
5. nodeA: send unlock to nodeB
nodeB: complains about an unlock during a conversion
The fix is to have nodeB not send a convert reply if it has already sent a
granted message. (We already do this for cases where the conversion is
granted when first processing it, but we don't in cases where the grant
is done after processing the convert.)
commit 3417dfe50b6acd71ce568bb427136d6f2093e105
Author: Ryan O'Hara <rohara@redhat.com>
Date: Fri Nov 2 21:08:50 2007 +0000
BZ 364661 - Fix ccs_tool to return EXIT_SUCCESS for most commands.
commit a25f4f8608fa941e2d9d1a3ce092c6e579e103a8
Author: Patrick Caulfield <pcaulfie@redhat.com>
Date: Thu Nov 1 10:44:25 2007 +0000
When checking for old messages, compare against the last ACKed message rather
than the last seen message (which could be a NOACK message). otherwise we end
up throwing awa messages if they arrive out of order.
bz#299061
commit 3738514b6c4fe15cc6380bdf6fe4df33ea99ec2c
Author: Lon Hohberger <lhh@redhat.com>
Date: Tue Oct 30 17:41:13 2007 +0000
Fix #338511 - dlm trying to connect from wrong IP
commit 9e715649d5341d3c537d4aa3f162398685270c94
Author: Ryan McCabe <rmccabe@redhat.com>
Date: Mon Oct 29 20:38:13 2007 +0000
Fix a bunch of format string errors.
commit f5bb753aed627c49a87564ff2ce6725ff6fe4296
Author: Lon Hohberger <lhh@redhat.com>
Date: Fri Oct 26 20:27:39 2007 +0000
Fix bugzilla #354391 - apps using magmamsg don't connect from correct source IP on multi-homed hosts
commit e68c57b32c5b664205321079100bf8dbb41028d4
Author: Lon Hohberger <lhh@redhat.com>
Date: Fri Oct 26 20:26:47 2007 +0000
Fix bugzilla #298831 / #298861 - apps using magmamsg don't connect from correct source IP on multi-homed hosts
commit 3df5ddd843b1ac58542d91635ff3ec4cf8efd4e2
Author: Lon Hohberger <lhh@redhat.com>
Date: Fri Oct 26 20:23:46 2007 +0000
Fix bugzilla #298831 - apps using magmamsg don't connect from correct source IP on multi-homed hosts
commit 4c1c792c84cd7cb8ab729c347835bcac6c64c753
Author: Lon Hohberger <lhh@redhat.com>
Date: Fri Oct 26 20:15:37 2007 +0000
Fix bugzillas #298831, #298861, #354391 - apps using magmamsg don't connect from correct source IP on multi-homed hosts
commit 7ecf055077aab17873a916ff3a46f4c0562513ff
Author: Jonathan Brassow <jbrassow@redhat.com>
Date: Fri Oct 26 18:46:10 2007 +0000
-take out annoying message.
commit ca4fbfc2aa0ae450667512399267a89eea926f06
Author: Ryan O'Hara <rohara@redhat.com>
Date: Fri Oct 12 18:16:30 2007 +0000
Update man page to explain 'cman_tool nodes' filter/format options.
commit cae74d3be6497e35dbd1603147eb4ae70ef73c9b
Author: Ryan O'Hara <rohara@redhat.com>
Date: Fri Oct 12 18:10:12 2007 +0000
Add ability to format output and filter based on node name.
commit ed08da2f721563274c8d782ed969708ee205384a
Author: Ryan McCabe <rmccabe@redhat.com>
Date: Wed Oct 10 17:13:57 2007 +0000
Don't use 'new' as an identifier in function prototypes to allow the header to be included by C++ programs.
commit effdfce319d004ccb14e214fa420cebb08bbc06c
Author: Lon Hohberger <lhh@redhat.com>
Date: Wed Oct 10 16:23:19 2007 +0000
Fix #326531 - fence_xvm using dprintf for debugging can cause corruption
commit 007ebe7da57b42c509c60f3a82e71e7076dd31a2
Author: James Parsons <jparsons@redhat.com>
Date: Tue Oct 9 14:56:15 2007 +0000
Fix for bz299191
commit cbe80ab7633b83a3682f744953aaaeced4ab227f
Author: Patrick Caulfield <pcaulfie@redhat.com>
Date: Fri Oct 5 10:06:24 2007 +0000
Don't overwrite castparam if /it/ is NULL.
Checking castaddr is wrong because it's not possible to affect that
from the userland API.
This is also consistent with the fix for bz#318061 in RHEL5.
This check-in fixes bz#318851 but TPTB will decide whether it gets included
in RHEL4.6
commit 52598fe04e07201e2c84f1f9d5abd9fa581111a2
Author: Lon Hohberger <lhh@redhat.com>
Date: Wed Oct 3 21:04:04 2007 +0000
Fix 310981
-----------------------------------------------------------------------
hooks/post-receive
--
Cluster Project
More information about the Cluster-cvs
mailing list