When a write to one of the devices of a RAID5/6 fails, the failure is
recorded in the metadata of the other devices so that after a restart
the data on the failed drive wont be trusted even if that drive seems
to be working again (maybe a cable was unplugged).
Similarly when we record a bad-block in response to a write failure,
we must not let the write complete until the bad-block update is safe.
Currently there is no interlock between the write request completing
and the metadata update. So it is possible that the write will
complete, the app will confirm success in some way, and then the
machine will crash before the metadata update completes.
This is an extremely small hole for a racy to fit in, but it is
theoretically possible and so should be closed.
So:
- set MD_CHANGE_PENDING when requesting a metadata update for a
failed device, so we can know with certainty when it completes
- queue requests that completed when MD_CHANGE_PENDING is set to
only be processed after the metadata update completes
- call raid_end_bio_io() on bios in that queue when the time comes.
Signed-off-by: NeilBrown <neilb@suse.com>
When a write to one of the legs of a RAID10 fails, the failure is
recorded in the metadata of the other legs so that after a restart
the data on the failed drive wont be trusted even if that drive seems
to be working again (maybe a cable was unplugged).
Currently there is no interlock between the write request completing
and the metadata update. So it is possible that the write will
complete, the app will confirm success in some way, and then the
machine will crash before the metadata update completes.
This is an extremely small hole for a racy to fit in, but it is
theoretically possible and so should be closed.
So:
- set MD_CHANGE_PENDING when requesting a metadata update for a
failed device, so we can know with certainty when it completes
- queue requests that experienced an error on a new queue which
is only processed after the metadata update completes
- call raid_end_bio_io() on bios in that queue when the time comes.
Signed-off-by: NeilBrown <neilb@suse.com>
When a write to one of the legs of a RAID1 fails, the failure is
recorded in the metadata of the other leg(s) so that after a restart
the data on the failed drive wont be trusted even if that drive seems
to be working again (maybe a cable was unplugged).
Similarly when we record a bad-block in response to a write failure,
we must not let the write complete until the bad-block update is safe.
Currently there is no interlock between the write request completing
and the metadata update. So it is possible that the write will
complete, the app will confirm success in some way, and then the
machine will crash before the metadata update completes.
This is an extremely small hole for a racy to fit in, but it is
theoretically possible and so should be closed.
So:
- set MD_CHANGE_PENDING when requesting a metadata update for a
failed device, so we can know with certainty when it completes
- queue requests that experienced an error on a new queue which
is only processed after the metadata update completes
- call raid_end_bio_io() on bios in that queue when the time comes.
Signed-off-by: NeilBrown <neilb@suse.com>
md_setup_cluster already calls try_module_get(), so this
try_module_get isn't needed.
Also, there is no matching module_put (except in error patch),
so this leaves an unbalanced module count.
Signed-off-by: NeilBrown <neilb@suse.com>
This code looks racy.
The only possible race is if two modules try to register at the same
time and that won't happen. But make the code look safe anyway.
Signed-off-by: NeilBrown <neilb@suse.com>
In gather_all_resync_info, we need to read the disk bitmap sb and
check if it needs recovery.
Reviewed-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: Guoqing Jiang <gqjiang@suse.com>
Signed-off-by: NeilBrown <neilb@suse.com>
Introduce MD_CLUSTER_BEGIN_JOIN_CLUSTER flag to make sure
complete(&cinfo->completion) is only be invoked when node
join cluster. Otherwise node failure could also call the
complete, and it doesn't make sense to do it.
Reviewed-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: Guoqing Jiang <gqjiang@suse.com>
Signed-off-by: NeilBrown <neilb@suse.com>
We also need to free the lock resource before goto out.
Reviewed-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: Guoqing Jiang <gqjiang@suse.com>
Signed-off-by: NeilBrown <neilb@suse.com>
The sb_lock is not used anywhere, so let's remove it.
Reviewed-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: Guoqing Jiang <gqjiang@suse.com>
Signed-off-by: NeilBrown <neilb@suse.com>
If the node just join the cluster, and receive the msg from other nodes
before init suspend_list, it will cause kernel crash due to NULL pointer
dereference, so move the initializations early to fix the bug.
md-cluster: Joined cluster 3578507b-e0cb-6d4f-6322-696cd7b1b10c slot 3
BUG: unable to handle kernel NULL pointer dereference at (null)
... ... ...
Call Trace:
[<ffffffffa0444924>] process_recvd_msg+0x2e4/0x330 [md_cluster]
[<ffffffffa0444a06>] recv_daemon+0x96/0x170 [md_cluster]
[<ffffffffa045189d>] md_thread+0x11d/0x170 [md_mod]
[<ffffffff810768c4>] kthread+0xb4/0xc0
[<ffffffff8151927c>] ret_from_fork+0x7c/0xb0
... ... ...
RIP [<ffffffffa0443581>] __remove_suspend_info+0x11/0xa0 [md_cluster]
Reviewed-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: Guoqing Jiang <gqjiang@suse.com>
Signed-off-by: NeilBrown <neilb@suse.com>
In complicated cluster environment, it is possible that the
dlm lock couldn't be get/convert on purpose, the related err
info is added for better debug potential issue.
For lockres_free, if the lock is blocking by a lock request or
conversion request, then dlm_unlock just put it back to grant
queue, so need to ensure the lock is free finally.
Signed-off-by: Guoqing Jiang <gqjiang@suse.com>
Signed-off-by: NeilBrown <neilb@suse.com>
We should init completion within lockres_init, otherwise
completion could be initialized more than one time during
it's life cycle.
Reviewed-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: Guoqing Jiang <gqjiang@suse.com>
Signed-off-by: NeilBrown <neilb@suse.com>
There is problem with previous communication mechanism, and we got below
deadlock scenario with cluster which has 3 nodes.
Sender Receiver Receiver
token(EX)
message(EX)
writes message
downconverts message(CR)
requests ack(EX)
get message(CR) gets message(CR)
reads message reads message
requests EX on message requests EX on message
To fix this problem, we do the following changes:
1. the sender downconverts MESSAGE to CW rather than CR.
2. and the receiver request PR lock not EX lock on message.
And in case we failed to down-convert EX to CW on message, it is better to
unlock message otherthan still hold the lock.
Reviewed-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: Lidong Zhong <ldzhong@suse.com>
Signed-off-by: Guoqing Jiang <gqjiang@suse.com>
Signed-off-by: NeilBrown <neilb@suse.com>
When node A stops an array while the array is doing a resync, we need
to let another node B take over the resync task.
To achieve the goal, we need the A send an explicit BITMAP_NEEDS_SYNC
message to the cluster. And the node B which received that message will
invoke __recover_slot to do resync.
Reviewed-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: Guoqing Jiang <gqjiang@suse.com>
Signed-off-by: NeilBrown <neilb@suse.com>
Make recover_slot as a wraper to __recover_slot, since the
logic of __recover_slot can be reused for the condition
when other nodes need to take over the resync job.
Signed-off-by: Guoqing Jiang <gqjiang@suse.com>
Signed-off-by: NeilBrown <neilb@suse.com>
We used to set up the safemode_timer timer in md_run. If md_run
would fail before the timer was set up we'd end up trying to modify
a timer that doesn't have a callback function when we access safe_delay_store,
which would trigger a BUG.
neilb: delete init_timer() call as setup_timer() does that.
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: NeilBrown <neilb@suse.com>
It is possible (though unlikely) for a reshape to be
interrupted between the time that end_reshape is called
and the time when raid5_finish_reshape is called.
This can leave conf->reshape_progress set to MaxSector,
but mddev->reshape_position not.
This combination confused reshape_request() when ->reshape_backwards.
As conf->reshape_progress is so high, it seems the reshape hasn't
really begun. But assuming MaxSector is a valid address only
leads to sorrow.
So ensure reshape_position and reshape_progress both agree,
and add an extra check in reshape_request() just in case they don't.
Signed-off-by: NeilBrown <neilb@suse.com>
There can be a small window between the moment that recovery
actually writes the last block and the time when various sysfs
and /proc/mdstat attributes report that it has finished.
During this time, 'sync_completed' can have the wrong value.
This can confuse monitoring software.
So:
- don't set curr_resync_completed beyond the end of the devices,
- set it correctly when resync/recovery has completed.
Signed-off-by: NeilBrown <neilb@suse.com>
While it generally shouldn't happen, it is not impossible for
curr_resync_completed to exceed resync_max.
This can particularly happen when reshaping RAID5 - the current
status isn't copied to curr_resync_completed promptly, so when it
is, it can exceed resync_max.
This happens when the reshape is 'frozen', resync_max is set low,
and reshape is re-enabled.
Taking a difference between two unsigned numbers is always dangerous
anyway, so add a test to behave correctly if
curr_resync_completed > resync_max
Signed-off-by: NeilBrown <neilb@suse.com>
This ensures that 'sync_action' will show 'recover' immediately the
array is started. If there is no spare the status will change to
'idle' once that is detected.
Clear MD_RECOVERY_RECOVER for a read-only array to ensure this change
happens.
This allows scripts which monitor status not to get confused -
particularly my test scripts.
Signed-off-by: NeilBrown <neilb@suse.com>
This code is calculating:
writepos, which is the furthest along address (device-space) that we
*will* be writing to
readpos, which is the earliest address that we *could* possible read
from, and
safepos, which is the earliest address in the 'old' section that we
might read from after a crash when the reshape position is
recovered from metadata.
The first is a precise calculation, so clipping at zero doesn't
make sense. As the reshape position is now guaranteed to always be
a multiple of reshape_sectors and as we already BUG_ON when
reshape_progress is zero, there is no point in this min_t() call.
The readpos and safepos are worst case - actual value depends on
precise geometry. That worst case could be negative, which is only
a problem because we are storing the value in an unsigned.
So leave the min_t() for those.
Signed-off-by: NeilBrown <neilb@suse.com>
When reshaping, we work in units of the largest chunk size.
If changing from a larger to a smaller chunk size, that means we
reshape more than one stripe at a time. So the required alignment
of reshape_position needs to take into account both the old
and new chunk size.
This means that both 'here_new' and 'here_old' are calculated with
respect to the same (maximum) chunk size, so testing if they are the
same when delta_disks is zero becomes pointless.
Signed-off-by: NeilBrown <neilb@suse.com>
The chunk_sectors and new_chunk_sectors fields of mddev can be changed
any time (via sysfs) that the reconfig mutex can be taken. So raid5
keeps internal copies in 'conf' which are stable except for a short
locked moment when reshape stops/starts.
So any access that does not hold reconfig_mutex should use the 'conf'
values, not the 'mddev' values.
Several don't.
This could result in corruption if new values were written at awkward
times.
Also use min() or max() rather than open-coding.
Signed-off-by: NeilBrown <neilb@suse.com>
NeilBrown
Guoqing Jiang
Sasha Levin