]> sourceware.org Git - lvm2.git/log
lvm2.git
9 years agoman: reference -S/--select in {vg,lv,pv}change, {vg,lv}remove, {vg,lv,pv}display...
Peter Rajnoha [Wed, 11 Feb 2015 09:03:37 +0000 (10:03 +0100)]
man: reference -S/--select in {vg,lv,pv}change, {vg,lv}remove, {vg,lv,pv}display and vg{import,export} man pages

9 years agotoollib: select: issue an error message for failed VG/LV/PV selection
Peter Rajnoha [Wed, 11 Feb 2015 08:36:09 +0000 (09:36 +0100)]
toollib: select: issue an error message for failed VG/LV/PV selection

9 years agoman: dumpconfig: mention LVM_COMMAND_PROFILE env var and add more references to ...
Peter Rajnoha [Wed, 11 Feb 2015 08:21:08 +0000 (09:21 +0100)]
man: dumpconfig: mention LVM_COMMAND_PROFILE env var and add more references to --mergedconfig option

The LVM_COMMAND_PROFILE env var is new - mention it in dumpconfig's
man page.

Also, dumpconfig always displays the top of the config cascade.
To display all the config found in the cascade merged (just like
it's used during LVM command processing), --mergedconfig option
must be used - this one's already described in that man page,
just make sure it's clear and add reference for this option also
in --profile/--commandprofile/--metadataprofile description.

9 years agoNIX: Drop Fedora 16 builds.
Petr Rockai [Tue, 10 Feb 2015 16:26:32 +0000 (17:26 +0100)]
NIX: Drop Fedora 16 builds.

9 years agoNIX: Adjust lists of extra RPMs to install.
Petr Rockai [Tue, 10 Feb 2015 16:26:02 +0000 (17:26 +0100)]
NIX: Adjust lists of extra RPMs to install.

9 years agotest: Fix harness not building on RHEL5 dev-mcsontos-rhel5-CLOEXEC
Marian Csontos [Tue, 10 Feb 2015 13:51:45 +0000 (14:51 +0100)]
test: Fix harness not building on RHEL5

9 years agoselect: initialize selection handle for process_each_* fns with initial report type
Peter Rajnoha [Tue, 10 Feb 2015 12:46:37 +0000 (13:46 +0100)]
select: initialize selection handle for process_each_* fns with initial report type

This is a followup patch for previous patchset that enables selection in
process_each_* fns to fix an issue where field prefixes are not
automatically used for fields in selection criteria.

Use initial report type that matches the intention of each process_each_* functions:
- _process_pvs_in_vg - PVS
- process_each_vg - VGS
- process_each_lv and process_each_lv_in_vg - LVS

This is not normally needed for the selection handle init, BUT we would
miss the field prefix matching, e.g.

lvchange -ay -S 'name=lvol0'

The "name" above would not work if we didn't initialize reporting with
the LVS type at its start. If we pass proper init type, reporting code
can deduce the prefix automatically ("lv_name" in this case).

This report type is then changed further based on what selection criteria we
have. When doing pure selection, not report output, the final report type
is purely based on combination of this initial report type and report types
of the fields used in selection criteria.

9 years agotools: allow -S|--select for vgexport and vgimport
Peter Rajnoha [Fri, 12 Dec 2014 11:03:34 +0000 (12:03 +0100)]
tools: allow -S|--select for vgexport and vgimport

9 years agotools: allow -S|--select for vgdisplay, lvdisplay and pvdisplay without -C
Peter Rajnoha [Fri, 12 Dec 2014 09:20:15 +0000 (10:20 +0100)]
tools: allow -S|--select for vgdisplay, lvdisplay and pvdisplay without -C

We already allowed -S|--select with {vg,lv,pv}display -C (which
was then equal to {vg,lv,pv}s command. Since we support selection
in toolib now, we can support -S also without using -C in *display
commands now.

9 years agotools: allow -S|--select for vgremove and lvremove
Peter Rajnoha [Fri, 12 Dec 2014 08:34:50 +0000 (09:34 +0100)]
tools: allow -S|--select for vgremove and lvremove

9 years agotools: allow -S|--select for vgchange, lvchange and pvchange
Peter Rajnoha [Wed, 3 Dec 2014 09:03:48 +0000 (10:03 +0100)]
tools: allow -S|--select for vgchange, lvchange and pvchange

9 years agopvchange: use processing_handle when processing items for pvchange
Peter Rajnoha [Wed, 3 Dec 2014 09:02:13 +0000 (10:02 +0100)]
pvchange: use processing_handle when processing items for pvchange

pvchange is an exception that does not use toollib yet for iterating
over the list of PVs (process_each_pv) so intialize the
processing_handle and use just like it's used in toollib.

9 years agoreport: implement report_for_selection
Peter Rajnoha [Mon, 1 Dec 2014 14:40:03 +0000 (15:40 +0100)]
report: implement report_for_selection

We have 3 input report types:
  - LVS (representing "_select_match_lv")
  - VGS (representing "_select_match_vg")
  - PVS (representing "_select_match_pv")

The input report type is saved in struct selection_handle's "orig_report_type"
 variable.

However, users can use any combination of fields of different report types in
selection criteria - the resulting report type can thus differ. The struct
selection_handle's "report_type" variable stores this resulting report type.

The resulting report_type can end up as one of:
  - LVS
  - VGS
  - PVS
  - SEGS
  - PVSEGS

This patch adds logic to report_for_selection based on (sensible) combination
of orig_report_type and report_type and calls appropriate reporting functions
or iterates over multiple items that need reporting to determine the selection
result.

9 years agoreport: add report_for_selection and use it in select_match_{pv,vg,lv}
Peter Rajnoha [Mon, 1 Dec 2014 13:19:30 +0000 (14:19 +0100)]
report: add report_for_selection and use it in select_match_{pv,vg,lv}

The report_for_selection does the actual "reporting for selection only".
The selection status will be saved in struct selection_handle's "selected"
variable.

9 years agorefactor: report: factor out code to determine final report type
Peter Rajnoha [Mon, 1 Dec 2014 13:04:29 +0000 (14:04 +0100)]
refactor: report: factor out code to determine final report type

The code to determine final report type based on combination of input
report type (determined from fields used for reporting to output and selection)
can be reused for pure reporting for selection - factor out this code into
_get_final_report_type function.

9 years agotoollib: select the whole structure if at least one of its items is selected
Peter Rajnoha [Wed, 3 Dec 2014 13:20:00 +0000 (14:20 +0100)]
toollib: select the whole structure if at least one of its items is selected

This applies to:
  - process_each_lv_in_vg - the VG is selected only if at least one of its LVs is selected
  - process_each_segment_in_lv - the LV is selected only if at least one of its LV segments is selected
  - process_each_pv_in_vg - the VG is selected only if at least one of its PVs is selected
  - process_each_segment_in_pv - the PV is selected only if at least one of its PV segments is selected

So this patch causes the selection result to be properly propagated up to callers.

9 years agotoollib: initialize handles (including reporting for selection) for _select_match_...
Peter Rajnoha [Fri, 28 Nov 2014 14:04:25 +0000 (15:04 +0100)]
toollib: initialize handles (including reporting for selection) for _select_match_* used in process_each_* fns

Call _init_processing_handle, _init_selection_handle and
_destroy_processing_handle in process_each_* and related functions to
set up and destroy handles used while processing items.

9 years agotoollib: add init_processing_handle, init_selection_handle and destroy_processing_han...
Peter Rajnoha [Fri, 28 Nov 2014 13:46:18 +0000 (14:46 +0100)]
toollib: add init_processing_handle, init_selection_handle and destroy_processing_handle helper functions

The init_processing_handle, init_selection_handle and
destroy_processing_handle are helper functions that allocate and
initialize the handles used when processing items in process_each_*
and related functions.

9 years agotoollib: pass struct processing_handle to _select_match_* functions
Peter Rajnoha [Fri, 28 Nov 2014 13:34:56 +0000 (14:34 +0100)]
toollib: pass struct processing_handle to _select_match_* functions

The "struct processing_handle" contains handles to drive the selection/matching
so pass it to the _select_match_* functions which are entry points to the
selection mechanism used in process_each_* and related functions.

This is revised and edited version of former Dave Teigland's patch which
provided starting point for all the select support in process_each_* fns.

9 years agoreport: add report_init_for_selection fn and modify report_object to support reportin...
Peter Rajnoha [Tue, 2 Dec 2014 12:14:12 +0000 (13:14 +0100)]
report: add report_init_for_selection fn and modify report_object to support reporting for selection only

The new "report_init_for_selection" is just a wrapper over
dm_report_init_with_selection that initializes reporting for selection
only. This means we're not going to do the actual reporting to output
for display and as such we intialize reporting as if no fields are reported
or sorted. The only fields "reported" are taken from the selection criteria
string and all such fields are marked as hidden automatically (FLD_HIDDEN flag).
These fields are used solely for selection criteria matching.

Also, modify existing report_object function that was used for reporting to
output for display. Now, it can either cause reporting to output or reporting
for selection only. The selection result is stored in struct selection_handle's
"selected" variable which can be handled further by any report_object caller.

9 years agotoollib: replace void *handle with struct processing_handle for use in processing...
Peter Rajnoha [Thu, 27 Nov 2014 14:02:13 +0000 (15:02 +0100)]
toollib: replace void *handle with struct processing_handle for use in processing functions (process_each_*/process_single_* and related)

This patch replaces "void *handle" with "struct processing_handle *handle"
in process_each_*, process_single_* and related functions.

The struct processing_handle consists of two handles inside now:

  - the "struct selection_handle *selection_handle" used for
    applying selection criteria while processing process_each_*,
    process_single_* and related functions (patches using this
    logic will follow)

  - the "void* custom_handle" (this is actually the original handle
    used before this patch - a pointer to custom data passed into
    process_each_*, process_single_* and related functions).

9 years agocleanup: make report type condition consistent with the rest of the code
Peter Rajnoha [Tue, 25 Nov 2014 13:13:04 +0000 (14:13 +0100)]
cleanup: make report type condition consistent with the rest of the code

It's just more readable this way (each condition resulting in a certain
report type is always on one line), just like it's used elsewhere in the
code.

9 years agotoollib: prepare select_match_{pv,vg,lv} hooks
David Teigland [Mon, 24 Nov 2014 10:08:41 +0000 (11:08 +0100)]
toollib: prepare select_match_{pv,vg,lv} hooks

These hooks will check currently processed PV/VG/LV with current selection
criteria and the processing continues only if there's a match.

9 years agolibdm: report: add dm_report_object_is_selected
Peter Rajnoha [Mon, 24 Nov 2014 10:07:39 +0000 (11:07 +0100)]
libdm: report: add dm_report_object_is_selected

The new dm_report_object_is_selected fn makes it possible to opt whether the
object reported should be displayed on output or not. Also, in addition to
that, it makes it possible to save the result of selection (either 0 or 1).

So dm_report_object_is_selected is simply  more general form of object
reporting fn - combinations now allow for:

  dm_report_object_is_selected(rh, object, 1, NULL):
  This is exactly the original dm_report_object fn and it's fully equal
  to it.

  dm_report_object_is_selected(rh, object, 0, selected):
  Do not display the result on output, but save info whether the object
  is selected or not in 'selected' variable.

  dm_report_object_is_selected(rh, object, 1, selected):
  Display the result on output (if it passes selection criteria) and save
  whether the object is selected or not in 'selected' variable.

  dm_report_object(rh, object, 0, NULL):
  This combination is not allowed - it will end up with internal error.
  We're either interested in selection status or we want to display the
  result on output or both, but never nothing of the two.

9 years agotest: Fix a typo in shell/pvmove-restart.sh.
Petr Rockai [Tue, 10 Feb 2015 11:42:05 +0000 (12:42 +0100)]
test: Fix a typo in shell/pvmove-restart.sh.

9 years agolvm: recognize LVM_COMMAND_PROFILE env var for default command profile to use in...
Peter Rajnoha [Mon, 9 Feb 2015 13:16:24 +0000 (14:16 +0100)]
lvm: recognize LVM_COMMAND_PROFILE env var for default command profile to use in LVM commands

Once LVM_COMMAND_PROFILE environment variable is specified, the profile
referenced is used just like it was specified using "<lvm command> --commandprofile".
If both --commandprofile cmd line option and LVM_COMMAND_PROFILE env
var is used, the --commandprofile cmd line option gets preference.

9 years agoreporter: properly check for _do_info_and_status return value and free the mempool...
Peter Rajnoha [Mon, 9 Feb 2015 10:48:21 +0000 (11:48 +0100)]
reporter: properly check for _do_info_and_status return value and free the mempool if created within the fn

9 years agoconfigure: Fix a warning about test/Makefile.in.
Petr Rockai [Thu, 5 Feb 2015 15:30:50 +0000 (16:30 +0100)]
configure: Fix a warning about test/Makefile.in.

9 years agoNIX: Request {fedora,centos}-release packages explicitly.
Petr Rockai [Thu, 5 Feb 2015 13:21:31 +0000 (14:21 +0100)]
NIX: Request {fedora,centos}-release packages explicitly.

9 years agoNIX: Adjust install_rpms (kernel rpm is now required).
Petr Rockai [Thu, 5 Feb 2015 12:30:18 +0000 (13:30 +0100)]
NIX: Adjust install_rpms (kernel rpm is now required).

9 years agoNIX: We need to install lcov for ./configure now as well.
Petr Rockai [Thu, 5 Feb 2015 12:30:02 +0000 (13:30 +0100)]
NIX: We need to install lcov for ./configure now as well.

9 years agoNIX: Move corosynclib-devel & valgrind-devel out of common.
Petr Rockai [Wed, 14 Jan 2015 16:29:13 +0000 (17:29 +0100)]
NIX: Move corosynclib-devel & valgrind-devel out of common.

9 years agoNIX: Allow adding/overriding test configurations.
Petr Rockai [Wed, 14 Jan 2015 16:28:44 +0000 (17:28 +0100)]
NIX: Allow adding/overriding test configurations.

9 years agoNIX: Set ulimit -c to unlimited, to get backtraces in logs.
Petr Rockai [Wed, 19 Nov 2014 10:32:23 +0000 (11:32 +0100)]
NIX: Set ulimit -c to unlimited, to get backtraces in logs.

9 years agoNIX: Add a CentOS 7.0 64b build.
Petr Rockai [Mon, 13 Oct 2014 07:18:23 +0000 (09:18 +0200)]
NIX: Add a CentOS 7.0 64b build.

9 years agotest: Fix a segfault in the test runner.
Petr Rockai [Wed, 1 Oct 2014 14:04:21 +0000 (16:04 +0200)]
test: Fix a segfault in the test runner.

9 years agotest: It is not an error for a test daemon to exit before teardown.
Petr Rockai [Wed, 1 Oct 2014 09:53:24 +0000 (11:53 +0200)]
test: It is not an error for a test daemon to exit before teardown.

9 years agotest: Remove spurious set -x in lvm-wrapper.
Petr Rockai [Wed, 1 Oct 2014 09:50:31 +0000 (11:50 +0200)]
test: Remove spurious set -x in lvm-wrapper.

9 years agotest: Properly detect warnings from testcases.
Petr Rockai [Wed, 1 Oct 2014 09:28:15 +0000 (11:28 +0200)]
test: Properly detect warnings from testcases.

9 years agotest: Only force an IO sync when the testcase has finished.
Petr Rockai [Wed, 1 Oct 2014 09:27:44 +0000 (11:27 +0200)]
test: Only force an IO sync when the testcase has finished.

9 years agotest: Allow a longer timeout for lvextend-snapshot-dmeventd.
Petr Rockai [Wed, 1 Oct 2014 07:52:09 +0000 (09:52 +0200)]
test: Allow a longer timeout for lvextend-snapshot-dmeventd.

9 years agotest: Make dmsetup remove more robust in pvmove-restart.
Petr Rockai [Wed, 1 Oct 2014 07:08:26 +0000 (09:08 +0200)]
test: Make dmsetup remove more robust in pvmove-restart.

9 years agotest: Make killing local (test) daemons more robust.
Petr Rockai [Wed, 1 Oct 2014 06:53:27 +0000 (08:53 +0200)]
test: Make killing local (test) daemons more robust.

9 years agotest: Disable lvmetad debug log on another bunch of tests.
Petr Rockai [Wed, 1 Oct 2014 06:31:53 +0000 (08:31 +0200)]
test: Disable lvmetad debug log on another bunch of tests.

9 years agotest: Call $VALGRIND directly in lvm-wrapper.
Petr Rockai [Wed, 1 Oct 2014 06:19:05 +0000 (08:19 +0200)]
test: Call $VALGRIND directly in lvm-wrapper.

9 years agoNIX: Add valgrind-devel to the build environment.
Petr Rockai [Wed, 1 Oct 2014 06:18:46 +0000 (08:18 +0200)]
NIX: Add valgrind-devel to the build environment.

9 years agomemlock: Do not check memory balance if RUNNING_ON_VALGRIND.
Petr Rockai [Wed, 1 Oct 2014 06:17:59 +0000 (08:17 +0200)]
memlock: Do not check memory balance if RUNNING_ON_VALGRIND.

9 years agoconfigure: Look for valgrind.h independently of VALGRIND_POOLS.
Petr Rockai [Wed, 1 Oct 2014 06:16:27 +0000 (08:16 +0200)]
configure: Look for valgrind.h independently of VALGRIND_POOLS.

9 years agotest: Add support for running LVM commands in tests through strace.
Petr Rockai [Tue, 30 Sep 2014 15:09:26 +0000 (17:09 +0200)]
test: Add support for running LVM commands in tests through strace.

9 years agotest: Make lvm-wrapper available in the installed testsuite, too.
Petr Rockai [Tue, 30 Sep 2014 15:08:34 +0000 (17:08 +0200)]
test: Make lvm-wrapper available in the installed testsuite, too.

9 years agoNIX: Install valgrind, gdb and strace in the test environment.
Petr Rockai [Tue, 30 Sep 2014 15:06:46 +0000 (17:06 +0200)]
NIX: Install valgrind, gdb and strace in the test environment.

9 years agoNIX: Make timeout and test envvars configurable via --arg.
Petr Rockai [Tue, 30 Sep 2014 15:06:19 +0000 (17:06 +0200)]
NIX: Make timeout and test envvars configurable via --arg.

9 years agotest: Fix prepare_scsi_debug_dev.
Petr Rockai [Tue, 30 Sep 2014 14:50:31 +0000 (16:50 +0200)]
test: Fix prepare_scsi_debug_dev.

9 years agotest: Avoid passing -c to flock.
Petr Rockai [Tue, 30 Sep 2014 14:32:11 +0000 (16:32 +0200)]
test: Avoid passing -c to flock.

9 years agotest: Re-try with loopback if a backing device is too small.
Petr Rockai [Tue, 30 Sep 2014 14:19:14 +0000 (16:19 +0200)]
test: Re-try with loopback if a backing device is too small.

9 years agoNIX: Detect that the VM booted up more directly.
Petr Rockai [Tue, 30 Sep 2014 09:02:25 +0000 (11:02 +0200)]
NIX: Detect that the VM booted up more directly.

9 years agoNIX: Avoid a spurious failure when all tests pass.
Petr Rockai [Tue, 30 Sep 2014 09:01:48 +0000 (11:01 +0200)]
NIX: Avoid a spurious failure when all tests pass.

9 years agotest: Fix failures due to undefined abs_top_builddir.
Petr Rockai [Sun, 14 Sep 2014 22:19:53 +0000 (00:19 +0200)]
test: Fix failures due to undefined abs_top_builddir.

9 years agotest: Avoid running 'dmsetup remove' with no devices.
Petr Rockai [Sun, 14 Sep 2014 21:01:31 +0000 (23:01 +0200)]
test: Avoid running 'dmsetup remove' with no devices.

9 years agotest: Fix 100%FREE RAID tests since test device sizes changed slightly.
Petr Rockai [Sun, 14 Sep 2014 09:46:13 +0000 (11:46 +0200)]
test: Fix 100%FREE RAID tests since test device sizes changed slightly.

9 years agotest: Update brick-shelltest (configurable timeout, better banner, exit code).
Petr Rockai [Sun, 14 Sep 2014 09:15:51 +0000 (11:15 +0200)]
test: Update brick-shelltest (configurable timeout, better banner, exit code).

9 years agoNIX: Mark builds with failed tests as such.
Petr Rockai [Sun, 14 Sep 2014 09:12:12 +0000 (11:12 +0200)]
NIX: Mark builds with failed tests as such.

9 years agotest: Do not rely on dev/mapper/* nodes in pvmove-restart.sh.
Petr Rockai [Sun, 14 Sep 2014 09:11:50 +0000 (11:11 +0200)]
test: Do not rely on dev/mapper/* nodes in pvmove-restart.sh.

9 years agotest: Fix implementation of check targets with new flavour system.
Petr Rockai [Sun, 14 Sep 2014 09:11:12 +0000 (11:11 +0200)]
test: Fix implementation of check targets with new flavour system.

9 years agoNIX: Update path to /xchg (was /tmp/xchg with older nixpkgs).
Petr Rockai [Sat, 13 Sep 2014 15:53:42 +0000 (17:53 +0200)]
NIX: Update path to /xchg (was /tmp/xchg with older nixpkgs).

9 years agoNIX: Do not fail aggregation when a sub-job fails to produce output.
Petr Rockai [Wed, 10 Sep 2014 16:18:45 +0000 (18:18 +0200)]
NIX: Do not fail aggregation when a sub-job fails to produce output.

9 years agoNIX: Run each test flavour as a separate derivation.
Petr Rockai [Tue, 29 Jul 2014 12:30:22 +0000 (14:30 +0200)]
NIX: Run each test flavour as a separate derivation.

9 years agoNIX: Add --kmsg to lvm2-testsuite invocations.
Petr Rockai [Tue, 22 Jul 2014 16:44:53 +0000 (18:44 +0200)]
NIX: Add --kmsg to lvm2-testsuite invocations.

9 years agotest: Update brick-shelltest.h.
Petr Rockai [Tue, 22 Jul 2014 16:44:32 +0000 (18:44 +0200)]
test: Update brick-shelltest.h.

9 years agoNIX: Keep the udev log out of tmpfs.
Petr Rockai [Tue, 22 Jul 2014 09:46:19 +0000 (11:46 +0200)]
NIX: Keep the udev log out of tmpfs.

9 years agoNIX: Some kernels need ide/ata modules (i.e. no builtin ata).
Petr Rockai [Mon, 21 Jul 2014 01:56:34 +0000 (03:56 +0200)]
NIX: Some kernels need ide/ata modules (i.e. no builtin ata).

9 years agoNIX: Use IDE drive emulation instead of virtio.
Petr Rockai [Sun, 20 Jul 2014 23:54:42 +0000 (01:54 +0200)]
NIX: Use IDE drive emulation instead of virtio.

9 years agolibdaemon: Use select to yield CPU on a blocked read or write.
Petr Rockai [Sun, 20 Jul 2014 23:53:48 +0000 (01:53 +0200)]
libdaemon: Use select to yield CPU on a blocked read or write.

9 years agotest: Update brick-shelltest, adding timestamps to logs.
Petr Rockai [Sun, 6 Jul 2014 14:22:23 +0000 (16:22 +0200)]
test: Update brick-shelltest, adding timestamps to logs.

9 years agoNIX: Use the new --heartbeat option on lvm2-testsuite.
Petr Rockai [Fri, 4 Jul 2014 16:28:10 +0000 (18:28 +0200)]
NIX: Use the new --heartbeat option on lvm2-testsuite.

9 years agotest: Package up the test runner in a reusable C++ "brick".
Petr Rockai [Fri, 4 Jul 2014 16:24:56 +0000 (18:24 +0200)]
test: Package up the test runner in a reusable C++ "brick".

9 years agoNIX: Pass --fatal-timeouts to the testsuite.
Petr Rockai [Fri, 4 Jul 2014 07:45:56 +0000 (09:45 +0200)]
NIX: Pass --fatal-timeouts to the testsuite.

9 years agotest: Implement --fatal-timeouts to allow a timely VM restart.
Petr Rockai [Tue, 1 Jul 2014 19:34:06 +0000 (21:34 +0200)]
test: Implement --fatal-timeouts to allow a timely VM restart.

9 years agotest: Make the runner's heartbeat more reliable.
Petr Rockai [Tue, 1 Jul 2014 19:22:58 +0000 (21:22 +0200)]
test: Make the runner's heartbeat more reliable.

9 years agotest: Disable lvmetad logging in lvconvert-{mirror,raid}*.
Petr Rockai [Mon, 21 Jul 2014 19:44:53 +0000 (21:44 +0200)]
test: Disable lvmetad logging in lvconvert-{mirror,raid}*.

9 years agoNIX: Re-work VM monitoring.
Petr Rockai [Sun, 29 Jun 2014 15:37:35 +0000 (17:37 +0200)]
NIX: Re-work VM monitoring.

9 years agoNIX: Add Fedora 20 jobs.
Petr Rockai [Sun, 29 Jun 2014 15:37:16 +0000 (17:37 +0200)]
NIX: Add Fedora 20 jobs.

9 years agoNIX: Existence of results list no longer indicates that the suite has finished.
Petr Rockai [Sat, 28 Jun 2014 10:51:18 +0000 (12:51 +0200)]
NIX: Existence of results list no longer indicates that the suite has finished.

9 years agoNIX: Drop -h from rpm invocations as it messes up logs.
Petr Rockai [Sat, 28 Jun 2014 10:50:57 +0000 (12:50 +0200)]
NIX: Drop -h from rpm invocations as it messes up logs.

9 years agoNIX: Use nixpkgs kernel for building, distro kernel for testing.
Petr Rockai [Sat, 28 Jun 2014 09:16:46 +0000 (11:16 +0200)]
NIX: Use nixpkgs kernel for building, distro kernel for testing.

9 years agoNIX: Adapt to the append-based test journal.
Petr Rockai [Sat, 28 Jun 2014 09:16:22 +0000 (11:16 +0200)]
NIX: Adapt to the append-based test journal.

9 years agoNIX: Implement a smarter VM monitoring loop.
Petr Rockai [Fri, 27 Jun 2014 21:09:30 +0000 (23:09 +0200)]
NIX: Implement a smarter VM monitoring loop.

9 years agoNIX: Use --batch mode for running the testsuite.
Petr Rockai [Fri, 27 Jun 2014 21:09:02 +0000 (23:09 +0200)]
NIX: Use --batch mode for running the testsuite.

9 years agoNIX: Refactor around the new test runner.
Petr Rockai [Thu, 26 Jun 2014 23:25:18 +0000 (01:25 +0200)]
NIX: Refactor around the new test runner.

- The RPM build and the tests are now executed in separate VMs.
- Run the testsuite by using the new lvm2-testsuite RPM.
- The VM running the tests is restarted from the outside if it hangs, and the
  runner keeps a journal to avoid running a bad test ad infinitum.
- TODO: lcov reports and more intelligent VM rebooting (track the journal)

9 years agospec: Add an (optional) lvm2-testsuite subpackage.
Petr Rockai [Thu, 26 Jun 2014 23:24:47 +0000 (01:24 +0200)]
spec: Add an (optional) lvm2-testsuite subpackage.

9 years agotest: Implement --only option to runner.
Petr Rockai [Tue, 1 Jul 2014 12:07:27 +0000 (14:07 +0200)]
test: Implement --only option to runner.

9 years agotest: Fix loading of an append-mode journal.
Petr Rockai [Sun, 29 Jun 2014 22:38:07 +0000 (00:38 +0200)]
test: Fix loading of an append-mode journal.

9 years agotest: Add a missing copyright/licence notice to journal.h.
Petr Rockai [Sun, 29 Jun 2014 12:30:49 +0000 (14:30 +0200)]
test: Add a missing copyright/licence notice to journal.h.

9 years agotest: Fix a precedence bug in the runner.
Petr Rockai [Sat, 28 Jun 2014 09:15:28 +0000 (11:15 +0200)]
test: Fix a precedence bug in the runner.

9 years agotest: Use append for journal maintenance instead of rename.
Petr Rockai [Sat, 28 Jun 2014 09:15:13 +0000 (11:15 +0200)]
test: Use append for journal maintenance instead of rename.

9 years agotest: Implement specialised --batch output style w/ rusage stats.
Petr Rockai [Fri, 27 Jun 2014 21:19:43 +0000 (23:19 +0200)]
test: Implement specialised --batch output style w/ rusage stats.

9 years agotest: Add time/rusage formatting to the runner.
Petr Rockai [Fri, 27 Jun 2014 21:19:21 +0000 (23:19 +0200)]
test: Add time/rusage formatting to the runner.

9 years agotest: Print a heartbeat file (every 20s) in the runner.
Petr Rockai [Fri, 27 Jun 2014 21:18:14 +0000 (23:18 +0200)]
test: Print a heartbeat file (every 20s) in the runner.

9 years agotest: Avoid sleeps in the runner monitoring loop.
Petr Rockai [Fri, 27 Jun 2014 21:17:47 +0000 (23:17 +0200)]
test: Avoid sleeps in the runner monitoring loop.

9 years agotest: Run in /tmp if LVM_TEST_DIR is unset.
Petr Rockai [Fri, 27 Jun 2014 21:16:49 +0000 (23:16 +0200)]
test: Run in /tmp if LVM_TEST_DIR is unset.

This page took 0.06481 seconds and 5 git commands to generate.