IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
Commit 80a6de616a versioned the dm_tree_node_add_raid_target_with_params()
and dm_tree_node_add_raid_target() APIs for compatibility reasons.
There's no user of the latter function, remove it.
Related: rhbz834579
Related: rhbz1191935
Related: rhbz1191978
Commit 27384c52cf lowered the maximum number of devices
back to 64 for compatibility.
Because more members have been added to the API in
'struct dm_tree_node_raid_params *', we have to version
the public libdm RAID API to not break any existing users.
Changes:
- keep the previous 'struct dm_tree_node_raid_params' and
dm_tree_node_add_raid_target_with_params()/dm_tree_node_add_raid_target()
in order to expose the already released public RAID API
- introduce 'struct dm_tree_node_raid_params_v2' and additional functions
dm_tree_node_add_raid_target_with_params_v2()/dm_tree_node_add_raid_target_v2()
to be used by the new lvm2 lib reshape extentions
With this new API, the bitfields for rebuild/writemostly legs in
'struct dm_tree_node_raid_params_v2' can be raised to 256 bits
again (253 legs maximum supported in MD kernel).
Mind that we can limit the maximum usable number via the
DEFAULT_RAID{1}_MAX_IMAGES definition in defaults.h.
Related: rhbz834579
Related: rhbz1191935
Related: rhbz1191978
In order to support striped raid5/6/10 LV reshaping (change
of LV type, stripesize or number of legs), this patch
introduces infrastructure prerequisites to be used
by raid_manip.c extensions in followup patches.
This base is needed for allocation of out-of-place
reshape space required by the MD raid personalities to
avoid writing over data in-place when reading off the
current RAID layout or number of legs and writing out
the new layout or to a different number of legs
(i.e. restripe)
Changes:
- add members reshape_len to 'struct lv_segment' to store
out-of-place reshape length per component rimage
- add member data_copies to struct lv_segment
to support more than 2 raid10 data copies
- make alloc_lv_segment() aware of both reshape_len and data_copies
- adjust all alloc_lv_segment() callers to the new API
- add functions to retrieve the current data offset (needed for
out-of-place reshaping space allocation) and the devices count
from the kernel
- make libdm deptree code aware of reshape_len
- add LV flags for disk add/remove reshaping
- support import/export of the new 'struct lv_segment' members
- enhance lv_extend/_lv_reduce to cope with reshape_len
- add seg_is_*/segtype_is_* macros related to reshaping
- add target version check for reshaping
- grow rebuilds/writemostly bitmaps to 246 bit to support kernel maximal
- enhance libdm deptree code to support data_offset (out-of-place reshaping)
and delta_disk (legs add/remove reshaping) target arguments
Related: rhbz834579
Related: rhbz1191935
Related: rhbz1191978
When we preload device with smaller size, we avoid its resume,
so later suspend/resume of full device tree my process all
existing in flight bios.
Also update comment and avoid using confusing opposite meaning.
Add:
- support for segment types raid6_{ls,rs,la,ra}_6
(striped raid with dedicated last Q-Syndrome SubLVs)
- conversion support from raid5_{ls,rs,la,ra} to/from raid6_{ls,rs,la,ra}_6
- setting convenient segtypes on conversions from/to raid4/5/6
- related tests to lvconvert-raid-takeover.sh factoring
out _lvcreate,_lvconvert funxtions
Related: rhbz1366296
Add:
- support for segment type raid6_n_6 (striped raid with dedicated last parity/Q-Syndrome SubLVs)
- conversion support from striped/raid0/raid0_meta/raid4 to/from raid6_n_6
- related tests to lvconvert-raid-takeover.sh
Related: rhbz1366296
Add:
- support for segment type raid5_n (striped raid with dedicated last parity SubLVs)
- conversion support from striped/raid0/raid0_meta/raid4 to/from raid5_n
- related tests to lvconvert-raid-takeover.sh
Related: rhbz1366296
The system is likely in some very inconsisten state.
Do not try to make it even more problematic with trying
to invoke tools like thin_check via callback.
When dm_tree_find_node_by_uuid() fails to find passed uuid,
report in lof_debug the complete original uuid,
not the one stripped of LVM- prefix.
TODO: inspect manipulation with LVM- prefix here.
Wrong thin-pool feature flag ordering in dm table: It will lead to
unnecessary table reload.
Fix it by placeing feature flags in order they are returned from the
kernel so current 'table line diff' code will not see a difference.
When preloading thin-pool device node for already
existing/running thin-pool do not resume such thin-pool.
This allows to properly schedule commit point for metadata,
when thin-pool data or metadata volume is resized.
Extra space between 'cache' target and metadata device caused
string comparation being not equal and thus always causing
table reload even when uneeded.
Add more functionality to size_changed function.
While 'existing' API only detected 0 for
unchanged, and !0 for changed,
new improved API will also detected if the
size has only went bigger - or there was
size reduction.
Function work for the whole dm-tree - so
no change is size is always 0.
only size extension 1.
and if some size reduction is there - returns -1.
This result can be used for better evaluation
whether we need to flush before suspend.
Avoid validation of free space in pool, when no messages are passed.
Patch a3c7e326c3 add new check for
pool overload - but this check should not be made if there are
no messages and transaction_id is still within 'bounds' (bigger by 1).
Revert back to already existing behavior which has been slightly
modified by a900d150e4.
At the end however it seem to be equal to change TID right with first
metadata write.
Existing code missed handling for 'unused' thin-pool which would
require to also check empty message list for TID==0.
So with the fix we now again preserve 'active' thin-pool volume
when first thin volume is created - this property was lost and caused
problems in cluster, where the lock was hold, but volume was no longer
active on the node.
Another missing part was the proper support for already increased,
but unfinished TID change.
So going back here with existing logic -
TID is increased with first MDA update.
Code allows start with either same TID or (TID-1).
If there are messages, TID must be lower by 1 for sending,
otherwise messages were already posted.
Existing messaging intarface for thin-pool has a few 'weak' points:
* Message were posted with each 'resume' operation, thus not allowing
activation of thin-pool with the existing state.
* Acceleration skipped suspend step has not worked in cluster,
since clvmd resumes only nodes which are suspended (have proper lock
state).
* Resume may fail and code is not really designed to 'fail' in this
phase (generic rule here is resume DOES NOT fail unless something serious
is wrong and lvm2 tool usually doesn't handle recovery path in this case.)
* Full thin-pool suspend happened, when taken a thin-volume snapshot.
With this patch the new method relocates message passing into suspend
state.
This has a few drawbacks with current API, but overal it performs
better and gives are more posibilities to deal with errors.
Patch introduces a new logic for 'origin-only' suspend of thin-pool and
this also relates to thin-volume when taking snapshot.
When suspend_origin_only operation is invoked on a pool with
queued messages then only those messages are posted to thin-pool and
actual suspend of thin pool and data and metadata volume is skipped.
This makes taking a snapshot of thin-volume lighter operation and
avoids blocking of other unrelated active thin volumes.
Also fail now happens in 'suspend' state where the 'Fail' is more expected
and it is better handled through error paths.
Activation of thin-pool is now not sending any message and leaves upto a tool
to decided later how to finish unfinished double-commit transaction.
Problem which needs some API improvements relates to the lvm2 tree
construction. For the suspend tree we do not add target table line
into the tree, but only a device is inserted into a tree.
Current mechanism to attach messages for thin-pool requires the libdm
to know about thin-pool target, so lvm2 currently takes assumption, node
is really a thin-pool and fills in the table line for this node (which
should be ensured by the PRELOAD phase, but it's a misuse of internal API)
we would possibly need to be able to attach message to 'any' node.
Other thing to notice - current messaging interface in thin-pool
target requires to suspend thin volume origin first and then send
a create message, but this could not have any 'nice' solution on lvm2
side and IMHO we should introduce something like 'create_after_resume'
message.
Patch also changes the moment, where lvm2 transaction id is increased.
Now it happens only after successful finish of kernel transaction id
change. This change was needed to handle properly activation of pool,
which is in the middle of unfinished transaction, and also this corrects
usage of thin-pool by external apps like Docker.
Add support for sending message in suspend tree for thin-pools.
When this operation is requested whole subtree suspend is then skipped.
This is experimantal support for new lvm2 code for sending message
in suspend phase where 'thin-pool origin-only suspend' will send
messages instead of really suspending thin-pool tree.
When suspening thin volume origin-only - only thin volume is suspended,
then messages are posted and thin-pool suspend is skipped.
_node_name() prepares into dm_tree internal buffer device
name and it (major:minor) for easy usage for debug messages.
To avoid any allocation a small buffer in struct dm_tree is preallocated
to store this message.
Support error_if_no_space feature for thin pools.
Report more info about thinpool status:
(out_of_data (D), metadata_read_only (M), failed (F) also as health
attribute.)
Do not use 'any' policy name as a value in config tree - so we stick
with 'policy_settings' and extra 'policy_name' for libdm params.
Update lvm2 API as well.
Example of supported metadata:
policy = "mq"
policy_settings {
migration_threshold = 2048
sequential_threshold = 512
random_threshold = 4
read_promote_adjustment = 10
}
Support new PASSTHROUGH 'feature' flag.
Add dm_config_node to pass in policy args.
Really use origin_uuid instead of using extra call
to pass seg_areas.
Switch to 64bit feature flag bit set so there is
enough space in future for new bits...
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.
Add extra safety detection for thin pool transaction id
and query pool status after confirmed message.
In case there is a missmatch, immeditelly abort further
processing.
2.02.106 added suffixes to some LV uuids in the kernel.
If any of these LVs is activated with 2.02.105 or earlier,
and then a later version is used, the LVs appear invisible and
activation commands fail.
The code now has to check the kernel for both old and new uuids.
As part of better error handling, remove DM devices that have been
sucessfully created but failed to load a table. This can happen
when pvmove'ing in a cluster and the cluster mirror daemon is not
running on a remote node - the mapping table failing to load as a
result. In this case, any revert would work on other nodes running
cmirrord because the DM devices on those nodes did succeed in loading.
However, because no table was able to load on the non-cmirrord nodes,
there is no table present that points to what needs to be reverted.
This causes the empty DM device to remain on the system without being
present in any LVM representation.
This patch should only be considered a partial fix to the overall
problem. This is because only the device which failed to load a
table is removed. Any LVs that may have been loaded as requirements
to the DM device that failed to load may be left in place. Complete
clean-up will require tracking those devices which have been created
as dependencies and removing them along with the device that failed
to load a table.
When the node enters dtree with implicit dependency, it
automatically has udev flags from parent node
and could not be changed later when the node has been
entered again via i.e lvm's preload tracking.
Resolve this by tracking whether the node has been
created by implicit dependency tracking or has been
entered explicitely. Implicit node could be later
upgraded by an explicit _add_dev() with proper udev_flags.
For implicit devices add special udev flags to avoid
any scan and udev rule processing if we resume such device.
Patch allows easier removing of orphan nodes.
Reuse _node_send_messages for just checking
for valid transaction_id with preload.
This allows earlier detection of incosistent thin pool.
Code does the same thing, except for sending messages.
Improve testing of transation_id to not allow other difference
then either kernel TID is equal or is lower by oned and there
are queued messages for transaction.
Mark messages as submitted if the transaction_id is already matching.
Do not try to deactivate node on failure here and leave it on
proper error path of the caller.
Deactivation of top level node has to happen,
before traversing subtree.
Swap list logic and rather append new nodes to the head
and then use normal iteration.
(in-release update)
Avoid introducing libdm structure allocated in library user.
Use direct call with all currently supported args.
When new arg is added, new function will cover it.
This patch defines a structure for holding all of the device-mapper
cache target's status information. The associated function provides
an easy way for higher levels (LVM) to consume the information.
This patch finishes the device-mapper interface for the cache and
cachepool segment types (i.e. the cache target).
This patch adds the cache segment type - the second of two necessary
to create cache logical volumes. This segment type references the
cachepool (the small fast device) and the origin (the large slow device);
linking them to create the cache device. The cache device is the
hierarchical device-mapper device that the user ulitmately makes use
of.
The cache segment sources the information necessary to construct the
device-mapper cache target from the origin and cachepool segments to
which it links.
Revert activated volumes if callback fails.
This is currently used only for thin_check failure support.
When thin_check detects failure in thin metadata device, it deactivate
volumes in reversed order that have been preloaded for thin pool activation.
After this change lvm command will not leave active pool subvolumes
in dm table.
Pass dnode pointer instead of rather unknown child pointer.
The pointer is currently unused and passing child pointer
is quite undefined, while dnode has at least some usability.
When images and their associated metadata are removed from a RAID1 LV,
the remaining sub-LVs are "shifted" down to fill the gaps. For
example, if there is a 3-way mirror:
[0][1][2]
and we remove device#0, the devices will be shifted down
[1][2]
and renamed.
[0][1]
This can create a problem for resume_lv (specifically,
dm_tree_activate_children) during the renaming process though. This
is because it will attempt to rename the higher indexed sub-LVs first
and find that it cannot because there are currently other sub-LVs with
that name. The solution is to check for a conflicting name before
attempting to rename. If a conflict is found and that conflicting
sub-LV is also in the process of renaming, we can defer the current
rename until the conflicting sub-LV has renamed and cleared the
conflict.
Now that resume_lv can handle these types of rename conflicts, we can
remove the workaround in RAID that was attempting to resume a RAID1
LV from the bottom-up in order to force a proper rename in assending
order before attempting a resume on the top-level LV. This "hack"
only worked for single machine use-cases of LVM. Clearing this up
paves the way for exclusive activation of RAID LVs in a cluster.
When resuming a node needed by a higher layer of the tree,
if the resume fails, only remove it if the node did not
originally have a live table.
Ref. 97f8454ecc
Clear send_messages flag when they have been delivered successfully.
There is no need to validate it for all other activations of the same
node in the dm_tree.
Also add extra debug message which shows the reason for skipping
sending of messages because the transaction_id has already the matching
value.
This patch adds the ability to set the minimum and maximum I/O rate for
sync operations in RAID LVs. The options are available for 'lvcreate' and
'lvchange' and are as follows:
--minrecoveryrate <Rate> [bBsSkKmMgG]
--maxrecoveryrate <Rate> [bBsSkKmMgG]
The rate is specified in size/sec/device. If a suffix is not given,
kiB/sec/device is assumed. Setting the rate to 0 removes the preference.
This patch may not be fully correct. It tries to solve
the imbalanced suspend counter.
The problem starts when some LV is created and fails in resume path.
(i.e. resuming to large PV (enforced) over small loop devices)
This fails in _resume_node() after dm_task_run(). And while
existing device with empty table is left in inactive table,
further calls are reporting this device is in suspend state.
When later the lvm2 tries to rollback created device and deactivate it,
it will end with internal error, when we try to decrement
never incremented suspend counter.
As an 'easy fix' for now update suspend counter only for live nodes.
TODO: explore better fix.
Since we use get_status also in dmeventd, which may use one pool
for a single device, in case it would be repeatedly returning error,
it may not be freeing the pool and would cause slow but steady growth.
To stay safe in the error path release any allocated memory.
'lvchange' is used to alter a RAID 1 logical volume's write-mostly and
write-behind characteristics. The '--writemostly' parameter takes a
PV as an argument with an optional trailing character to specify whether
to set ('y'), unset ('n'), or toggle ('t') the value. If no trailing
character is given, it will set the flag.
Synopsis:
lvchange [--writemostly <PV>:{t|y|n}] [--writebehind <count>] vg/lv
Example:
lvchange --writemostly /dev/sdb1:y --writebehind 512 vg/raid1_lv
The last character in the 'lv_attr' field is used to show whether a device
has the WriteMostly flag set. It is signified with a 'w'. If the device
has failed, the 'p'artial flag has priority.
Example ("nosync" raid1 with mismatch_cnt and writemostly):
[~]# lvs -a --segment vg
LV VG Attr #Str Type SSize
raid1 vg Rwi---r-m 2 raid1 500.00m
[raid1_rimage_0] vg Iwi---r-- 1 linear 500.00m
[raid1_rimage_1] vg Iwi---r-w 1 linear 500.00m
[raid1_rmeta_0] vg ewi---r-- 1 linear 4.00m
[raid1_rmeta_1] vg ewi---r-- 1 linear 4.00m
Example (raid1 with mismatch_cnt, writemostly - but failed drive):
[~]# lvs -a --segment vg
LV VG Attr #Str Type SSize
raid1 vg rwi---r-p 2 raid1 500.00m
[raid1_rimage_0] vg Iwi---r-- 1 linear 500.00m
[raid1_rimage_1] vg Iwi---r-p 1 linear 500.00m
[raid1_rmeta_0] vg ewi---r-- 1 linear 4.00m
[raid1_rmeta_1] vg ewi---r-p 1 linear 4.00m
A new reportable field has been added for writebehind as well. If
write-behind has not been set or the LV is not RAID1, the field will
be blank.
Example (writebehind is set):
[~]# lvs -a -o name,attr,writebehind vg
LV Attr WBehind
lv rwi-a-r-- 512
[lv_rimage_0] iwi-aor-w
[lv_rimage_1] iwi-aor--
[lv_rmeta_0] ewi-aor--
[lv_rmeta_1] ewi-aor--
Example (writebehind is not set):
[~]# lvs -a -o name,attr,writebehind vg
LV Attr WBehind
lv rwi-a-r--
[lv_rimage_0] iwi-aor-w
[lv_rimage_1] iwi-aor--
[lv_rmeta_0] ewi-aor--
[lv_rmeta_1] ewi-aor--
Revert commit 31c24dd9f2. This commit
was used to force a RAID device-mapper table to be loaded into the
kernel despite the fact that it was identical to the one already
loaded. The effect allowed a RAID array with a transiently failed
device to refresh and reintegrate the failed device. This operation
is better done in the kernel on a 'resume'. Since,
'lvchange --refresh' already performs a suspend/resume cycle, the
above commit is not needed once the kernel change is made. Reverting
the commit removes an unnecessary (at least for now) change to the
device-mapper interface.
I've updated the dm_status_raid structure and dm_get_status_raid()
function to make it handle the new kernel status fields that will
be coming in dm-raid v1.5.0. It is backwards compatible with the
old status line - initializing the new fields to '0'. The new
structure is also more amenable to future changes. It includes a
'reserved' field that is currently initialized to zero but could
be used to hold flags describing new features. It also now uses
pointers for the character strings instead of attempting to allocate
their space along with the structure (causing the size of the
structure to be variable). This allows future fields to be appended.
The new fields that are available are:
- sync_action : shows what the sync thread in the kernel is doing
(idle, frozen, resync, recover, check, repair, or
reshape)
- mismatch_count: shows the number of discrepancies which were
found or repaired by a "check" or "repair"
process, respectively.
Previous commit included changes to WHATSNEW, but the code changes
were missing. Here is the description from the previous commit:
commit bbc6378b73
Author: Jonathan Brassow <jbrassow@redhat.com>
Date: Thu Feb 21 11:31:36 2013 -0600
RAID: Make 'lvchange --refresh' restore transiently failed RAID PVs
A new function (dm_tree_node_force_identical_table_reload) was added to
avoid the suppression of identical table reloads. This allows RAID LVs
to reload the on-disk superblock information that contains which devices
have failed and the bitmaps. If the failed device has returned, this has
the effect of restoring the device and initiating recovery. Without this
patch, the user had to completely deactivate their RAID LV and re-activate
it in order to restore the failed device. Now they simply need to
suspend and resume (which is done by 'lvchange --refresh').
The identical table suppression is only avoided if the LV is not PARTAIL
(i.e. all of it's devices can be seen and read by LVM) and the kernel
status of the array contains failed devices. In other words, the function
will only be called in the case where we may have success in restoring
a failed device in the array.
Similar to the way thin* accesses its kernel status, we add a method
for RAID to grab the various values in its status output without the
higher levels (LVM) having to understand how to parse the output.
Added functions include:
- lib/activate/dev_manager.c:dev_manager_raid_status()
Pulls the status line from the kernel
- libdm/libdm-deptree.c:dm_get_status_raid()
Parses status line and puts components into dm_status_raid struct
- lib/activate/activate.c:lv_raid_dev_health()
Accesses dm_status_raid to deliver raid dev_health string
The new structure and functions can provide a more unified way to access
status information. ('lv_raid_percent' could switch to using these
functions, for example.)
If the resume of preloaded node fails, do not leave such
node in the table - since it may not be easy to detach such
node later when the node is i.e. internal.
i.e. failing activation of the thin pool with mismatching
chunk size may leave -tpool device in the table, which
could have been then removed only by dmsetup command.
Patch clears the flag if thin pool is stacked over mirror.
Since thin pool could be used to stack device over mirrors,
it needs resume properly i.e. mirrors with corelog which are otherwise
unconditionally skipped (for pvmove functionality).
This patch adds support for RAID10. It is not the default at this
stage. The user needs to specify '--type raid10' if they would like
RAID10 instead of stacked mirror over stripe.
if the thin_check fail on thin pool - still return successful deactivation,
since lvremove would currently fail.
TODO: find some way to not run check with lvremove.
