Peter Rajnoha [Tue, 21 Oct 2014 10:01:57 +0000 (12:01 +0200)]
report: add new LVSSTATUS and SEGSSTATUS report type
Similar to LVSINFO type which gathers LV + its DM_DEVICE_INFO, the
new LVSSTATUS/SEGSSTATUS report type will gather LV/segment + its
DM_DEVICE_STATUS.
Since we can report status only for certain segment, in case
of LVSSTATUS we need to choose which segment related to the LV
should be processed that represents the "LV status". In case of
SEGSSTATUS type it's clear - the status is reported for the
segment just processed.
Peter Rajnoha [Mon, 20 Oct 2014 11:46:50 +0000 (13:46 +0200)]
refactor: rename struct lv_with_info used in reporting code to lv_with_info_and_seg_status
The former struct lv_with_info is renamed to lv_with_info_and_seg_status as it can
hold more than just "info", there's lv's segment status now in addition:
Where lv_seg points to lv's segment that is being reported or
processed in general.
New struct lv_seg_status keeps the information about segment status -
the status retrieved via DM_DEVICE_STATUS ioctl. This information will
be used for reporting dm device target status for the LV segment
specified.
So this patch introduces third level of LV information that is
kept for reuse while reporting fields within one reporting line,
causing only one DM_DEVICE_STATUS ioctl call per LV segment line
reported (otherwise we'd need to call the DM_DEVICE_STATUS for each
segment status field in one LV segment/reporting line which is not
efficient).
So currently we have three levels of information that can be used
to report an LV/LV segment:
- LV metadata itself (struct logical_volume *lv)
- LV's DM_DEVICE_INFO ioctl result (struct lvinfo *info)
- LV's segment DM_DEVICE_STATUS ioctl result (this status must be
bound to a segment, not the whole LV as the whole LV may be
composed of several segments of course)
(this is the new struct lv_seg_status *seg_status)
Zdenek Kabelac [Wed, 5 Nov 2014 14:15:07 +0000 (15:15 +0100)]
snapshot: virtual save commit
More efficient spare volume creation. Save 1 extra commit
and properly activate this volume according to our cluster
activation rules (using lv_active_change() for this).
Zdenek Kabelac [Sun, 9 Nov 2014 18:48:39 +0000 (19:48 +0100)]
lv_manip: remove_layer_from_lv restore lv name
Since we 'layer' for cache origin which and we support dropping
cache layer - we need to restore origin name in case
the origin LV is more complex target - i.e. raid.
Zdenek Kabelac [Wed, 5 Nov 2014 19:27:51 +0000 (20:27 +0100)]
pools: after fail of vg_write no restore
Revert part of commit 51a29e60564d537,
it's probably bad idea to continue with any recovery, when
vg_write() or vg_commit() fail - so it's better to leave it as it is.
Zdenek Kabelac [Wed, 5 Nov 2014 10:53:11 +0000 (11:53 +0100)]
activate: check all snap segs are inactive
When deactivating origin, we may have possibly left table in broken state,
where origin is not active, but snapshot volume is still present.
Let's ensure deactivation of origin detects also all associated
snapshots are inactive - otherwise do not skip deactivation.
(so i.e. 'vgchange -an' would detect errors)
Zdenek Kabelac [Wed, 5 Nov 2014 14:14:58 +0000 (15:14 +0100)]
lv: lv_active_change add needs_exclusive flag
Let's use this function for more activations in the code.
'needs_exlusive' will enforce exlusive type for any given LV.
We may want to activate LV in exlusive mode, even when we know
the LV (as is) supports non-exlusive activation as well.
lvcreate -ay -> exclusive & local
lvcreate -aay -> exclusive & local
lvcreate -aly -> exclusive & local
lvcreate -aey -> exclusive (might be on any node).
Peter Rajnoha [Wed, 5 Nov 2014 09:35:38 +0000 (10:35 +0100)]
report: cleanup: simplify LVSINFO detection
LVSINFO is just a subtype of LVS report type with extra "info" ioctl
called for each LV reported (per output line) so include its processing
within "case LVS" switch, not as completely different kind of reporting
which may be misleading when reading the code.
There's already the "lv_info_needed" flag set in the _report fn, so
call the approriate reporting function based on this flag within the
"case LVS" switch line.
Actually the same is already done for LV is reported per segments
within the "case SEGS" switch line. So this patch makes the code more
consistent so it's processed the same way for all cases.
Also, this is a preparation for another and new subtype that will
be introduced later - the "LVSSTATUS" and "SEGSSTATUS" report type.
When responding to DM_EVENT_CMD_GET_REGISTERED_DEVICE no longer
ignore threads that have already been unregistered but which
are still present.
This means the caller can unregister a device and poll dmeventd
to ensure the monitoring thread has gone away before removing
the device. If a device was registered and unregistered in quick
succession and then removed, WAITEVENT could run in parallel with
the REMOVE.
Threads are moved to the _thread_registry_unused list when they
are unregistered.
Zdenek Kabelac [Tue, 4 Nov 2014 09:33:35 +0000 (10:33 +0100)]
thin: new pool is activated without overlay
Activate of new/unused/empty thin pool volume skips
the 'overlay' part and directly provides 'visible' thin-pool LV to the user.
Such thin pool still gets 'private' -tpool UUID suffix for easier
udev detection of protected lvm2 devices, and also gets udev flags to
avoid any scan.
Such pool device is 'public' LV with regular /dev/vgname/poolname link,
but it's still 'udev' hidden device for any other use.
To display proper active state we need to do few explicit tests
for this condition.
Before it's used for any lvm2 thin volume, deactivation is
now needed to avoid any 'race' with external usage.
Zdenek Kabelac [Tue, 4 Nov 2014 14:08:15 +0000 (15:08 +0100)]
thin: check for new pool before creating thin volume
Call check_new_thin_pool() to detect in-use thin-pool.
Save extra reactivation of thin-pool when thin pool is not active.
(it's now a bit more expensive to invoke thin_check for new pools.)
For new pools:
We now active locally exclusively thin-pool as 'public' LV.
Validate transaction_id is till 0.
Deactive.
Prepare create message for thin-pool and exclusively active pool.
Active new thin LV.
And deactivate thin pool if it used to be inactive.
Zdenek Kabelac [Tue, 4 Nov 2014 08:42:15 +0000 (09:42 +0100)]
libdm: allow to activate any pool with tid == 0
When transaction_id is set 0 for thin-pool, libdm avoids validation
of thin-pool, unless there are real messages to be send to thin-pool.
This relaxes strict policy which always required to know
in front transaction_id for the kernel target.
It now allows to activate thin-pool with any transaction_id
(when transaction_id is passed in)
It is now upto application to validate transaction_id from life
thin-pool volume with transaction_id within it's own metadata.
Zdenek Kabelac [Mon, 3 Nov 2014 11:52:29 +0000 (12:52 +0100)]
cache: lv_cache_status
Replace lv_cache_block_info() and lv_cache_policy_info()
with lv_cache_status() which directly returns
dm_status_cache structure together with some calculated
values.
After use mem pool stored inside lv_status_cache structure
needs to be destroyed.
Zdenek Kabelac [Sun, 2 Nov 2014 16:48:21 +0000 (17:48 +0100)]
cache: allow deactivation of empty pool
Tool will use internal activation of unused cache pool to
clear metadata area before next use of cache-pool.
So allow to deactivation unused pool in case some error
case happend and we were not able to deactivation pool
right after metadata wipe.
Zdenek Kabelac [Thu, 30 Oct 2014 22:43:12 +0000 (23:43 +0100)]
metadata: add extents_from_percent_size
Add API call to calculate extents from percentage value.
Size is based in DM_PERCENT_1 units.
(Supporting decimal point number).
This commit is preparing functionality for more global
usage of % with i.e. --size option.
Zdenek Kabelac [Tue, 28 Oct 2014 10:20:33 +0000 (11:20 +0100)]
lvmcmdline: extra support for decimal point
If root has locales with different decimal point then '.'
(i.e. Czech with ',') lets be tolerant and retry with
"C" locales in the case '.' is found during parse of number.
Peter Rajnoha [Mon, 27 Oct 2014 10:25:08 +0000 (11:25 +0100)]
report: selection: fix selection criteria to not match reserved values when using >, <, >=, <
Some values are reserved for special purpose like 'undefined', 'unmanaged' etc.
When using >, <, >= and < comparison operators where the range is considered,
do not include reserved values as proper values in this range which
would otherwise result in not so obvious criteria match (as the reserved value is
actually transparent for the user). It's incorrect.
However, we still need to fix situaton where the reserved value may be
included in resulting range:
Before this patch:
$ vgs -o vg_name,vg_mda_copies vg1 vg2 -S 'vg_mda_copies >= 1'
VG #VMdaCps
vg1 1
vg2 unmanaged
With this patch applied:
$ vgs -o vg_name,vg_mda_copies vg1 vg2 -S 'vg_mda_copies >= 1'
VG #VMdaCps
vg1 1
From the examples above, we can see that without this patch applied,
the vg_mda_copies >= 1 also matched the reserved value 18446744073709551615
(which is represented by the "unamanged" string on report). When
applying the operators, such values must be skipped! They're meant to
be matched only against their string representation only, e.g.: