glusterfs.git/xlators/mgmt/glusterd/src/glusterd-utils.c, branch v3.5.1beta2

glusterd: On gaining quorum spawn_daemons in new thread

2014-06-08T14:50:18+00:00

  Backport of http://review.gluster.org/7703 from master

During startup, if a glusterd has peers, it waits till quorum is
obtained to spawn bricks and other services. If peers are not present,
the daemons are started during glusterd' startup itself.

The spawning of daemons as a quorum action was done without using a
seperate thread, unlike the spawn on startup. Since, quotad was launched
using the blocking runner_run api, this leads to the thread being
blocked. The calling thread is almost always the epoll thread and this
leads to a deadlock. The runner_run call blocks the epoll thread waiting
for quotad to start, as a result glusterd cannot serve any requests. But
the startup of quotad is blocked as it cannot fetch the volfile from
glusterd.

The fix for this is to launch the spawn daemons task in a seperate
thread. This will free up the epoll thread and prevents the above
deadlock from happening.

BUG: 1105188
Change-Id: Idad1e96fbe1411dfd4b1a542fb5fa115673636c0
Signed-off-by: Kaushal M 
Reviewed-on: http://review.gluster.org/7995
Tested-by: Gluster Build System 
Reviewed-by: Niels de Vos

mgmt/glusterd: make sure quota enforcer has established connection with quotad before marking quota as enabled.

2014-01-29T06:46:35+00:00

without this patch there is a window of time when quota is marked as
enabled in quota-enforcer, but connection to quotad wouldn't have been
established. Any checklimit done during this period can result in a
failed fop because of unavailability of quotad.

Change-Id: I0d509fabc434dd55ce9ec59157123524197fcc80
Signed-off-by: Raghavendra G 
BUG: 969461
Signed-off-by: Raghavendra G 
Reviewed-on: http://review.gluster.org/6572
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur 
Reviewed-on: http://review.gluster.org/6820

glusterd/geo-rep: more glusterd and cli fixes for geo-rep.

2014-01-27T17:44:33+00:00

    -> handle option validation cases in reset case.
    -> Creating valid conf path when glusterd restarts.
    -> Reading the gsyncd worker thread status and displaying it.
    -> Displaying status-detail per worker.
    -> Fetch checkpoint info in geo-rep status.
    -> use-tarssh value validation added.

misc: misc geo-rep fixes based on cluster, logrotate etc..
    -> cluster/dht: fix 'stime' getxattr getting overwritten.
    -> cluster/afr: return max of 'stime' values in subvol.
    -> geo-rep-logrotate: Sending SIGHUP to geo-rep auxiliary.
    -> cluster/dht: fix convoluted logic while aggregating.
    -> cluster/*: fix 'stime' min/max fetch logic.

Change-Id: I811acea0bbd6194797a3e55d89295d1ea021ac85
BUG: 1036552
Signed-off-by: Ajeet Jha 
Reviewed-on: http://review.gluster.org/6405
Tested-by: Gluster Build System 
Reviewed-by: Amar Tumballi 
Reviewed-by: Anand Avati 
Reviewed-on: http://review.gluster.org/6810
Reviewed-by: Vijay Bellur

glusterd: Save/restore/sync rebalance dict

2014-01-03T11:11:04+00:00

A dictionary was added to store additional information of a rebalance
process, like the bricks being removed in case of a rebalance started
by remove-brick. This dictionary wasn't being stored/restored or synced
during volume sync, leading to errors like a volume status command
failing. These issues have been fixed in this patch. The rebalance dict
is now stored/restored and also exported/imported during volume sync.

Also, this makes sure that the rebalance dict is only create on
remove-brick start. This adds a bricks decommissioned status to the
information imported/exported during volume sync.

BUG: 1040809
Change-Id: I46cee3e4e34a1f3266a20aac43368854594f01bc
Signed-off-by: Kaushal M 
Backport-of: http://review.gluster.org/6492
Reviewed-on: http://review.gluster.org/6524
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

glusterd: rebalance to ref volinfo before starting

2013-12-23T15:00:37+00:00

        Backport of http://review.gluster.org/6522

Change-Id: Ib316897dcbd0748bfb3bfcda186b9fe30c07f80f
BUG: 1038051
Signed-off-by: Krishnan Parthasarathi 
Reviewed-on: http://review.gluster.org/6570
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

glusterd: make volinfo a refcnt'ed object.

2013-12-23T15:00:24+00:00

        Backport of http://review.gluster.org/6521

Add glusterd_volinfo_remove(..) which removes @volinfo from the list
of volumes in the cluster and performs an unref on @volinfo

Change-Id: I5f546ca58f61bc334ab1bab4c51c4a21e1f66161
BUG: 1038051
Signed-off-by: Krishnan Parthasarathi 
Reviewed-on: http://review.gluster.org/6569
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

glusterd: ignore failure to stop a stopped service.

2013-12-23T14:59:09+00:00

        Backport of http://review.gluster.org/6525

kill(2) returns -1 with errno set to ESRCH when the pid of the process
being killed doesn't exist. Failing glusterd_brick_stop on a stopped
brick could result in volume-stop failing, in commit phase.
This fix prevents that from happening.

Change-Id: I00f46fa06e489a671efbb8e4119f545f8ccea329
BUG: 1038051
Signed-off-by: Krishnan Parthasarathi 
Reviewed-on: http://review.gluster.org/6568
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

glusterd: Fix stopping of stale rebalance processes

2013-12-23T14:58:28+00:00

        Backport of http://review.gluster.org/6531

Trying to stop rebalance process via RPC using the GD_SYNCOP macro,
could lead to glusterd crashing. In case of an implicit volume update,
which happens when a peer comes back up, the stop function would be
called in the epoll thread. This would lead to glusterd crashing as the
epoll thread doesn't have synctasks for the GD_SYNCOP macro to make use
of.

Instead of using the RPC method, we now terminate the rebalance process
by kill(). The rebalance process has been designed to be resistant to
interruption, so this will not lead to any data corruption.

Also, when checking for stale rebalance task, make sure that the old
task-id is not null.

Change-Id: I54dd93803954ee55316cc58b5877f38d1ebc40b9
BUG: 1044327
Signed-off-by: Kaushal M 
Signed-off-by: Krishnan Parthasarathi 
Reviewed-on: http://review.gluster.org/6567
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

rpc,glusterd: Use rpc_clnt notifyfn to cleanup mydata

2013-12-23T14:58:18+00:00

        Backport of http://review.gluster.org/5512

rpc:
- On a RPC_TRANSPORT_CLEANUP event, rpc_clnt_notify calls the registered
  notifyfn with a RPC_CLNT_DESTROY event. The notifyfn should properly
  cleanup the saved mydata on this event.
- Break the reconnect chain when an rpc client is disabled. This will
  prevent new disconnect events which can lead to crashes.

glusterd:
- Added support for RPC_CLNT_DESTROY in glusterd_brick_rpc_notify
- Use a common glusterd_rpc_clnt_unref() function throught glusterd in
  place of rpc_clnt_unref(). This function correctly gives up the
  big-lock before performing the unref.

Change-Id: I93230441c5089039643fc9f5632477ef1b695348
BUG: 962619
Signed-off-by: Kaushal M 
Signed-off-by: Krishnan Parthasarathi 
Reviewed-on: http://review.gluster.org/6566
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

glusterd: Save/restore/sync rebalance dict

2013-12-23T14:58:03+00:00

        Backport of http://review.gluster.org/6492

A dictionary was added to store additional information of a rebalance
process, like the bricks being removed in case of a rebalance started
by remove-brick. This dictionary wasn't being stored/restored or synced
during volume sync, leading to errors like a volume status command
failing. These issues have been fixed in this patch. The rebalance dict
is now stored/restored and also exported/imported during volume sync.

Also, this makes sure that the rebalance dict is only create on
remove-brick start. This adds a bricks decommissioned status to the
information imported/exported during volume sync.

Change-Id: I56fed23dc2de80a96648055fe705e9c3ffd55227
BUG: 1040809
Signed-off-by: Kaushal M 
Signed-off-by: Krishnan Parthasarathi 
Reviewed-on: http://review.gluster.org/6565
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur