glusterfs.git/xlators/mgmt/glusterd/src/glusterd-sm.c, branch v3.10.2

tier : Tier as a service

2017-01-17T04:49:47+00:00

tierd is implemented by separating from rebalance process.

The commands affected:

1) Attach tier will trigger this process instead of old one
2) tier start and tier start force will also trigger this process.
3) volume status [tier] will show tier daemon as a process instead
of task and normal tier status and tier detach status works.
4) tier stop implemented.
5) detach tier implemented separately along with new detach tier
status
6) volume tier volname status will work using the changes.
7) volume set works

This patch has separated the tier translator from the legacy
DHT rebalance code. It now sends the RPCs from the CLI
to glusterd separate to the DHT rebalance code.
The daemon is now a service, similar to the snapshot daemon,
and can be viewed using the volume status command.

The code for the validation and commit phase are the same
as the earlier tier validation code in DHT rebalance.

The “brickop” phase has been changed so that the status
command can use this framework.

The service management framework is now used.
DHT rebalance does not use this framework.

This service framework takes care of :

*) spawning the daemon, killing it and other such processes.
*) volume set options , which are written on the volfile.
*) restart and reconfigure functions. Restart is to restart
the daemon at two points
        1)after gluster goes down and comes up.
        2) to stop detach tier.
*) reconfigure is used to make immediate volfile changes.
By doing this, we don’t restart the daemon.
it has the code to rewrite the volfile for topological
changes too (which comes into place during add and remove brick).

With this patch the log, pid, and volfile are separated
and put into respective directories.

Change-Id: I3681d0d66894714b55aa02ca2a30ac000362a399
BUG: 1313838
Signed-off-by: hari gowtham 
Reviewed-on: http://review.gluster.org/13365
Smoke: Gluster Build System 
Tested-by: hari gowtham 
CentOS-regression: Gluster Build System 
NetBSD-regression: NetBSD Build System 
Reviewed-by: Dan Lambright 
Reviewed-by: Atin Mukherjee

glusterd: Add a new event to handle multi-net probes

2016-03-29T04:43:35+00:00

This allows GlusterD to send updates to all other nodes when attaching
new addresses using multi-net peer probe.

Change-Id: I62846be750ab3721912e7b49656594347ea61723
BUG: 1320458
Signed-off-by: Kaushal M 
Reviewed-on: http://review.gluster.org/13817
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Atin Mukherjee

glusterd: fixing few memory leak in glusterd

2016-02-24T04:19:05+00:00

Current glusterd code base having memory leak. This is because of
memory allocate by dict_allocate_and_serialize function in
"gd_syncop_mgmt_v3_lock" and "gd_syncop_mgmt_v3_unlock"
function is not freeing up meory upon exit.

Fix is to free the memory after exit of the above function.

Thanx Carlos and Roman for finding out the issue and fix.

Change-Id: Id67aa794c84969830ca7ea8c2374f80c64d7a639
BUG: 1287517
Signed-off-by: Gaurav Kumar Garg 
Signed-off-by: Carlos Chinea 
Signed-off-by: Roman Tereshonkov 
Reviewed-on: http://review.gluster.org/12927
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Jeff Darcy

glusterd: reduce friend update flood

2015-12-23T03:52:19+00:00

When in a befriended state, glusterd would broadcast friend updates to
all other peers whenver a ACC or LOCAL_ACC event occurred.

When a downed glusterd came back up and established connections again,
this lead to a flood of friend updates to happen on the order of N^2 (N
is the number of peers in the cluster)

In larger clusters this was problematic, and could lead to very long
times for the cluster to settle down when a peer came back up. Multiple
peers coming back up at the same time would compound the problem.

Broadcasting of friend updates doesn't have much use in places other
that during a peer probe. Instead of broadcasting friend updates on
connection re-establishment, updates can just be exchanged between the
peers involved in the connection.

This patch changes the glusterd friend state-machine to send updates
only to the required peer for ACC or LOCAL_ACC events when in befriended
state. The number of updates sent now is in the order of N.

For a 10 node cluster, the number of updates reduced by 5 times. When
creating the 10 node cluster, the updates reduced from ~500 to ~150.
When a glusterd restarted, the number of exchanges reduced from ~160 to
~35.

BUG: 1292749
Change-Id: Ib6072090c7069b081d018cdaa3dc878819ab1d18
Signed-off-by: Kaushal M 
Reviewed-on: http://review.gluster.org/12999
Reviewed-by: Atin Mukherjee 
Tested-by: NetBSD Build System 
Tested-by: Gluster Build System

glusterd: stop daemon services upon peer detach correctly

2015-12-03T11:28:04+00:00

Problem:
Currently glusterd is stopping all the daemons service upon peer detach.
If user have multi node cluster and if user want to detach any node
from the cluster, and detached node having stand alone volume then upon
detaching glusterd stopping all the daemon's of the detached node,
which is having running volume.

Fix:
Upon peer detach it should do peer detach cleanup properly and it should
stop only those daemon on the node on which it require.

Change-Id: I98b8099166f82e235ded6d02261f59a6511a003b
BUG: 1287455
Signed-off-by: Gaurav Kumar Garg 
Reviewed-on: http://review.gluster.org/12838
Reviewed-by: Atin Mukherjee 
Tested-by: NetBSD Build System 
Tested-by: Gluster Build System

snapshot:cleanup snaps during unprobe

2015-08-26T10:47:48+00:00

When doing an unprobe, the volume that doesnot
contain any brick of the particular node will be
deleted. So the snaps associated with that volume
should also delete

Change-Id: I9f3d23bd11b254ebf7d7722cc1e12455d6b024ff
BUG: 1203185
Signed-off-by: Mohammed Rafi KC 
Reviewed-on: http://review.gluster.org/9930
Tested-by: NetBSD Build System 
Reviewed-by: Avra Sengupta 
Reviewed-by: Rajesh Joseph

glusterd: stop all the daemons services on peer detach

2015-08-25T05:18:55+00:00

Currently glusterd is not stopping all the deamon service on peer detach

With this fix it will do peer detach cleanup properlly and will stop all
the daemon which was running before peer detach on the node.

Change-Id: Ifed403ed09187e84f2a60bf63135156ad1f15775
BUG: 1255386
Signed-off-by: Gaurav Kumar Garg 
Reviewed-on: http://review.gluster.org/11509
Tested-by: NetBSD Build System 
Tested-by: Gluster Build System 
Reviewed-by: Atin Mukherjee

glusterd: Send friend update even for EVENT_RCVD_ACC

2015-07-11T07:18:09+00:00

In a multi-network cluster, a new peer being probed into the cluster
will not get all the addresses of the peer that initiated the probe in
some cases as it doesn't recieve friend updates from other peers in the
cluster.

This happens when the new peer establishes connection with the other
peers before the other peers connect to the new peer.

Assuming, F is the initiator peer, O is one of the other peers in the
cluster and N is the new peer, the following series of events occur on O
when N establishes the connection first. N is already in the BEFRIENDED
state on O, and the actions happening will refer the BEFRIENDED state
table.

  EVENT_RCV_FRIEND_REQ -> results in handle_friend_add_req being called
                          which injects a LOCAL_ACC
  EVENT_RCVC_LOCAL_ACC -> results in send_friend_update being called
                          which should have sent an update to N, but O
                          has still not established a connection to N,
                          so the update isn't sent
  EVENT_CONNECTED      -> O now connects to N, this results in O sending a
                          friend_add req to N
  EVENT_RCVD_ACC       -> friend_add_cbk inject this event, but this event
                          results in a NOOP when in BEFRIENDED

As a result this O doesn't recieve all the addresses of F. If the
cluster contains any volumes with bricks attached to the missing
addresses of F and O is restarted in this condition, GlusterD will fail
to start as it wouldn't be able to resolve those bricks.

This commit changes the EVENT_RCVD_ACC action for the BEFRIENDED state
from a NOOP to send_friend_update. This makes sure that the new peer
recieves the updates from the other existing peers, irrespective of who
establishes the connection first, thus solving the problem.

Change-Id: Id807bc3032cf4cb13a5ba83819f2d50c96e76e96
BUG: 1241882
Signed-off-by: Kaushal M 
Reviewed-on: http://review.gluster.org/11625
Tested-by: NetBSD Build System 
Reviewed-by: Krishnan Parthasarathi

glusterd: Porting left out log messages to new framework

2015-06-27T06:32:01+00:00

Change-Id: I70d40ae3b5f49a21e1b93f82885cd58fa2723647
BUG: 1235538
Signed-off-by: Nandaja Varma 
Reviewed-on: http://review.gluster.org/11388
Tested-by: NetBSD Build System 
Reviewed-by: Anand Nekkunti 
Tested-by: Gluster Build System 
Reviewed-by: Kaushal M

sm/glusterd: Porting messages to new logging framework

2015-06-12T09:05:29+00:00

Change-Id: I391d1ac6a7b312461187c2e8c6f14d09a0238950
BUG: 1194640
Signed-off-by: Nandaja Varma 
Reviewed-on: http://review.gluster.org/9927
Reviewed-by: Atin Mukherjee 
Tested-by: NetBSD Build System 
Tested-by: Gluster Build System 
Reviewed-by: Krishnan Parthasarathi