[Bugs] [Bug 1259078] New: should not spawn another migration daemon on graph switch

[bugzilla at redhat.com]

https://bugzilla.redhat.com/show_bug.cgi?id=1259078

            Bug ID: 1259078
           Summary: should not spawn another migration daemon on graph
                    switch
           Product: GlusterFS
           Version: 3.7.5
         Component: tiering
          Severity: high
          Priority: high
          Assignee: bugs at gluster.org
          Reporter: dlambrig at redhat.com
        QA Contact: bugs at gluster.org
                CC: bugs at gluster.org
        Depends On: 1226005
            Blocks: 1227469



+++ This bug was initially created as a clone of Bug #1226005 +++

When we did a graph switch on a rebalance daemon, a second call to
gf_degrag_start() was done. This lead to multiple threads doing migration. When
multiple threads try to move the same file there can be deadlocks.

--- Additional comment from Anand Avati on 2015-05-28 14:01:39 EDT ---

REVIEW: http://review.gluster.org/10977 (When we did a graph switch on a
rebalance daemon, a second call to gf_degrag_start() was done. This lead to
multiple threads doing migration. When multiple threads try to move the same
file there can be deadlocks.) posted (#1) for review on master by Dan Lambright
(dlambrig at redhat.com)

--- Additional comment from Anand Avati on 2015-05-28 17:29:10 EDT ---

REVIEW: http://review.gluster.org/10977 (When we did a graph switch on a
rebalance daemon, a second call to gf_degrag_start() was done. This lead to
multiple threads doing migration. When multiple threads try to move the same
file there can be deadlocks.) posted (#3) for review on master by Dan Lambright
(dlambrig at redhat.com)

--- Additional comment from Anand Avati on 2015-05-29 15:56:51 EDT ---

REVIEW: http://review.gluster.org/10977 (When we did a graph switch on a
rebalance daemon, a second call to gf_degrag_start() was done. This lead to
multiple threads doing migration. When multiple threads try to move the same
file there can be deadlocks.) posted (#4) for review on master by Dan Lambright
(dlambrig at redhat.com)

--- Additional comment from Anand Avati on 2015-05-30 06:34:53 EDT ---

REVIEW: http://review.gluster.org/10977 (When we did a graph switch on a
rebalance daemon, a second call to gf_degrag_start() was done. This lead to
multiple threads doing migration. When multiple threads try to move the same
file there can be deadlocks.) posted (#5) for review on master by Niels de Vos
(ndevos at redhat.com)

--- Additional comment from Anand Avati on 2015-05-30 11:59:50 EDT ---

REVIEW: http://review.gluster.org/10977 (cluster/dht: maintain start state of
rebalance daemon across graph switch.) posted (#6) for review on master by Dan
Lambright (dlambrig at redhat.com)

--- Additional comment from Anand Avati on 2015-06-01 14:12:41 EDT ---

COMMIT: http://review.gluster.org/10977 committed in master by Vijay Bellur
(vbellur at redhat.com) 
------
commit 3f11b8e8ec6d78ebe33636b64130d5d133729f2c
Author: Dan Lambright <dlambrig at redhat.com>
Date:   Thu May 28 14:00:37 2015 -0400

    cluster/dht: maintain start state of rebalance daemon across graph switch.

    When we did a graph switch on a rebalance daemon, a second call
    to gf_degrag_start() was done. This lead to multiple threads
    doing migration. When multiple threads try to move the same
    file there can be deadlocks.

    Change-Id: I931ca7fe600022f245e3dccaabb1ad004f732c56
    BUG: 1226005
    Signed-off-by: Dan Lambright <dlambrig at redhat.com>
    Reviewed-on: http://review.gluster.org/10977
    Tested-by: NetBSD Build System <jenkins at build.gluster.org>
    Reviewed-by: Shyamsundar Ranganathan <srangana at redhat.com>

--- Additional comment from Anand Avati on 2015-06-26 19:22:05 EDT ---

REVIEW: http://review.gluster.org/11372 (cluster/tier: stop tier migration
after graph switch) posted (#10) for review on master by Dan Lambright
(dlambrig at redhat.com)

--- Additional comment from Anand Avati on 2015-06-26 19:36:02 EDT ---

REVIEW: http://review.gluster.org/11372 (cluster/tier: stop tier migration
after graph switch) posted (#11) for review on master by Dan Lambright
(dlambrig at redhat.com)

--- Additional comment from Anand Avati on 2015-06-26 20:11:22 EDT ---

COMMIT: http://review.gluster.org/11372 committed in master by Dan Lambright
(dlambrig at redhat.com) 
------
commit 875aa01ec80e56d85d0bc6028c6f1417f6ab140f
Author: Dan Lambright <dlambrig at redhat.com>
Date:   Tue Jun 23 16:35:03 2015 -0400

    cluster/tier: stop tier migration after graph switch

    On a graph switch, a new xlator and private structures are
    created. The tier migration daemon must stop using the
    old xlator and private structures and begin using the
    new ones. Otherwise, when RPCs arrive (such as counter
    queries from glusterd), the new xlator will be consulted
    but it will not have up to date information. The fix
    detects a graph switch and exits the daemon in this
    case. Typical graph switches for the tier case would
    be turning off performance translators.

    Change-Id: Ibfbd4720dc82ea179b77c81b8f534abced21e3c8
    BUG: 1226005
    Signed-off-by: Dan Lambright <dlambrig at redhat.com>
    Reviewed-on: http://review.gluster.org/11372


Referenced Bugs:

https://bugzilla.redhat.com/show_bug.cgi?id=1226005
[Bug 1226005] should not spawn another migration daemon on graph switch
https://bugzilla.redhat.com/show_bug.cgi?id=1227469
[Bug 1227469] should not spawn another migration daemon on graph switch
-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are on the CC list for the bug.
You are the assignee for the bug.