Whamcloud - gitweb
thantry [Fri, 27 Dec 2002 08:14:06 +0000 (08:14 +0000)]
Merging from HEAD onto b_intel
This is an intermediate step. Need to update further tomorrow.
adilger [Thu, 26 Dec 2002 10:29:26 +0000 (10:29 +0000)]
Damn CVS ate my nice long comment...
Changes related to "fast" LOV stripe MD storage on the MDS:
- Store small MD data inside the inode to avoid an extra seek. We don't need
the super-hack of overloading the inode delete method like we did before,
because we don't store the file size on the MDS any more, which means we
don't truncate the fake blocks from the file on unlink.
- Don't send size updates to MDS to avoid needless RPC and also avoid messing
with MD data in inode per above.
- Reduce MDS MD size by removing redundant fields and reducing the size of
others, so that it is more likely we can fit this data into the inode. We
can use the lmm_magic field to determine different stripe types, which will
probably need different data fields anyways.
This changes the ioctl interface, so I probably should create a lov_user_md
struct again, now that lov_setstripe() iocontrol exists.
- Pack MDS MD data only up to the last-allocated OST index, to make it smaller.
- Change lov_create() OST selection code (per bug 470). This wasn't strictly
related to this change, but it was in my tree and overlapping and I didn't
want to re-do the changes. It seems to work, and it will help us handle
errors that happen just at the wrong time (and also OSTs with no space/objs).
- Check MD input data more carefully.
pschwan [Tue, 24 Dec 2002 05:14:50 +0000 (05:14 +0000)]
lfind --obd aborted at the first file that didn't contain that OST. fixed.
pschwan [Fri, 20 Dec 2002 19:07:19 +0000 (19:07 +0000)]
merge HEAD into b_md
adilger [Fri, 20 Dec 2002 13:24:57 +0000 (13:24 +0000)]
Lots of things here:
- hopefully fix for bug 519 (don't write stripe MD to the MDS each open)
- cleanup of delayed-file-create code
- proper abstraction of user-supplied stripe data into LOV/OSC
- fix bogus LBUG in ll_file_release when lsm is NULL (just means that
no objects were allocated for this file yet)
- make lstripe help less verbose
- remove old lovstripe program, it is obsoleted by lstripe.
Sadly, lstripe does not yet work, although acceptance-small.sh runs
fine so I'm checking this in so I can debug lstripe under UML.
rread [Thu, 19 Dec 2002 06:48:37 +0000 (06:48 +0000)]
* add fstype support
* update scripts for last minute interface change
rread [Wed, 18 Dec 2002 23:58:44 +0000 (23:58 +0000)]
* update HEAD to b_config
nfshp [Tue, 17 Dec 2002 13:19:40 +0000 (13:19 +0000)]
merge HEAD to b_io branch. (seems quite a large changes from
b_md and b_recover)
cvs2svn [Mon, 16 Dec 2002 17:01:18 +0000 (17:01 +0000)]
This commit was manufactured by cvs2svn to create branch 'unlabeled-1.4.6'.
braam [Mon, 16 Dec 2002 17:01:10 +0000 (17:01 +0000)]
- Phil's intent fix for middle of symlink lookups.
- Update version
shaver [Mon, 16 Dec 2002 02:04:25 +0000 (02:04 +0000)]
Land b_recovery (other than WIP lock-replay work).
- b=512,others: fully reactivate OST imports after reconnection.
- b=513(partial): make sure client sees out -ENOTCONN from mds_handle
- b=515: more graceful error handling for truncating on dead OST
- b=474: don't error out just because a file is striped across a dead OST; only
if we actually try to access it
- b=410: fix garbage sizes when stat(2)ing a file that includes a stripe on a
dead OST
- print console diagnostic for completion-timeout client recovery
- MSG_RESENT for requests that were retransmitted
braam [Mon, 16 Dec 2002 01:17:16 +0000 (01:17 +0000)]
- add more patch series files. We have now obsoleted most of the
patches directory with the new regime.
yanhao [Sun, 15 Dec 2002 16:16:01 +0000 (16:16 +0000)]
yanhao: change to new lmc syntax
hitao [Sun, 15 Dec 2002 07:40:12 +0000 (07:40 +0000)]
file lin-2.5.44.txt was initially added on branch b_lin25.
hitao [Sun, 15 Dec 2002 07:40:11 +0000 (07:40 +0000)]
file lin-2.5.44 was initially added on branch b_lin25.
hitao [Sun, 15 Dec 2002 07:40:10 +0000 (07:40 +0000)]
file lin-2.5.44.pc was initially added on branch b_lin25.
hitao [Sun, 15 Dec 2002 07:40:09 +0000 (07:40 +0000)]
file lin-2.5.44.patch was initially added on branch b_lin25.
shaver [Sat, 14 Dec 2002 23:09:31 +0000 (23:09 +0000)]
Updating from HEAD, in preparation for landing.
pschwan [Sat, 14 Dec 2002 19:40:17 +0000 (19:40 +0000)]
land b_md onto HEAD. the highlights:
- fstat didn't correctly synchronize attributes (399)
- server must handle lock cancellation during blocking AST prep (487)
- bulk descriptors were free()d too soon (511)
- fix paths in lconf, which would load incorrect modules (451, 507)
- fix confusing lconf 'host not found' error message (386)
braam [Sat, 14 Dec 2002 13:56:10 +0000 (13:56 +0000)]
- this time added with -ko
braam [Sat, 14 Dec 2002 13:55:13 +0000 (13:55 +0000)]
- redo the addition of the patch, forgot -ko
braam [Sat, 14 Dec 2002 13:43:37 +0000 (13:43 +0000)]
add patch fixups for Red Hat 2.4.18-18.8.0 kernel
thantry [Sat, 14 Dec 2002 02:49:44 +0000 (02:49 +0000)]
file llparser.pm was initially added on branch b_intel.
shorthair [Thu, 12 Dec 2002 14:11:26 +0000 (14:11 +0000)]
file lustre-2.5-from811-812.patch was initially added on branch b_lin25.
shorthair [Thu, 12 Dec 2002 14:11:25 +0000 (14:11 +0000)]
file lustre-2.5-from809-812.patch was initially added on branch b_lin25.
nfshp [Thu, 12 Dec 2002 12:53:09 +0000 (12:53 +0000)]
file iod.c was initially added on branch b_io.
pschwan [Wed, 11 Dec 2002 16:21:47 +0000 (16:21 +0000)]
backout #2
narasimd [Wed, 11 Dec 2002 06:53:25 +0000 (06:53 +0000)]
Merging to head, BUG#405
yanhao [Wed, 11 Dec 2002 01:07:56 +0000 (01:07 +0000)]
yanhao: check in again for bug 461 wrong path in lconf
pschwan [Tue, 10 Dec 2002 16:36:30 +0000 (16:36 +0000)]
Back out yanhao's configure.in change -- this broke the build, and belongs
on a branch, please be more careful.
yanhao [Tue, 10 Dec 2002 15:50:38 +0000 (15:50 +0000)]
yanhao: fix the bug 461 wrong paths in lconf
fix the bug 386 lconf error message confusing
yanhao [Tue, 10 Dec 2002 15:50:35 +0000 (15:50 +0000)]
file lconf.in was initially added on branch b_md.
amrutjoshi [Tue, 10 Dec 2002 11:39:08 +0000 (11:39 +0000)]
Needed more fixes to path
amrutjoshi [Tue, 10 Dec 2002 11:20:56 +0000 (11:20 +0000)]
Fixed the file name
adilger [Tue, 10 Dec 2002 08:25:57 +0000 (08:25 +0000)]
file lustre_fsfilt.h was initially added on branch b_md.
adilger [Tue, 10 Dec 2002 08:25:56 +0000 (08:25 +0000)]
file fsfilt_extN.c was initially added on branch b_md.
adilger [Tue, 10 Dec 2002 08:25:55 +0000 (08:25 +0000)]
file fsfilt_ext3.c was initially added on branch b_md.
adilger [Tue, 10 Dec 2002 08:25:54 +0000 (08:25 +0000)]
file fsfilt.c was initially added on branch b_md.
braam [Tue, 10 Dec 2002 08:16:01 +0000 (08:16 +0000)]
fix the pathnames in the patch
eeb [Mon, 9 Dec 2002 21:50:22 +0000 (21:50 +0000)]
file checkstat.c was initially added on branch b_md.
braam [Sun, 8 Dec 2002 10:01:53 +0000 (10:01 +0000)]
- remove the README file Zach wrote. I placed it on the project wiki
- add two initial patches for 2.5 as in use with 2.5.44
- these require breaking up in components etc as Zach suggests
yanhao [Sun, 8 Dec 2002 06:15:07 +0000 (06:15 +0000)]
*** empty log message ***
zab [Sat, 7 Dec 2002 00:21:14 +0000 (00:21 +0000)]
- bring in akpm's patch management scripts and a first cut at
patch-2.4.18-14 and patch-2.4.18-um against 2.4.18-17.8.0. the README
should hopefully make it clear how these are intended to be used. email
to follow.
pschwan [Fri, 6 Dec 2002 10:19:14 +0000 (10:19 +0000)]
Update the IA-64 patch
cvs2svn [Fri, 6 Dec 2002 10:11:55 +0000 (10:11 +0000)]
This commit was manufactured by cvs2svn to create branch 'unlabeled-1.4.6'.
pschwan [Fri, 6 Dec 2002 10:11:13 +0000 (10:11 +0000)]
merge b_md into HEAD:
- lfind did not preserve OST order in output (443)
- symlinks cause hung clients, incorrect data (439)
- stop dereferencing request after dropping refcount (457)
- don't LASSERT(spin_is_locked) on non-SMP (455)
- fixes for many rename() bugs
morrone [Fri, 6 Dec 2002 01:26:50 +0000 (01:26 +0000)]
Removed all files in this tree, I goofed on the import
morrone [Fri, 6 Dec 2002 01:15:13 +0000 (01:15 +0000)]
This commit was generated by cvs2svn to compensate for changes in r5101,
which included commits to RCS files with non-trunk default branches.
morrone [Fri, 6 Dec 2002 01:15:13 +0000 (01:15 +0000)]
First checkin of auto-test buffalo scripts
braam [Thu, 5 Dec 2002 14:48:19 +0000 (14:48 +0000)]
file multifstat.c was initially added on branch b_md.
eeb [Thu, 5 Dec 2002 14:22:12 +0000 (14:22 +0000)]
file patch-2.4.18-hp1_pnnl18.2.8qsnet was initially added on branch b_md.
pschwan [Thu, 5 Dec 2002 08:42:29 +0000 (08:42 +0000)]
land b_md onto HEAD:
- LRU counters were broken, causing constant lock purge (433, 432)
- the first part of the fixes for the ldlm_handle2lock race (419)
- cleanup unused ll_inode_info field (442)
- keep ioctl semaphore from serializing test_getattr/brw
braam [Thu, 5 Dec 2002 07:54:26 +0000 (07:54 +0000)]
file local2-hack.xml was initially added on branch b_md.
braam [Thu, 5 Dec 2002 07:54:25 +0000 (07:54 +0000)]
file llmount2-hackcleanup.sh was initially added on branch b_md.
braam [Thu, 5 Dec 2002 07:54:24 +0000 (07:54 +0000)]
file sanityN.sh was initially added on branch b_md.
shaver [Thu, 5 Dec 2002 00:31:32 +0000 (00:31 +0000)]
Landing b_recovery.
b=441: garbage on read from stripes with failed OSTs.
b=438: mark OSCs as active before reconnecting during recovery.
b=403: lov_enqueue and lov_cancel need to handle inactive OSTs
sravi [Wed, 4 Dec 2002 14:13:21 +0000 (14:13 +0000)]
file genXml was initially added on branch b_ravi_devel.
pschwan [Wed, 4 Dec 2002 03:15:15 +0000 (03:15 +0000)]
b=442
Remove lli_intent_lock_handle that none of us seem to have an explanation for.
thantry [Wed, 4 Dec 2002 00:14:12 +0000 (00:14 +0000)]
Changes made to LProcFS after internal review.
Changes reduce the code size, remove redundant functions
statfs calls added for OSC/MDC/LOV in ll_pack.c (??)
Change set provided by Rumi.
adilger [Mon, 2 Dec 2002 18:06:14 +0000 (18:06 +0000)]
Fix typo in acceptance-small.sh
adilger [Mon, 2 Dec 2002 08:43:34 +0000 (08:43 +0000)]
Update for new release
cvs2svn [Mon, 2 Dec 2002 08:41:04 +0000 (08:41 +0000)]
This commit was manufactured by cvs2svn to create branch 'unlabeled-1.4.6'.
adilger [Mon, 2 Dec 2002 08:39:50 +0000 (08:39 +0000)]
Merge of b_md to HEAD:
- fixes of many metadata bugs:
- multiple client file opens
- missing/extra intent releases
- symlink fixes
- OST/MDS server threads clobbering each other
- dbench 2 works mostly
- many others too numerous to mention here (ChangeLog is a good record)
adilger [Sun, 1 Dec 2002 22:39:02 +0000 (22:39 +0000)]
Remove this .xml file yet one more time.
adilger [Sun, 1 Dec 2002 22:36:51 +0000 (22:36 +0000)]
Ignore all generated .xml files.
adilger [Sun, 1 Dec 2002 22:32:10 +0000 (22:32 +0000)]
Fix ba-mount.sh so that it saves the config in $config instead of config.
Remove explicit port number.
adilger [Sun, 1 Dec 2002 22:21:44 +0000 (22:21 +0000)]
Fix compile warnings (hopefully nothing else is missing...)
shaver [Sun, 1 Dec 2002 03:47:10 +0000 (03:47 +0000)]
Landing b_lock_replay so that Phil can use my ldlm iterators and whatnot for his
upcoming locking work.
Highlights:
- lock replay infrastructure (needs much more testing, but didn't regress
anything outside recovery)
- b=421: ldlm iterators
- b=348: imports now have service levels, replacing connections' c_level
- replace c_delayed_head with imp_delayed_list
- split imp_request_list into imp_replay_list for retained requests and
imp_sending_list for inflight reqs
- as a side-effect, clean up rq_refcount story materially
- client-side recovery is now dispatched via a per-import handler function,
for better layering and modularity
- wire imports up to recovery before attempting mounts, for better handling of
mount-time failures
shaver [Thu, 28 Nov 2002 20:21:24 +0000 (20:21 +0000)]
Landing of b_recovery (at last).
Highlights:
- b=324: MDS recovery must replay transactions in strict transno sequence
- b=325: getattr after OST failure returns -EIO
- b=326: unlink after OST failure returns -EIO
- b=400: new client can't join cluster after OST failure
- b=403: multi-client access failure when OST fails
- b=410: After an OST failure, lfind incorrectly displays file information
- b=417: Freeing unreplayable requests twice (aed's fix from b_md)
- b=402: (partial) give error for lstripe request that exceeds configured OSTs
- much better support for reconnecting to MDS after network partition
(still some lock-repeating issues to be resolved for some requests)
- better support for connecting to multiple MDSes on one host (xid and
transno and request_list are all per-import now)
- track disconnecting clients in last_rcvd, for more reliable recovery
- also, sync last_rcvd after connect/disconnect
- reduced syslog/CERROR output for recovery (hi, Terry!)
- server (DLM) timeout is half the system-wide timeout, to avoid cascading
failure in the face of a dead client
- don't wait for recovery to finish in order to send disconnect messages
- removal of c_dying_head
- don't wait for timeout to trigger recovery after ptl_send_rpc error
- strict MDS transno ordering via mds_transno_sem (non-optimal, but correct)
- many !handle -> IS_ERR(handle) fixes around mds_fs_start callers.
- turn on client-eviction for bulk-timeouts in OST and MDS
braam [Tue, 26 Nov 2002 20:59:42 +0000 (20:59 +0000)]
file llmountcleanup2-hack.sh was initially added on branch b_md.
braam [Tue, 26 Nov 2002 20:59:41 +0000 (20:59 +0000)]
file llmount2-hack.sh was initially added on branch b_md.
braam [Mon, 25 Nov 2002 22:20:33 +0000 (22:20 +0000)]
file setuid.c was initially added on branch b_md.
adilger [Mon, 25 Nov 2002 20:26:02 +0000 (20:26 +0000)]
Whitespace.
amrutjoshi [Mon, 25 Nov 2002 15:08:38 +0000 (15:08 +0000)]
Added sanity.sh & rundbench tests to Makefile.am EXTRA_DIST
pschwan [Mon, 25 Nov 2002 09:17:17 +0000 (09:17 +0000)]
merge b_md onto HEAD. as best as I can remember:
- fix dentry->d_it clobbering
- credentials weren't being shipped for readdir/getattr operations
- remove invalid assertions triggered during some concurrent MD updates
- many, many updates to llanalyse
adilger [Sun, 24 Nov 2002 19:45:00 +0000 (19:45 +0000)]
Don't dump bogus locks, we can deref null pointers therein.
adilger [Sun, 24 Nov 2002 18:26:58 +0000 (18:26 +0000)]
Include the portal request/reply number so we know the source of the bad
export data.
adilger [Sat, 23 Nov 2002 02:24:12 +0000 (02:24 +0000)]
update
adilger [Sat, 23 Nov 2002 02:23:10 +0000 (02:23 +0000)]
Fix the many client simultaneous startup bug (b=392).
We just ignore exports with NULL mcd, since they are just being set up.
The clients have a semaphore to prevent multiple setups there. We could
alternately have a semaphore on the MDS instead of the current spinlock.
pschwan [Fri, 22 Nov 2002 12:37:06 +0000 (12:37 +0000)]
Print more information when we get a callback with a bogus export.
Also, try to see if we can proceed--maybe the rest of the packet is OK.
rread [Fri, 22 Nov 2002 01:33:01 +0000 (01:33 +0000)]
use correct name for toenal_ni
rread [Fri, 22 Nov 2002 00:36:06 +0000 (00:36 +0000)]
* add cleanup code for toenal
adilger [Thu, 21 Nov 2002 18:05:17 +0000 (18:05 +0000)]
Change over to new echo client like runregression-net.sh
amrutjoshi [Thu, 21 Nov 2002 10:57:26 +0000 (10:57 +0000)]
file version_tag.pl was initially added on branch b_phil.
adilger [Thu, 21 Nov 2002 08:06:48 +0000 (08:06 +0000)]
Remove generated file.
adilger [Thu, 21 Nov 2002 07:18:56 +0000 (07:18 +0000)]
Remove bogus LBUG and error return code. This was triggered by a non-root
user trying to create a file and getting -EACCES from lookup_one_len().
rread [Thu, 21 Nov 2002 00:34:39 +0000 (00:34 +0000)]
teach runregression-net.sh about echo_client
adilger [Wed, 20 Nov 2002 23:27:23 +0000 (23:27 +0000)]
Omission in checkin.
adilger [Wed, 20 Nov 2002 22:57:51 +0000 (22:57 +0000)]
Fix bad awk in mcr-routed-config.sh, which grabbed every OST UUID instead
of just the one we wanted.
adilger [Wed, 20 Nov 2002 22:56:08 +0000 (22:56 +0000)]
Remove PORT specification, so we get the (correct) default if it changes
(and people haven't copied the fixed port number to dozens of configs
like I did.
adilger [Wed, 20 Nov 2002 20:33:57 +0000 (20:33 +0000)]
Sadly, CVS doesn't handle adding/removing/moving files very well from
branch merges.
sravi [Tue, 19 Nov 2002 18:58:05 +0000 (18:58 +0000)]
file ha_assist2 was initially added on branch b_ravi_devel.
sravi [Tue, 19 Nov 2002 18:48:38 +0000 (18:48 +0000)]
file lutils.py was initially added on branch b_ravi_devel.
pschwan [Mon, 18 Nov 2002 05:28:30 +0000 (05:28 +0000)]
b=191
b=214
b=306
b=385
landing b_brw, which _includes_ all fixes from b_lov.
Roughly, this fixes bugs in 3 general areas:
- OSC was locking one byte too many; fixed.
- several LOV bug fixes, to support adding LOVs and striping over fewer
than the maximum number
- the brw code was rewritten
See the bugs listed above for more details. This will be immediately
tagged as 0.5.17.
pschwan [Sun, 17 Nov 2002 04:50:12 +0000 (04:50 +0000)]
file echo_client.c was initially added on branch b_brw.
adilger [Sat, 16 Nov 2002 08:27:40 +0000 (08:27 +0000)]
Don't dereference last_xid and last_committed in a request we just freed.
adilger [Fri, 15 Nov 2002 20:48:54 +0000 (20:48 +0000)]
Revert buffalization of acceptance-small.sh in HEAD, that lives in ltest.
alokbhagwat [Fri, 15 Nov 2002 15:40:02 +0000 (15:40 +0000)]
Modified to report to buffalo (Buffalized)
adilger [Thu, 14 Nov 2002 23:32:53 +0000 (23:32 +0000)]
Decrement the LOV module use count on connect errors.
thantry [Thu, 14 Nov 2002 19:30:23 +0000 (19:30 +0000)]
Bug Fix for Bug #369