1
0
mirror of git://sourceware.org/git/lvm2.git synced 2024-12-22 17:35:59 +03:00
Commit Graph

8918 Commits

Author SHA1 Message Date
Petr Rockai
05a421960f NIX: Avoid test-installing lvm2-sysvinit (fails on recent Fedoras). 2014-04-07 20:38:43 +02:00
Petr Rockai
0996b5ccf4 spec: Add lvmthin.7.gz to packages.inc. 2014-04-07 20:38:42 +02:00
Alasdair G Kergon
f13977e8fe autoreconf 2014-04-07 16:47:07 +01:00
Alasdair G Kergon
eb1602406b configure: Warn if old thin_check tool found.
If thin_check is auto-detected, issue a warning if the version is older
than 0.3.x.
2014-04-07 16:44:29 +01:00
David Teigland
2fd5c67f32 man: lvmthin clean up lvm.conf references 2014-04-04 09:51:56 -05:00
Peter Rajnoha
a7c930b18d tools: don't require --major to be specified when using -My option on kernels > 2.4
Since kernel > 2.4 have dynamically assigned major numbers.

[0] raw/~ $ lvcreate -l1 -My --minor 10 vg
  Logical volume "lvol0" created

[0] raw/~ $ lvcreate -l1 -My --major 254 --minor 11 vg
  Ignoring supplied major number - kernel assigns major numbers dynamically. Using major number 253 instead.
  Logical volume "lvol1" created

[0] raw/~ $ lvs --profile out -o+major,minor
  lvol0 vg     -wima-----    4.00 253  10
  lvol1 vg     -wima-----    4.00 253  11
2014-04-04 13:29:39 +02:00
Alasdair G Kergon
4d0c3ed0f8 configure: regenerate
Used rhel7 autoreconf.
2014-04-04 02:36:47 +01:00
Alasdair G Kergon
5d7614fcf9 format_text: Report failed close. 2014-04-04 02:28:10 +01:00
Alasdair G Kergon
cc72f340d7 thin: Support thin_check --clear-needs-check-flag.
Update thin provisioning tools to version 0.3.2 or later!
2014-04-04 02:22:40 +01:00
Alasdair G Kergon
b38a3d8c85 lvmetad: Update retry logic vars and comments.
Avoid using variables with same name as functions.
Use lvm_even_rand for random numbers.
2014-04-04 01:46:53 +01:00
Alasdair G Kergon
d36f721bb9 reports: FIXME for unnecessary cache wipes. 2014-04-04 01:41:06 +01:00
Alasdair G Kergon
3f0f558786 dev-cache: Improve open device check messages. 2014-04-04 01:39:42 +01:00
Alasdair G Kergon
c16c1a9f70 clvmd: Update new remove_info INTERNAL_ERRORS. 2014-04-04 01:37:43 +01:00
Alasdair G Kergon
12ddaa5f10 lib: Share lvm_even_rand for random numbers. 2014-04-04 01:26:19 +01:00
Alasdair G Kergon
e7b8e0a10c vgsplit: Mark cache moving code NOTREACHED. 2014-04-04 01:19:04 +01:00
Alasdair G Kergon
6d2a26f6b6 man: Add lvmthin(7). 2014-04-04 01:14:25 +01:00
Jonathan Brassow
6c6468f91d RAID: Improve an error message
When down-converting a RAID1 LV, if the user specifies too few devices,
they will get a confusing message.
Ex:
[root]# lvcreate -m 2 --type raid1 -n raid -L 500M taft
  Logical volume "raid" created

[root]# lvconvert -m 0 taft/raid /dev/sdd1
  Unable to extract enough images to satisfy request
  Failed to extract images from taft/raid

This patch makes the error message a bit clearer by telling the user
the count they are trying to remove and the number of devices they
supplied.

[root@bp-01 lvm2]# lvcreate --type raid1 -m 3 -L 200M -n lv vg
  Logical volume "lv" created

[root@bp-01 lvm2]# lvconvert -m -3 vg/lv /dev/sdb1
  Unable to remove 3 images:  Only 1 device given.
  Failed to extract images from vg/lv

[root@bp-01 lvm2]# lvconvert -m -3 vg/lv /dev/sd[bc]1
  Unable to remove 3 images:  Only 2 devices given.
  Failed to extract images from vg/lv

[root@bp-01 lvm2]# lvconvert -m -3 vg/lv /dev/sd[bcd]1
[root@bp-01 lvm2]# lvs -a -o name,attr,devices vg
  LV   Attr       Devices
  lv   -wi-a----- /dev/sde1(1)

This patch doesn't work in all cases.  The user can specify the right
number of devices, but not a sufficient amount of devices from the LV.
This will produce the old error message:
[root@bp-01 lvm2]# lvconvert -m -3 vg/lv /dev/sd[bcf]1
  Unable to extract enough images to satisfy request
  Failed to extract images from vg/lv
However, I think this error message is sufficient for this case.
2014-04-03 16:57:41 -05:00
David Teigland
f242c0611d man: lvmthin pool space exhaustion
Initial description and instructions for
data/metadata space exhaustion.
2014-04-03 16:20:15 -05:00
David Teigland
7b11b1b4b5 man: add lvmthin man page 2014-04-01 16:04:08 -05:00
Zdenek Kabelac
c2876ee1c9 cache: enforce local exlusive activation
For cache flushing local exlusive activation is needed.
2014-04-01 21:29:28 +02:00
Zdenek Kabelac
c95d43b28c tests: update lvcreate-cache 2014-04-01 20:54:10 +02:00
Zdenek Kabelac
dc5a3c9964 debug: add internal error for passed LV
TODO: in fact we should parameter LV.
2014-04-01 20:54:09 +02:00
Zdenek Kabelac
1eaef2b705 cleanup: just rename for interal function 2014-04-01 20:54:09 +02:00
Zdenek Kabelac
9cb053339e cleanup: cache updates messages
Passing non cached device is an internal error.
Print messages at non-error level.
Shorten sleep delay for cache flush.
2014-04-01 20:54:09 +02:00
Zdenek Kabelac
e2ea3cd7ba cleanup: cache use const char policy
Policy should be const char pointer.
2014-04-01 20:54:09 +02:00
Zdenek Kabelac
d58cc2c0fc cleanup: cache reuse code for pool test
Using same error message for pool associated devices.
2014-04-01 20:18:05 +02:00
Zdenek Kabelac
e72dea55bf cache: fix order of metadata change
Start to change metadata after they are archived.
And since cache_pool is virtual skip deactivation
call for this LV.
2014-04-01 20:17:10 +02:00
Zdenek Kabelac
d3004479cc cache: use remove_layer_from_lv
Since the usability problem were fixed, we can use this function.
Cleanup orphan LVs with TEMPORARY flags
(reduces couple blkid error reports, but couple of them
is still left...)
2014-04-01 20:17:10 +02:00
Zdenek Kabelac
0a15a210a5 cache: cache segment is non-discardable
Since cache segment is purely virtual mapping, it has nothing for
discard. Discardable is cache origin here which is now
properly removed on 'delete' phase.

Plain lv_empty() call needs to only detach cache origin and leave
origin unchanged.
2014-04-01 20:17:10 +02:00
Zdenek Kabelac
a018c57f0b cache: never activate cache pool
Since cache-pool is purely lvm abstraction layer LV, it never
need any device node, so do not add even 'error' device for it.
2014-04-01 20:17:10 +02:00
Peter Rajnoha
504c328e3d tests: add profiles-thin and prepare_profiles helper fn 2014-04-01 15:52:54 +02:00
Petr Rockai
af73a0a518 libdm: Add missing UDEV_CFLAGS. 2014-04-01 15:26:26 +02:00
Petr Rockai
d5682ccbac lvm-wrappers: Remove spurious #include <libudev.h>. 2014-04-01 15:26:26 +02:00
Petr Rockai
490226fc47 pvs: Wipe persistent filters when given a device argument. 2014-04-01 14:46:22 +02:00
Peter Rajnoha
a30219a8c0 tests: update profiles.sh for recent changes 2014-04-01 11:48:49 +02:00
Peter Rajnoha
13009adbf2 man: add a note about --profile when using dumpconfig --mergedconfig 2014-04-01 11:19:38 +02:00
Peter Rajnoha
68bb639a08 config: fix compile error in config-settings.h if using --disable-dmeventd
config/config_settings.h:257:102: error: 'DMEVENTD_PATH' undeclared here (not in a function)
2014-03-31 15:57:42 +02:00
Zdenek Kabelac
367e4551a1 tests: add more rename test
Drop test, which are now in other files (listings.sh)
2014-03-31 12:05:35 +02:00
Zdenek Kabelac
ee8708cc8e tests: more listing tests 2014-03-31 12:05:35 +02:00
Zdenek Kabelac
f93c9b447b cleanup: skip eval of minor when major is true
Use normal '||'.
2014-03-31 12:05:34 +02:00
Zdenek Kabelac
a33d4355a1 cleanup: lvdisplay indent 2014-03-31 12:05:34 +02:00
Zdenek Kabelac
84beba5d7f vg_validate: check size of lv_name + vg_name
Since the whole dm device name may not exceed 127 characters,
validate no LV names is bigger then this limit.
2014-03-31 12:05:32 +02:00
Zdenek Kabelac
6570d36ad5 lv_rename: resume fail is certainly error
Failing resume path means command has failed,
even when commit was ok.
2014-03-31 12:03:25 +02:00
Zdenek Kabelac
d3e4934a15 lvrename: fix name length validation
This test for LV name restriction check name of device is below 128
chars (which is enforced by dm target).
Thus it should not count with device name.
(Though the test for PATH_MAX size should be probably also added,
but this is runtime test, since theoretically devpath might differ in cluster)
2014-03-31 12:02:18 +02:00
Zdenek Kabelac
caa429da2e lvdiplay: prohibit use of -c and -m
It's unclear why we should prohibit use of -v output.
So reenable (like with other 'display' tools)
But -c -m is really unsupported - return invalid cmd.
2014-03-31 12:00:45 +02:00
Zdenek Kabelac
8f9150c241 tests: add more pvdisplay tests 2014-03-30 23:45:00 +02:00
Zdenek Kabelac
1216aa7826 tests: workaround clvmd .cache consistency
May need futher fixes, but it's needed only on non-udev systems
(systems which still needs .cache - just like test suite)
2014-03-30 23:45:00 +02:00
Zdenek Kabelac
2c8f0c9800 tests: more vgexport tests
cover more options
2014-03-30 23:45:00 +02:00
Zdenek Kabelac
cc82dc7b23 cleanup: pvdisplay indent
Reindent else path, since other branch already exits with return.
2014-03-30 23:44:59 +02:00
Zdenek Kabelac
d7b6334079 cleanup: vgexport drops pv pointer
PV pointer is not really needed.
2014-03-30 23:44:59 +02:00