]> sourceware.org Git - lvm2.git/log
lvm2.git
14 years agoMove regex printing code from test to main tree (may use in debug messages).
Alasdair Kergon [Thu, 22 Apr 2010 17:42:38 +0000 (17:42 +0000)]
Move regex printing code from test to main tree (may use in debug messages).
Yet another optimiser fix attempt.

14 years agoThe following tests in the testsuite have race conditions:
Jonathan Earl Brassow [Thu, 22 Apr 2010 15:39:40 +0000 (15:39 +0000)]
The following tests in the testsuite have race conditions:
1) Test that the primary mirror image cannot be removed while
   the mirror set is sync'ing.
2) Test that you cannot start a second mirror up-convert while
   one is already in progress.

The trouble is that if the sync/conversion finishes before the
tests occur, the tests will fail by why of success where there
should have been failure.  This means the sync/conversion must
happen very quickly, but this is possible because the test
mirrors we are creating are so small.

In order to decrease the likelyhood of these test failing (or
more correctly, failing to test the right thing), I've increase
the size of the mirrors.  It will still be remotely possible that
the tests will fail (by way of failing to test the right thing).
If this continues to happen, more involved mechanisms will need
to be put in place.  (Perhaps these will still be created, but
this change should be a remedy until that time.)

14 years agoDon't walk rightmost through NULL pointers.
Alasdair Kergon [Thu, 22 Apr 2010 14:33:14 +0000 (14:33 +0000)]
Don't walk rightmost through NULL pointers.

14 years agoFix rightmost rotation, and use LEFT and RIGHT to make symmetry more obvious.
Alasdair Kergon [Thu, 22 Apr 2010 13:42:34 +0000 (13:42 +0000)]
Fix rightmost rotation, and use LEFT and RIGHT to make symmetry more obvious.

14 years agofix leftmost rotation
Alasdair Kergon [Thu, 22 Apr 2010 13:18:27 +0000 (13:18 +0000)]
fix leftmost rotation

14 years agoisprint
Alasdair Kergon [Thu, 22 Apr 2010 03:42:00 +0000 (03:42 +0000)]
isprint

14 years agoStill not satisfactory...
Alasdair Kergon [Thu, 22 Apr 2010 03:24:24 +0000 (03:24 +0000)]
Still not satisfactory...

14 years agoadd -r to print out closer to original regex format
Alasdair Kergon [Thu, 22 Apr 2010 03:12:01 +0000 (03:12 +0000)]
add -r to print out closer to original regex format

14 years agoDisallow the addition of mirror images while a mirror up-convert
Jonathan Earl Brassow [Wed, 21 Apr 2010 14:04:24 +0000 (14:04 +0000)]
Disallow the addition of mirror images while a mirror up-convert
is already occurring.  The addition of new legs can be retried
once the current conversion is complete.

14 years agoDisallow the primary mirror image from being removed when the
Jonathan Earl Brassow [Wed, 21 Apr 2010 13:55:08 +0000 (13:55 +0000)]
Disallow the primary mirror image from being removed when the
mirror is not in-sync.  This restriction is not extended to
repair operations (i.e. it will not limit what 'lvconvert --repair'
can do).

14 years agoMake matcher_t and parse_t compilable
Zdenek Kabelac [Wed, 21 Apr 2010 08:01:51 +0000 (08:01 +0000)]
Make matcher_t and parse_t compilable

These two old-test/regex utils are usable for testing output of regex
processing at core level. For getting them usable configure need to create
Makefile. This is currently disable by default.

14 years agoAdd a regex optimisation pass for shared character prefixes.
Alasdair Kergon [Tue, 20 Apr 2010 22:31:22 +0000 (22:31 +0000)]
Add a regex optimisation pass for shared character prefixes.

14 years agoRe-enable t-topology-support.sh
Mike Snitzer [Tue, 20 Apr 2010 18:18:59 +0000 (18:18 +0000)]
Re-enable t-topology-support.sh

Reintroduce split teardown (teardown() calls teardown_devs()) because
t-topology-support.sh only needs the teardown_devs() subset of the full
teardown() between each iteration of the topology tests -- in particular
the $TESTDIR must not get removed between each topology test iteration.

prepare_loop() must return if prepare_scsi_debug_dev() already
established $LOOP.

Also fix (and simplify) the unsafe scsi-debug device discovery in
prepare_scsi_debug_dev().

14 years agoRemove -n argument from vgcfgrestore.
Milan Broz [Tue, 20 Apr 2010 18:17:56 +0000 (18:17 +0000)]
Remove -n argument from vgcfgrestore.

This is old lvm1 syntax, because arguments changed
from PVs to VGs compatibility is broken already.

name_ARG is not used in code.

14 years agoAdd device creation to basic nightly test.
Dave Wysochanski [Tue, 20 Apr 2010 15:19:36 +0000 (15:19 +0000)]
Add device creation to basic nightly test.

Ensure we can create devices for use in tests before running the real
tests.  This may in various cases, and may involve machine configuration
rather than a failure in some specific test.
For example, if a test is run with LVM_TEST_DIR=/tmp, selinux is enabled,
and the default security context is set to "<<none>>" for /tmp, all the
tests will fail, unable to create devices, since dmsetup will fail, a
result of machpathcon() returning an error code.

Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
14 years agoAdd -S command to clvmd, so it can restart itself and still
Christine Caulfield [Tue, 20 Apr 2010 14:07:37 +0000 (14:07 +0000)]
Add -S command to clvmd, so it can restart itself and still
preserve exlusive LV locks.

14 years agoAdd dm_bitset_equal to libdevmapper.
Alasdair Kergon [Tue, 20 Apr 2010 13:58:22 +0000 (13:58 +0000)]
Add dm_bitset_equal to libdevmapper.

14 years agotypo in comment
Alasdair Kergon [Tue, 20 Apr 2010 12:18:31 +0000 (12:18 +0000)]
typo in comment

14 years agoMove function up file
Alasdair Kergon [Tue, 20 Apr 2010 12:14:28 +0000 (12:14 +0000)]
Move function up file

14 years agomissed files from last commit
Alasdair Kergon [Tue, 20 Apr 2010 10:58:18 +0000 (10:58 +0000)]
missed files from last commit

14 years agoAdd additional test to start the snapshot merge (which had just failed
Mike Snitzer [Mon, 19 Apr 2010 22:44:42 +0000 (22:44 +0000)]
Add additional test to start the snapshot merge (which had just failed
because an FS was mounted -- origin was still open).

14 years agoAdd dm_bit_and. (ejt)
Alasdair Kergon [Mon, 19 Apr 2010 21:23:01 +0000 (21:23 +0000)]
Add dm_bit_and. (ejt)

14 years agofix last commit
Alasdair Kergon [Mon, 19 Apr 2010 21:10:20 +0000 (21:10 +0000)]
fix last commit

14 years agoSimplify dm_bitset_create. (ejt)
Alasdair Kergon [Mon, 19 Apr 2010 21:08:32 +0000 (21:08 +0000)]
Simplify dm_bitset_create. (ejt)

14 years agoSpeed up dm_bit_get_next with ffs().
Alasdair Kergon [Mon, 19 Apr 2010 17:17:55 +0000 (17:17 +0000)]
Speed up dm_bit_get_next with ffs().

14 years agoChange lvm2app version number from 1 to 2.
Dave Wysochanski [Mon, 19 Apr 2010 15:24:00 +0000 (15:24 +0000)]
Change lvm2app version number from 1 to 2.

This version number change reflects the memory handling change
for string-based pv/vg/lv string based attributes.
In addition, when adding support for tags, I forgot to increase
the version number.

Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
14 years agoUse vg->vgmem to allocate vg/lv/pv string properties instead of dm_malloc/fr
Dave Wysochanski [Mon, 19 Apr 2010 15:22:24 +0000 (15:22 +0000)]
Use vg->vgmem to allocate vg/lv/pv string properties instead of dm_malloc/fr

Everywhere else in the API the caller can rely on lvm2app taking care of
memory allocation and free, so make the 'name' and 'uuid' properties of a
vg/lv/pv use the vg handle to allocate memory.

Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
14 years agoChange daemon lock filename from lvm2_monitor to lvm2-monitor for consistency.
Alasdair Kergon [Thu, 15 Apr 2010 19:15:03 +0000 (19:15 +0000)]
Change daemon lock filename from lvm2_monitor to lvm2-monitor for consistency.

14 years agoInstall symbolic .so links with relative paths between usrlibdir and libdir.
Zdenek Kabelac [Thu, 15 Apr 2010 15:12:20 +0000 (15:12 +0000)]
Install symbolic .so links with relative paths between usrlibdir and libdir.

14 years agoFix double DESTDIR usage for infodir and mandir.
Zdenek Kabelac [Thu, 15 Apr 2010 15:09:35 +0000 (15:09 +0000)]
Fix double DESTDIR usage for infodir and mandir.

14 years agoAdded awk script relpath.awk to calculate relative paths.
Zdenek Kabelac [Thu, 15 Apr 2010 15:06:26 +0000 (15:06 +0000)]
Added awk script relpath.awk to calculate relative paths.

14 years agopost-release
Alasdair Kergon [Wed, 14 Apr 2010 21:47:48 +0000 (21:47 +0000)]
post-release

14 years agotouch file
Alasdair Kergon [Wed, 14 Apr 2010 20:03:46 +0000 (20:03 +0000)]
touch file

14 years agofix broken 'make install' for lvm_dump.sh
Alasdair Kergon [Wed, 14 Apr 2010 20:03:15 +0000 (20:03 +0000)]
fix broken 'make install' for lvm_dump.sh

14 years agopre-release
Alasdair Kergon [Wed, 14 Apr 2010 18:54:37 +0000 (18:54 +0000)]
pre-release

14 years agopre-release
Alasdair Kergon [Wed, 14 Apr 2010 18:53:04 +0000 (18:53 +0000)]
pre-release

14 years ago.
Alasdair Kergon [Wed, 14 Apr 2010 18:10:30 +0000 (18:10 +0000)]
.

14 years agopre-release
Alasdair Kergon [Wed, 14 Apr 2010 17:50:49 +0000 (17:50 +0000)]
pre-release

14 years agoAdd missing readline linkage to lvm2app test.
Dave Wysochanski [Wed, 14 Apr 2010 16:13:34 +0000 (16:13 +0000)]
Add missing readline linkage to lvm2app test.

14 years agoWhat's new.
Petr Rockai [Wed, 14 Apr 2010 13:52:20 +0000 (13:52 +0000)]
What's new.

14 years agoAllow incomplete mirror restore in lvconvert --repair upon insufficient space.
Petr Rockai [Wed, 14 Apr 2010 13:51:58 +0000 (13:51 +0000)]
Allow incomplete mirror restore in lvconvert --repair upon insufficient space.

14 years agoDo not reset position in metadata ring buffer on vgrename and vgcfgrestore.
Peter Rajnoha [Wed, 14 Apr 2010 13:09:16 +0000 (13:09 +0000)]
Do not reset position in metadata ring buffer on vgrename and vgcfgrestore.

We should write metadata into next position in the ring buffer while calling
vgrename and vgcfgrestore. At this code level (_vg_write_raw), we were not able
to determine if this is a rename or not. If yes, then accompanying VG structure
passed here has a new name set, not the old one.

When looking for a location where to put metadata next, we were given a NULL
value because of failed VG name comparison (in _find_vg_rlocn) between the
name in existing metadata and metadata we're just about to write.

This resets the position in the ring buffer, overwriting any existing metadata
(and also incorrectly updates the cache to "orphan" afterwards).

This patch just adds old_name item in struct volume_group that we can check and use
if necessary and detect renames at lower layers as well.

The same applies for vgcfgrestore, but here we're using a special value of
old_name, an empty string, to disable the check with existing metadata totally.

14 years agoAllow VGs with active LVs to be renamed.
Peter Rajnoha [Wed, 14 Apr 2010 13:03:06 +0000 (13:03 +0000)]
Allow VGs with active LVs to be renamed.

14 years agoUse UUIDs instead of names while processing event handlers.
Peter Rajnoha [Wed, 14 Apr 2010 13:01:38 +0000 (13:01 +0000)]
Use UUIDs instead of names while processing event handlers.

Internally, we used DM names instead of UUIDs while processing event
handlers. This caused problems while trying to vgrename a VG with active LVs
where the names are being changed and so the devices were not found then.
The patch also contains a little bit of refactoring, moving "build_dlid" code
found in dev_manager.c to "build_dm_uuid", now in lvm-string.c (so we have
build_dm_uuid and build_dm_name at one place).

14 years agoSIGTERM testing clvmd before SIGKILL-ing it.
Petr Rockai [Wed, 14 Apr 2010 12:04:23 +0000 (12:04 +0000)]
SIGTERM testing clvmd before SIGKILL-ing it.

14 years ago Only pass visible LVs to tools in cmdline VG name/tag expansions without -a
Alasdair Kergon [Wed, 14 Apr 2010 02:19:49 +0000 (02:19 +0000)]
  Only pass visible LVs to tools in cmdline VG name/tag expansions without -a

14 years agoUse typedefs for toollib process_each functions.
Alasdair Kergon [Tue, 13 Apr 2010 23:57:41 +0000 (23:57 +0000)]
Use typedefs for toollib process_each functions.

14 years agoFix teardown in t-pvcreate-operation-md.sh nightly test.
Dave Wysochanski [Tue, 13 Apr 2010 21:42:44 +0000 (21:42 +0000)]
Fix teardown in t-pvcreate-operation-md.sh nightly test.

14 years agoAdd error diagnostic for setenv failure.
Zdenek Kabelac [Tue, 13 Apr 2010 20:54:57 +0000 (20:54 +0000)]
Add error diagnostic for setenv failure.

14 years agoUse C locales and use_mlockall for clvmd.
Zdenek Kabelac [Tue, 13 Apr 2010 19:54:16 +0000 (19:54 +0000)]
Use C locales and use_mlockall for clvmd.

Use same steps for clvmd as for dmeventd - using C locales to avoid reading
large mmaps and use mlockall() for threaded version.

14 years agoUpdate WHATS_NEW
Dave Wysochanski [Tue, 13 Apr 2010 18:18:54 +0000 (18:18 +0000)]
Update WHATS_NEW

14 years agoAdd pv->vg to solidify link between a pv and a vg.
Dave Wysochanski [Tue, 13 Apr 2010 17:26:36 +0000 (17:26 +0000)]
Add pv->vg to solidify link between a pv and a vg.

lvm2app needs a link back to the vg in order to use the vg handle for
memory allocations as well as other things.  This patch adds the field
to struct physical_volume, and sets pv->vg when reading a vg from disk or
extending a vg by using the helper function previously added,
add_pvl_to_vgs().  Moves and renames are handled with separate code
inside move_pv() and vgmerge().  Add pv->vg check to vg_validate().

A NULL value in pv->vg signifies membership in the orphan VG.
Note though in the case of pv_read() on a device with metadatacopies == 0,
more devices may need to be read for an authoritative answer.

Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
14 years agoUse del_pvl_from_vgs() in vgreduce paths.
Dave Wysochanski [Tue, 13 Apr 2010 17:26:20 +0000 (17:26 +0000)]
Use del_pvl_from_vgs() in vgreduce paths.

Somehow these got missed in earlier patches.

Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
14 years agoCall add_pvl_to_vgs() and del_pvl_from_vgs() from more places.
Dave Wysochanski [Tue, 13 Apr 2010 17:26:03 +0000 (17:26 +0000)]
Call add_pvl_to_vgs() and del_pvl_from_vgs() from more places.

Now that we have library functions to add/delete a pv from the vg->pvs
list, call them from everywhere.

Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
14 years agoAdd del_pvl_from_vgs() and move prototypes into metadata-exported.h
Dave Wysochanski [Tue, 13 Apr 2010 17:25:44 +0000 (17:25 +0000)]
Add del_pvl_from_vgs() and move prototypes into metadata-exported.h

Add a delete function to manage the vg->pvs list.

NOTE: It may be possible to do further cleanup to these add/del functions
by passing a 'pv' as input instead of 'pv_list'.  The pv_list is used for
functions which do allocations (lvcreate) while other places in the code
just manage a list of 'pv' (e.g. import functions, vgextend, etc).

Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
14 years agoMove increment of vg->pv_count from import_pool_vg() to import_pool_pvs().
Dave Wysochanski [Tue, 13 Apr 2010 17:25:26 +0000 (17:25 +0000)]
Move increment of vg->pv_count from import_pool_vg() to import_pool_pvs().

Move the increment of vg->pv_count next to the place where we add to
vg->pvs.  It looks safe to do this since the only caller of import_pool_vg()
calls import_pool_pvs() immediately afterward, and there is no way
import_pool_vg() can fail (always returns 1).  However, if there's a
memory allocation failure inside import_pool_pvs(), we will end up with
a different count in vg->pv_count that with the original code.  In any
case, vg->pv_count should be as close to dm_list_size(&vg->pvs) as
possible, as is the case everywhere else in the code.

Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
14 years agoRemove unnecessary parameter from import_pool_pvs().
Dave Wysochanski [Tue, 13 Apr 2010 17:25:13 +0000 (17:25 +0000)]
Remove unnecessary parameter from import_pool_pvs().

The dm_list * parameter is unnecessary since we are passing in 'vg'
and the only caller of import_pool_pvs() passes '&vg->pvs' in the
dm_list * parameter.  Just use vg->pvs directly in the function.

Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
14 years agoArguments to the --mirrorlog option are described in the synopsis as
Jonathan Earl Brassow [Tue, 13 Apr 2010 16:13:08 +0000 (16:13 +0000)]
Arguments to the --mirrorlog option are described in the synopsis as
{core|disk|mirrored}.  However, under the options section they are
described as {core|disk} - even though the 'mirrored' argument is
described.

s/'{core|disk}'/'{core|disk|mirrored}'/

14 years agoMask LCK_HOLD flag in cluster VG locks for compatibility reasons.
Milan Broz [Tue, 13 Apr 2010 14:36:24 +0000 (14:36 +0000)]
Mask LCK_HOLD flag in cluster VG locks for compatibility reasons.

14 years agoAdd check_cluster and check_local targets in addition to just "check".
Petr Rockai [Tue, 13 Apr 2010 13:28:52 +0000 (13:28 +0000)]
Add check_cluster and check_local targets in addition to just "check".

14 years agoProcess all core dumps that exist upon a test failure (this would hopefully
Petr Rockai [Tue, 13 Apr 2010 08:01:53 +0000 (08:01 +0000)]
Process all core dumps that exist upon a test failure (this would hopefully
include eg. clvmd cores, apart from the usual lvm ones).

14 years agoFix a reversed exit status from test harness.
Petr Rockai [Tue, 13 Apr 2010 07:34:19 +0000 (07:34 +0000)]
Fix a reversed exit status from test harness.

14 years agoOptionally disable the verbose repeat of a failed test (export
Petr Rockai [Tue, 13 Apr 2010 07:33:34 +0000 (07:33 +0000)]
Optionally disable the verbose repeat of a failed test (export
LVM_TEST_NOVERBOSE=1).

14 years agoDisable the rather-dangerous prepare_scsi_debug_dev in the testsuite.
Petr Rockai [Tue, 13 Apr 2010 07:02:14 +0000 (07:02 +0000)]
Disable the rather-dangerous prepare_scsi_debug_dev in the testsuite.

14 years agoMake it easy to run just a subset of tests by saying `make check T="regex"`.
Petr Rockai [Tue, 13 Apr 2010 06:45:53 +0000 (06:45 +0000)]
Make it easy to run just a subset of tests by saying `make check T="regex"`.

14 years agoFix t-lvcreate-usage, as lvcreate ... -i1 no longer prints "Redundant stripes
Petr Rockai [Tue, 13 Apr 2010 06:25:08 +0000 (06:25 +0000)]
Fix t-lvcreate-usage, as lvcreate ... -i1 no longer prints "Redundant stripes
argument: default is 1".

14 years agoFor the testsuite, set a default polling interval to 0 in lvm.conf. Cuts down
Petr Rockai [Tue, 13 Apr 2010 06:24:20 +0000 (06:24 +0000)]
For the testsuite, set a default polling interval to 0 in lvm.conf. Cuts down
testing time considerably.

14 years agoAdd --stripes to lvconvert too.
Alasdair Kergon [Tue, 13 Apr 2010 01:54:32 +0000 (01:54 +0000)]
Add --stripes to lvconvert too.

14 years ago Add activation/polling_interval to lvm.conf as --interval default.
Alasdair Kergon [Tue, 13 Apr 2010 01:43:56 +0000 (01:43 +0000)]
 Add activation/polling_interval to lvm.conf as --interval default.

14 years agoMore testsuite helpers for check.sh.
Petr Rockai [Mon, 12 Apr 2010 19:33:58 +0000 (19:33 +0000)]
More testsuite helpers for check.sh.

14 years agoAbsorb t-mirror-lvconvert-usage into t-lvconvert-mirror.
Petr Rockai [Mon, 12 Apr 2010 19:32:58 +0000 (19:32 +0000)]
Absorb t-mirror-lvconvert-usage into t-lvconvert-mirror.

14 years agoExplode t-mirror-basic into a number of smaller tests.
Petr Rockai [Mon, 12 Apr 2010 19:16:24 +0000 (19:16 +0000)]
Explode t-mirror-basic into a number of smaller tests.

14 years agoAdd a new helper script for the testsuite, that makes checking for various
Petr Rockai [Mon, 12 Apr 2010 19:02:59 +0000 (19:02 +0000)]
Add a new helper script for the testsuite, that makes checking for various
(lvm) properties easier and produces nicer output.

14 years agoDo not pollute /tmp with testing byproducts.
Petr Rockai [Mon, 12 Apr 2010 19:00:23 +0000 (19:00 +0000)]
Do not pollute /tmp with testing byproducts.

14 years agoDon't ignore error if resuming any LV fails in resume_lvs.
Alasdair Kergon [Mon, 12 Apr 2010 11:52:53 +0000 (11:52 +0000)]
Don't ignore error if resuming any LV fails in resume_lvs.
Skip closing persistent filter cache file if open failed.

14 years agoUpdate install rules for udev.
Zdenek Kabelac [Fri, 9 Apr 2010 21:44:28 +0000 (21:44 +0000)]
Update install rules for udev.

Fix unwanted modification of $(top_builddir)/make.tmpl.

Using dependency rules to install rules for udev.
There is minor problem, with concurent usage of builddir
and srcdir could lead to missuse of 10-dm.rules which
could be found in VPATH from different builddir.
However current solution uses intermediate target so
the generated 10-dm.rules exists only for short period of time
during make install execution.

14 years agoINSTALL rules updates
Zdenek Kabelac [Fri, 9 Apr 2010 21:42:48 +0000 (21:42 +0000)]
INSTALL rules updates

Patch is inspired by Debian's extra patch.

- removes OWNER & GROUP make vars they are parts of INSTALL command.
- adds INSTALL_PROGRAM for executable, uses $(INSTALL)
- adds INSTALL_DATA for non-executable data, uses ($INSTALL)
- adds INSTALL_WDATA for writable non-executable data, uses ($INSTALL)
- adds configure option --enable-write_install - to support
  installatin of writable files used by distribution
- replaces usage of ifeq @LIB_SUFFIX@ with $(LIB_SUFFIX)
- installs .a files from static builds without executable flag
- installs .a files to $(usrlibdir) instead of $(libdir)
- installs all static binaries to $(staticdir)
- create .so links for devel package in $(usrlibdir) instead of
  $(libdir)
- makes .so and .so.LIB_VERSION files within builddir
- removes VERSIONED_SHLIB and created versioned LIB_SHARED automagicaly
- install LIB_SHARED via install_lib_shared target
- install plugins via install_lib_shared_plugin target
- prints whole 'install' command during installation instead of less
  informative "Installing  $(something) $(somewhere)"
- install multiple man pages with one INSTALL command
- use DISTCLEAN_TARGETS instead of creating multiple distclean targets

14 years agoUse vpath instead of VPATH.
Zdenek Kabelac [Fri, 9 Apr 2010 21:34:25 +0000 (21:34 +0000)]
Use vpath instead of VPATH.

Usage of VPATH makes troubles when used within $(builddir).
Not only source files are being found through VPATH,
but targets as well. (make --debug=v)

Thus if user builds the code in $(srcdir) and also in some $(builddir)
he gets mangled results as some generated files (i.e. .export.sym)
are 'reused' from $(srcdir) instead of $(builddir).

This patch switches to use vpath were we could explicitly name
suffixes that should be looked via vpath - we must take care,
we do not generate files with these suffixes:
.c, .in, .po, .exported_symbols

14 years agoPermit mimage LVs to be striped in lvcreate and lvresize.
Alasdair Kergon [Fri, 9 Apr 2010 01:00:10 +0000 (01:00 +0000)]
Permit mimage LVs to be striped in lvcreate and lvresize.

14 years agoCheck for duplicate paths (pvids) on the commandline of vgcreate.
Dave Wysochanski [Thu, 8 Apr 2010 15:18:35 +0000 (15:18 +0000)]
Check for duplicate paths (pvids) on the commandline of vgcreate.

A user specifying duplicate paths on the cmdline of vgcreate will
get a message similar to the following:
vgcreate vgtest2 /dev/loop3 /dev/loop5
  Found duplicate PV jk1lXsKzwyOKlXq6bhaFFKMQQ06oPgu8: using /dev/loop5 not /dev/loop3
  Found duplicate PV jk1lXsKzwyOKlXq6bhaFFKMQQ06oPgu8: using /dev/loop3 not /dev/loop5
  Internal error: Duplicate PV id jk1lXs-Kzwy-OKlX-q6bh-aFFK-MQQ0-6oPgu8 detected for /dev/loop3 in vgtest2.

This is caught by vg_validate(), but it would be good to find
this condition earlier in the vgcreate code.  add_pv_to_vg()
currently checks by pvname, but does not look for duplcate pvids.
This patch adds the check for duplicate pvids and results in new
error output as follows:
vgcreate vgtest2 /dev/loop3 /dev/loop5
  Found duplicate PV jk1lXsKzwyOKlXq6bhaFFKMQQ06oPgu8: using /dev/loop5 not /dev/loop3
  Found duplicate PV jk1lXsKzwyOKlXq6bhaFFKMQQ06oPgu8: using /dev/loop3 not /dev/loop5
  Physical volume '/dev/loop5 (jk1lXs-Kzwy-OKlX-q6bh-aFFK-MQQ0-6oPgu8)' listed more than once.
  Unable to add physical volume '/dev/loop5' to volume group 'vgtest2'.

Signed-off-by: Dave Wysochanski <dwysocha@redhat.com>
14 years agoAdd cleandir target.
Zdenek Kabelac [Thu, 8 Apr 2010 09:15:37 +0000 (09:15 +0000)]
Add cleandir target.

Using non-recursive cleandir target for resursive distclean and clean.
Avoids duplicated clean invocation during distclean.
Indent $(RM) parameters.

14 years agoAnd another test that should have been failing - normal allocation policy
Alasdair Kergon [Thu, 8 Apr 2010 01:58:06 +0000 (01:58 +0000)]
And another test that should have been failing - normal allocation policy
does not put stripes on the same disk as each other!

14 years agoanother incorrect test that should have been failing before
Alasdair Kergon [Thu, 8 Apr 2010 01:43:46 +0000 (01:43 +0000)]
another incorrect test that should have been failing before

14 years agomissing ?:
Alasdair Kergon [Thu, 8 Apr 2010 00:56:26 +0000 (00:56 +0000)]
missing ?:

14 years agosuppress bogus compiler warning
Alasdair Kergon [Thu, 8 Apr 2010 00:52:41 +0000 (00:52 +0000)]
suppress bogus compiler warning

14 years agoFix pvmove allocation to take existing parallel stripes into account.
Alasdair Kergon [Thu, 8 Apr 2010 00:28:57 +0000 (00:28 +0000)]
Fix pvmove allocation to take existing parallel stripes into account.

When moving parts of striped LVs, pvmove wouldn't care about leaving you with
two stripes on the same disk.  Now --alloc anywhere is needed for that.
(Tried and gave up on two alternative approaches before the one committed here.)

14 years agoOnly fail if the top-level LV fails to be deactivated - allow deactivation
Alasdair Kergon [Wed, 7 Apr 2010 23:51:34 +0000 (23:51 +0000)]
Only fail if the top-level LV fails to be deactivated - allow deactivation
of its dependencies to fail.

14 years agoDon't forget to cd into $TESTDIR in test-utils.sh / prepare_testdir.
Petr Rockai [Wed, 7 Apr 2010 21:38:01 +0000 (21:38 +0000)]
Don't forget to cd into $TESTDIR in test-utils.sh / prepare_testdir.

14 years agoIssue a message if the new type of deactivation failure happens.
Alasdair Kergon [Wed, 7 Apr 2010 21:25:09 +0000 (21:25 +0000)]
Issue a message if the new type of deactivation failure happens.
If this can happen during 'normal' operations, I need to know.

14 years agoFix a mis-override of $PWD in test-utils.sh.
Petr Rockai [Wed, 7 Apr 2010 21:19:20 +0000 (21:19 +0000)]
Fix a mis-override of $PWD in test-utils.sh.

14 years ago Fix incorrect removal of symlinks after LV deactivation fails.
Alasdair Kergon [Wed, 7 Apr 2010 20:04:41 +0000 (20:04 +0000)]
  Fix incorrect removal of symlinks after LV deactivation fails.

14 years agoSet ulimit -c to unlimited to allow coredumps to be collected and analysed in
Petr Rockai [Wed, 7 Apr 2010 16:04:22 +0000 (16:04 +0000)]
Set ulimit -c to unlimited to allow coredumps to be collected and analysed in
the testsuite.

14 years agoAvoid spurious skips in the testsuite due to obsoleted
Petr Rockai [Wed, 7 Apr 2010 16:00:19 +0000 (16:00 +0000)]
Avoid spurious skips in the testsuite due to obsoleted
dmsetup_has_dm_devdir_support_.

14 years agoWipe dm-ioctl parameters in memory after use.
Milan Broz [Wed, 7 Apr 2010 15:57:20 +0000 (15:57 +0000)]
Wipe dm-ioctl parameters in memory after use.

14 years agoRefactor the test utilities, dropping the legacy test-lib.sh and curtailing
Petr Rockai [Wed, 7 Apr 2010 14:46:26 +0000 (14:46 +0000)]
Refactor the test utilities, dropping the legacy test-lib.sh and curtailing
lvm-utils.sh. Clears up lots of unused code, should have little observable
impact (it does change test directory layout slightly).

14 years agoIn test harness, use fwrite(3) in place of write(2) to avoid mixing fd-level
Petr Rockai [Wed, 7 Apr 2010 09:48:11 +0000 (09:48 +0000)]
In test harness, use fwrite(3) in place of write(2) to avoid mixing fd-level
and FILE-level IO on stdout.

14 years agoKeep the testsuite stats correct in spite of failed-test repetition.
Petr Rockai [Wed, 7 Apr 2010 09:41:33 +0000 (09:41 +0000)]
Keep the testsuite stats correct in spite of failed-test repetition.

14 years agoMove a fragile testcase toward the end of t-mirror-lvconvert.sh.
Petr Rockai [Tue, 6 Apr 2010 23:04:45 +0000 (23:04 +0000)]
Move a fragile testcase toward the end of t-mirror-lvconvert.sh.

14 years agoPut back the `lvconvert -m+1 --mirrorlog disk' test in t-lvconvert-mirror.sh.
Petr Rockai [Tue, 6 Apr 2010 22:22:26 +0000 (22:22 +0000)]
Put back the `lvconvert -m+1 --mirrorlog disk' test in t-lvconvert-mirror.sh.

14 years agodisable another broken check in a test
Alasdair Kergon [Tue, 6 Apr 2010 18:13:43 +0000 (18:13 +0000)]
disable another broken check in a test

This page took 0.066061 seconds and 5 git commands to generate.