Commit Graph

39 Commits

Author SHA1 Message Date
Lars Ellenberg 13d42685be drbd: add explicit drbd_md_sync to drbd_resync_finished
As we usually update the generation UUIDs here, we should explicitly
sync them to disk.  So far this has been done only implicitly by related
code paths.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 18:38:52 +02:00
Lars Ellenberg 22cc37a943 drbd: fix unlikely access after free and list corruption
Various cleanup paths have been incomplete, for the very unlikely case
that we cannot allocate enough bios from process context when submitting
on behalf of the peer or resync process.

Never observed.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 18:38:49 +02:00
Lars Ellenberg af85e8e83d drbd: fix for spurious fullsync (uuids rotated too fast)
If it was an "empty" resync, the SyncSource may have already "finished"
the resync and rotated the UUIDs, before noticing the connection loss
(and generating a new uuid, if Primary, rotating again), while the
SyncTarget did not change its uuids at all, or only got to the previous
sync-uuid.
This would then again lead to a full sync on next handshake
(see also Bug #251).

Fix:
Use explicit resync finished notification even for empty resyncs,
do not finish an empty resync implicitly on the SyncSource.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 18:38:48 +02:00
Philipp Reisner f10f262349 drbd: Fixed a stupid copy and paste error
This caused rs_planed to be not in sync with the content of the fifo.
That in turn could cause that the resync comes to a complete halt.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 18:38:43 +02:00
Lars Ellenberg 78db89287c drbd: DIV_ROUND_UP not needed here
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 18:38:39 +02:00
Philipp Reisner 5a75cc7cfb drbd: Fixed compatibility with protocol versions smaller than 95
Forgot to consider the max size for the resync requests.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 18:38:38 +02:00
Lars Ellenberg 1d53f09e17 drbd: fix potential kernel BUG (NULL deref)
BUG trace would look like:
 lc_find
 drbd_rs_complete_io
 got_OVResult
 drbd_asender

Could be triggered by explicit, or IO-error policy based,
detach during online-verify.

We may only dereference mdev->resync, if we first get_ldev(), as the
disk may break any time, causing mdev->resync to disappear once all
ldev references have been returned.
Already in flight online-verify requests or replies may still come in,
which we then need to ignore.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 18:38:34 +02:00
Lars Ellenberg 435f07402b drbd: don't count sendpage()d pages only referenced by tcp as in use
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 18:38:33 +02:00
Philipp Reisner 63106d3c6c drbd: Removed a race that could cause unexpected execution of w_make_resync_request()
The actual race happened int the drbd_start_resync() function. Where
drbd_resync_finished() -> __drbd_set_state() set STOP_SYNC_TIMER and
armed the timer.

If the timer fired before execution reaches the mod_timer statement
at the end of drbd_start_resync() the latter would cause an
unexpected call to w_make_resync_request().

Removed the STOP_SYNC_TIMER bit, and base it on the connection state.

The STOP_SYNC_TIMER bit probably originates probably the time before
the state engine.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 18:38:29 +02:00
Philipp Reisner 0778286a13 drbd: Disable activity log updates when the whole device is out of sync
When the complete device is marked as out of sync, we can disable
updates of the on disk AL. Currently AL updates are only disabled
if one uses the "invalidate-remote" command on an unconnected,
primary device, or when at attach time all bits in the bitmap are
set.

As of now, AL updated do not get disabled when a all bits becomes
set due to application writes to an unconnected DRBD device.
While this is a missing feature, it is not considered important,
and might get added later.

BTW, after initializing a "one legged" DRBD device
drbdadm create-md resX
drbdadm -- --force primary resX
AL updates also get disabled, until the first connect.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 18:38:26 +02:00
Philipp Reisner 0b70a13dac drbd: Sending of big packets, for payloads from 64KByte to 4GByte
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 18:38:23 +02:00
Philipp Reisner 204bba9965 drbd: Bugfix for regression introduced with f9bc8913c06022e
If we intent to use the block_id member of an epoch entry,
we may not use the digest member.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 18:38:22 +02:00
Lars Ellenberg 0f0601f4ea drbd: new configuration parameter c-min-rate
We now track the data rate of locally submitted resync related requests,
and can thus detect non-resync activity on the lower level device.

If the current sync rate is above c-min-rate, and the lower level device
appears to be busy, we throttle the resyncer.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 18:38:20 +02:00
Lars Ellenberg 80a40e439e drbd: reduce code duplication when receiving data requests
also canonicalize the return values of read_for_csum
and drbd_rs_begin_io to return -ESOMETHING, or 0 for success.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 18:38:19 +02:00
Lars Ellenberg 1d7734a0df drbd: use rolling marks for resync speed calculation
The current resync speed as displayed in /proc/drbd fluctuates a lot.
Using an array of rolling marks makes this calculation much more stable.
We used to have this (a long time ago with 0.7), but it got lost somehow.

If "stalled", do not discard the rest of the information, just add a
" (stalled)" tag to the progress line.

This patch also shortens a spinlock critical section somewhat, and
reduces the number of atomic operations in put_ldev.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 18:38:18 +02:00
Lars Ellenberg c36c3ced69 drbd: let drbd_free_ee implicitly free any digest
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 18:38:16 +02:00
Philipp Reisner 85719573dd drbd: Replaced some casts by an union. Improved comments
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 18:38:15 +02:00
Philipp Reisner d207450cf2 drbd: Bugfix: rs_in_flight could become wrong if read_for_csum() requested reschedule later
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 18:38:14 +02:00
Philipp Reisner 778f271dfe drbd: The new, smarter resync speed controller
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 18:38:14 +02:00
Philipp Reisner 8e26f9ccb9 drbd: New sync_param packet, that includes the parameters of the new controller
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 18:38:13 +02:00
Philipp Reisner 265be2d098 drbd: Finished the "on-no-data-accessible suspend-io;" functionality
When no data is accessible (no connection to the peer, nor a local disk)
allow the user to select to freeze all IO operations instead of getting
IO errors.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-10-14 14:52:53 +02:00
Lars Ellenberg e7f52dfb4f drbd: revert "delay probes", feature is being re-implemented differently
It was a now abandoned attempt to throttle resync bandwidth
based on the delay it causes on the bulk data socket.
It has no userbase yet, and has been disabled by
9173465ccb51c09cc3102a10af93e9f469a0af6f already.
This removes the now unused code.

The basic feature, namely using up "idle" bandwith
of network and disk IO subsystem, with minimal impact
to application IO, is being reimplemented differently.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2010-08-07 18:53:57 +02:00
Philipp Reisner 2a0ab2cd73 drbd: Reduce verbosity
The "Local READ/WRITE failed" messages are too verbose.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2010-06-01 11:12:27 +02:00
Lars Ellenberg d255e5ff5f drbd: fix hang on local read errors while disconnected
"canceled" w_read_retry_remote never completed, if they have been
canceled after drbd_disconnect connection teardown cleanup has already
run (or we are currently not connected anyways).

Fixed by not queueing a remote retry if we already know it won't work
(pdsk not uptodate), and cleanup ourselves on "cancel", in case we hit a
race with drbd_disconnect.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2010-06-01 11:12:27 +02:00
Philipp Reisner 32fa7e91f9 drbd: Removed the now empty w_io_error() function
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
2010-06-01 11:12:27 +02:00
Jens Axboe ee9a3607fb Merge branch 'master' into for-2.6.35
Conflicts:
	fs/ext3/fsync.c

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2010-05-21 21:27:26 +02:00
Philipp Reisner fc8ce1941d drbd: Fix: Do not detach, if a bio with a barrier fails
Introduced a few days ago:
  commit 45bb912bd5
  Author: Lars Ellenberg <lars.ellenberg@linbit.com>
  Date:   Fri May 14 17:10:48 2010 +0200

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2010-05-21 21:12:00 +02:00
Lars Ellenberg bb3d000cb9 drbd: allow resync requests to be larger than max_segment_size
this should allow for better background resync performance.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-05-18 02:02:36 +02:00
Lars Ellenberg 45bb912bd5 drbd: Allow drbd_epoch_entries to use multiple bios.
This should allow for better performance if the lower level IO stack
of the peers differs in limits exposed either via the queue,
or via some merge_bvec_fn.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-05-18 02:01:23 +02:00
Philipp Reisner cdd67a7460 drbd: Control the actual resync rate based on the queuing delay of data packets
In a setup with a high bandwidth and high latency network, eventually
involving deep queues in routers, it is beneficial to only fill those
queues up to an limited extend with resync data.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-05-18 01:25:47 +02:00
Lars Ellenberg 5c3c7e64bb drbd: don't expose failed local READ to upper layers
fix regression introduced in 8.3.3:
 commit a9b17323f2875f5d9b132c2b476a750bf44b10c7
 Author: Lars Ellenberg <lars.ellenberg@linbit.com>
 Date:   Wed Aug 12 15:18:33 2009 +0200

     out-of-spinlock completion of master bio

 : (bio_rw(bio) == READA)
    ? read_completed_with_error
    : read_ahead_completed_with_error;

is obviously not what was intended.

No one noticed because of
 * page-cache at work,
 * local RAIDs

Impact:
Failed local READs are not retried remotely,
but errored to upper layers, causing filesystems
to remount read-only, or worse.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-05-03 22:40:16 +02:00
Philipp Reisner d0c3f60f36 drbd: Make sure we do not send state updates during an empty resync [Bugz 271]
This is a race condition that existed for ages.
The previous commit reduces the window, this one closes it.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-03-11 16:10:40 +01:00
Philipp Reisner 309d1608cc drbd: Reduce the time an empty resync takes usually
This mitigates changes introduced with commit:
http://git.drbd.org/?p=drbd-8.3.git;a=commit;h=4b6803a3276652da3737

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-03-11 16:09:03 +01:00
Lars Ellenberg 676396d545 fix unit of rs_same_csums accounting
Depending on resync request size,
we need to account for more than one bit.

Impact: cosmetic

If SyncTarget reported correctly 100% equal checksums,
the SyncSource usually reported 12% equal checksums instead,
because it only counted requests, we typically do 32k resync requests,
and the bitmap granularity is still 4k.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2010-03-11 16:01:38 +01:00
Huang Weiyi 820cd61a28 drbd: remove unused #include <linux/version.h>
Remove unused #include <linux/version.h>('s) in
  drivers/block/drbd/drbd_main.c
  drivers/block/drbd/drbd_receiver.c
  drivers/block/drbd/drbd_worker.c

Signed-off-by: Huang Weiyi <weiyi.huang@gmail.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
2009-12-21 13:41:16 +01:00
Huang Weiyi 7b886f4f7a drbd: remove duplicated #include
Remove duplicated #include('s) in
  drivers/block/drbd/drbd_worker.c

Signed-off-by: Huang Weiyi <weiyi.huang@gmail.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
2009-12-21 13:41:11 +01:00
Jens Axboe 6a0afdf58d drbd: remove tracing bits
They should be reimplemented in the current scheme.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-10-01 21:17:58 +02:00
Lars Ellenberg ab8fafc2e1 dropping unneeded include autoconf.h
It is force-included on the gcc command line since at least 2.6.15.
Explicit include lines seem to break compilation now in certain configurations.

Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
Signed-off-by: Kamalesh Babulal <kamalesh@linux.vnet.ibm.com>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
2009-10-01 21:17:54 +02:00
Philipp Reisner b411b3637f The DRBD driver
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
2009-10-01 21:17:49 +02:00