fsgqa [Mon, 18 Nov 2002 05:13:03 +0000 (05:13 +0000)]
Add in test 070 - EA writing via fsstress.
Add in test 070 - EA writing via fsstress. Test case that currently
causes a panic.
Nathan Scott [Fri, 15 Nov 2002 05:21:22 +0000 (05:21 +0000)]
Allow timestamp display to be configured on/off instead of always off.
Allow timestamp display to be configured on/off instead of always off.
Nathan Scott [Thu, 14 Nov 2002 20:10:01 +0000 (20:10 +0000)]
Update entries for hosts troppo and goldfish.
fsgqa [Thu, 14 Nov 2002 03:39:42 +0000 (03:39 +0000)]
Very minor fixups.
Very minor fixups.
fsgqa [Tue, 12 Nov 2002 09:34:57 +0000 (09:34 +0000)]
Test for quota support the same way the user tools do.
Nathan Scott [Mon, 11 Nov 2002 02:48:59 +0000 (02:48 +0000)]
two is an interesting dbench case, discrepency between 1 and 10 is quite
large atm.
fsgqa [Mon, 11 Nov 2002 02:42:14 +0000 (02:42 +0000)]
unzip the tarfile in the tar measure - too much wasted space otehrwise.
fsgqa [Mon, 11 Nov 2002 00:25:59 +0000 (00:25 +0000)]
fix a typo - missing closing quote.
fsgqa [Sun, 10 Nov 2002 23:32:19 +0000 (23:32 +0000)]
Minor QA/bench updates.
echo mkfs & mount options into the output.
fsgqa [Sun, 10 Nov 2002 20:40:00 +0000 (20:40 +0000)]
Very minor tidyups in the report headers for bonnie runs.
Very minor tidyups in the report headers for bonnie runs.
fsgqa [Sat, 9 Nov 2002 20:13:42 +0000 (20:13 +0000)]
Fix up stderr handling in bench runs.
Send stderr to the $FULL file, not inline with stdout.
fsgqa [Sat, 9 Nov 2002 05:49:41 +0000 (05:49 +0000)]
Small updates to dbench test runs, and integrate some bonnie++ runs too.
Add a header showing mount and mkfs options.
fsgqa [Thu, 7 Nov 2002 05:42:19 +0000 (05:42 +0000)]
Crank up dbench, now that everything seems to be quite stable under load.
We now do 1, 10, 50, and 100 client runs each night.
Make sure the path back to the scripts directory is available in the
environment for scripts to get at if need be.
Christoph Hellwig [Mon, 4 Nov 2002 16:23:02 +0000 (16:23 +0000)]
Add hch's Munich test config
Add defaults for lab343
Nathan Scott [Thu, 10 Oct 2002 19:47:44 +0000 (19:47 +0000)]
Fix up minor build issues in some xfstests sources.
Fix build after recent xfsprogs header changes.
Nathan Scott [Sun, 6 Oct 2002 12:18:43 +0000 (12:18 +0000)]
Allow a specific benchmark run to be requested, rather than all of them
(that is still the default though).
Nathan Scott [Fri, 20 Sep 2002 07:47:55 +0000 (07:47 +0000)]
Refine QA testing scripts.
Don't be so verbose on each test.
Nathan Scott [Fri, 20 Sep 2002 01:35:55 +0000 (01:35 +0000)]
Additional refinements of auto-bench stuff.
Fix up producing the results files.
Nathan Scott [Thu, 19 Sep 2002 07:26:17 +0000 (07:26 +0000)]
Further work on a generic benchmarking framework.
make important argument mandatory, do a better job of collecting results.
fsgqa [Wed, 18 Sep 2002 08:15:56 +0000 (08:15 +0000)]
also report the size of the tar file we're untarring.
fsgqa [Wed, 18 Sep 2002 08:08:03 +0000 (08:08 +0000)]
remove shell diagnostics which was accidentally included.
fsgqa [Wed, 18 Sep 2002 08:03:37 +0000 (08:03 +0000)]
trivial updates to get more bits to work together.
Nathan Scott [Wed, 18 Sep 2002 03:28:20 +0000 (03:28 +0000)]
First cut at a framework for running benchmarks along with all of the
functional tests we have. Aim is to get wind of performance regressions
as changes are checked in, rather than at some arbitrary point down the
track. Still a bit of work to do wrt this framework though.
Steve Lord [Thu, 5 Sep 2002 20:12:34 +0000 (20:12 +0000)]
update my test config
fsgqa [Wed, 4 Sep 2002 11:14:37 +0000 (11:14 +0000)]
several fixes, should now work properly. dont go too crazy with big
files cos the scratch device may be of limited size.
fsgqa [Wed, 4 Sep 2002 10:58:49 +0000 (10:58 +0000)]
bit more cleanup here and there - don't toss out _all_ the error (stderr)
messages - some of em might help diagnose problems.
fsgqa [Wed, 4 Sep 2002 07:44:40 +0000 (07:44 +0000)]
exercises O_APPEND write functionality using a test case which was once
known to be broken.
fsgqa [Wed, 4 Sep 2002 07:00:33 +0000 (07:00 +0000)]
add in some contributed O_APPEND test helper programs.
Nathan Scott [Tue, 3 Sep 2002 23:15:38 +0000 (23:15 +0000)]
several fixes from Greg. (btw, keeping width to 80 chars for readability).
fsgqa [Tue, 3 Sep 2002 06:34:06 +0000 (06:34 +0000)]
revert to using the pid for tmp file names - mktemp -d gives an error on
early (6.2) redhat versions at least.
Nathan Scott [Tue, 3 Sep 2002 01:17:30 +0000 (01:17 +0000)]
Do a better job of cleaning up after autoconf.
Nathan Scott [Mon, 2 Sep 2002 06:04:01 +0000 (06:04 +0000)]
A LVM snapshot regression test by Greg Freemyer.
fsgqa [Mon, 2 Sep 2002 04:06:19 +0000 (04:06 +0000)]
add missing end-of-line to log parameters setting.
fsgqa [Mon, 2 Sep 2002 03:45:25 +0000 (03:45 +0000)]
really fix so this can still function when MKFS_OPTIONS include v2 logs.
still just a workaround, real fix for v2 logs is much more involved.
Nathan Scott [Mon, 2 Sep 2002 00:27:36 +0000 (00:27 +0000)]
update the _notrun message for v2 logs.
fsgqa [Sun, 1 Sep 2002 11:48:05 +0000 (11:48 +0000)]
simple change to make this work for v2 logs - should be revisited. loggen
tool needs real v2 log support...
Nathan Scott [Sun, 1 Sep 2002 11:11:13 +0000 (11:11 +0000)]
changes for supporting the version 2 log format.
Eric Sandeen [Mon, 19 Aug 2002 21:38:12 +0000 (21:38 +0000)]
Deal with the new "(NULLFSINO)" that repair might print
Nathan Scott [Thu, 8 Aug 2002 06:35:44 +0000 (06:35 +0000)]
arch.h is gone, dmapi headers moved around a little.
Eric Sandeen [Thu, 1 Aug 2002 15:04:44 +0000 (15:04 +0000)]
Whoops, use a fixed 4k for alloc's "bsize" parameter, seems
that this is independent of the fs block size.
Nathan Straz [Wed, 31 Jul 2002 22:21:07 +0000 (22:21 +0000)]
Fix more filters that will fail on xscsi names
Nathan Straz [Wed, 31 Jul 2002 21:58:36 +0000 (21:58 +0000)]
Fix regexes in _filter_mkfs to understand xscsi device names.
Eric Sandeen [Wed, 31 Jul 2002 21:32:56 +0000 (21:32 +0000)]
Make this test work for fs block size < page size
Dean Roehrich [Tue, 30 Jul 2002 15:22:14 +0000 (15:22 +0000)]
increase sizes of sidbuf[] and tokenbuf[].
Nathan Scott [Tue, 30 Jul 2002 02:23:42 +0000 (02:23 +0000)]
some additional diagnostics were added to invutil with recent irix merge.
Nathan Scott [Tue, 30 Jul 2002 01:24:34 +0000 (01:24 +0000)]
there's an extra message on matching with invutil now - add it in.
Steve Lord [Fri, 12 Jul 2002 21:24:23 +0000 (21:24 +0000)]
white space cleanup
fsgqa [Fri, 21 Jun 2002 04:11:38 +0000 (04:11 +0000)]
ensure testfile doesn't exist before we start.
Nathan Scott [Fri, 21 Jun 2002 00:14:20 +0000 (00:14 +0000)]
small fixups - might help Steves problems with this test failing.
Nathan Scott [Thu, 20 Jun 2002 04:51:39 +0000 (04:51 +0000)]
minor reorg of some bits of code to match kernel changes, no biggie.
Nathan Scott [Wed, 19 Jun 2002 02:15:08 +0000 (02:15 +0000)]
remove quota header, no point keeping syncd anymore.
Steve Lord [Tue, 18 Jun 2002 20:32:20 +0000 (20:32 +0000)]
Fix test for v2 striped logs
Steve Lord [Tue, 18 Jun 2002 11:55:12 +0000 (11:55 +0000)]
make this run at a reasonable speed - say a few hundred times faster
Nathan Scott [Tue, 18 Jun 2002 04:37:18 +0000 (04:37 +0000)]
add fstest.c and mmapcat.c - keep a copy of these here so they don't get
lost and so that I can write some QA tests which make use of em. authors
are listed at the head of each file, minor mods made to fix warnings, etc.
Dean Roehrich [Tue, 11 Jun 2002 18:02:16 +0000 (18:02 +0000)]
No Message Supplied
Dean Roehrich [Tue, 11 Jun 2002 18:00:30 +0000 (18:00 +0000)]
remove handle_write_invis/handle_read_invis
Dean Roehrich [Tue, 11 Jun 2002 17:59:14 +0000 (17:59 +0000)]
No Message Supplied
Dean Roehrich [Tue, 11 Jun 2002 17:54:03 +0000 (17:54 +0000)]
No Message Supplied
Dean Roehrich [Thu, 6 Jun 2002 18:51:10 +0000 (18:51 +0000)]
No Message Supplied
Nathan Scott [Thu, 6 Jun 2002 02:04:29 +0000 (02:04 +0000)]
remove no longer used file xfs_rtbit.c.
Nathan Scott [Thu, 6 Jun 2002 01:55:15 +0000 (01:55 +0000)]
remove no longer needed xfs_bit.c file.
Eric Sandeen [Wed, 5 Jun 2002 20:14:16 +0000 (20:14 +0000)]
Make scripts executable again... sigh...
fsgqa [Wed, 5 Jun 2002 02:24:23 +0000 (02:24 +0000)]
loop may not always be a module - if not, modprobe will barf, but we should still proceed.
Eric Sandeen [Tue, 4 Jun 2002 23:07:56 +0000 (23:07 +0000)]
Update copyright dates (again)
Eric Sandeen [Tue, 4 Jun 2002 22:53:09 +0000 (22:53 +0000)]
Undoes mod: xfs-cmds:slinx:
120772a
Undo xfs-cmds:slinx:
120772a, inadvertently whacked a previous mod.
Eric Sandeen [Tue, 4 Jun 2002 17:58:21 +0000 (17:58 +0000)]
Update copyright dates
fsgqa [Tue, 4 Jun 2002 06:52:25 +0000 (06:52 +0000)]
allow for non-64K allocations on the xfs IO path, which can cause quota
calcs to come out differently (closer to the block limit for example).
also helps the multiple blocksize situation too for this test.
Steve Lord [Sun, 2 Jun 2002 13:20:48 +0000 (13:20 +0000)]
probe for loop module when looking for loopfs support
Nathan Scott [Thu, 30 May 2002 05:53:15 +0000 (05:53 +0000)]
fix a typo in pkg test.
Nathan Scott [Thu, 30 May 2002 00:15:39 +0000 (00:15 +0000)]
use local, renamed quota header.
John Kihonge [Tue, 28 May 2002 16:20:19 +0000 (16:20 +0000)]
Adding dmnt2 and defining its parameters.
Tim Shimmin [Wed, 22 May 2002 02:21:06 +0000 (02:21 +0000)]
Delete the notrun call.
067 no longer panics for me.
Nathan Scott [Wed, 22 May 2002 01:53:39 +0000 (01:53 +0000)]
allow a couple more environment variables to be set before running QA.
fsgqa [Tue, 21 May 2002 06:25:36 +0000 (06:25 +0000)]
fix typo in location of xfs_support directory.
fsgqa [Tue, 21 May 2002 05:27:09 +0000 (05:27 +0000)]
use our own copy of stat(1) cos we know it'll be there & it has a set
output format.
Nathan Scott [Tue, 21 May 2002 04:28:30 +0000 (04:28 +0000)]
changes to fold xfs_support code back into XFS (the other sgi projects
using this will be / have been fixed). reduces our footprint on the
base kernel. NB: xfs_support.o as a distinct module no longer exists.
Nathan Scott [Thu, 16 May 2002 23:54:04 +0000 (23:54 +0000)]
update config for couple of hosts.
fsgqa [Tue, 14 May 2002 00:35:23 +0000 (00:35 +0000)]
make sure we can handle both the fhs and non-fhs compliant inventory
directory locations in this test.
Nathan Scott [Fri, 10 May 2002 07:56:20 +0000 (07:56 +0000)]
cosmetic change to the status email subject line.
fsgqa [Fri, 10 May 2002 05:56:02 +0000 (05:56 +0000)]
put the diff output in seq.full, where its more useful for debugging.
fsgqa [Fri, 10 May 2002 03:40:10 +0000 (03:40 +0000)]
fix a typo, INODE -> INO.
fsgqa [Fri, 10 May 2002 02:50:17 +0000 (02:50 +0000)]
sanitize the notrun messages.
fsgqa [Fri, 10 May 2002 02:43:53 +0000 (02:43 +0000)]
express the logsize at mkfs time in units of bytes, not blocks; allows the
test to work for non-4K blocksizes.
fsgqa [Fri, 10 May 2002 02:33:23 +0000 (02:33 +0000)]
make test work for variable blocksizes.
fsgqa [Thu, 9 May 2002 23:34:47 +0000 (23:34 +0000)]
be smarter about figuring out blocksize.
fsgqa [Thu, 9 May 2002 08:40:09 +0000 (08:40 +0000)]
clean up the diagnostic message a bit.
fsgqa [Thu, 9 May 2002 07:24:12 +0000 (07:24 +0000)]
updates for supporting blocksizes smaller than the pagesize.
fsgqa [Thu, 9 May 2002 06:09:58 +0000 (06:09 +0000)]
fix typo in header comment.
fsgqa [Thu, 9 May 2002 06:03:32 +0000 (06:03 +0000)]
fix to work with multiple block sizes - we were running mkfs with a log
size which was dependent on block size, and became too small for 1024k
filesystems.
Tim Shimmin [Thu, 9 May 2002 05:26:20 +0000 (05:26 +0000)]
move some acl funcs into common.attr
fsgqa [Wed, 8 May 2002 23:44:54 +0000 (23:44 +0000)]
Fix a typo.
fsgqa [Wed, 8 May 2002 05:45:22 +0000 (05:45 +0000)]
fix bruce's remote tape setup, remove couple of old QA machines now used
for other duties.
Tim Shimmin [Wed, 8 May 2002 04:29:55 +0000 (04:29 +0000)]
Test for handling of large ACLs with the limit and beyond
number of ACEs.
fsgqa [Tue, 7 May 2002 23:21:16 +0000 (23:21 +0000)]
fix typo, bozo -- TEST_MNT -> TEST_DIR; extra notrun checks now work.
fsgqa [Tue, 7 May 2002 23:11:58 +0000 (23:11 +0000)]
clarify the notrun message a little.
fsgqa [Tue, 7 May 2002 21:33:35 +0000 (21:33 +0000)]
don't unilaterally notrun this test -- check whether libc is busted, and
run the test if not.
fsgqa [Tue, 7 May 2002 05:58:13 +0000 (05:58 +0000)]
get some default MKFS_OPTIONS and MOUNT_OPTIONS happening so its more
obvious on how to use these variables.
Nathan Scott [Mon, 6 May 2002 21:52:50 +0000 (21:52 +0000)]
only notrun on Linux, update notrun message a little.
fsgqa [Mon, 6 May 2002 21:38:49 +0000 (21:38 +0000)]
got a bit carried away with use of $here, broke 063, now fixed.
Dean Roehrich [Mon, 6 May 2002 16:40:30 +0000 (16:40 +0000)]
Allow print_event to mimick minimal hsm.