[Pacemaker] why pacemaker does not control the resources

Andrey Groshev greenx at yandex.ru
Thu Nov 14 06:06:04 UTC 2013



14.11.2013, 02:22, "Andrew Beekhof" <andrew at beekhof.net>:
> On 14 Nov 2013, at 6:13 am, Andrey Groshev <greenx at yandex.ru> wrote:
>
>>  13.11.2013, 03:22, "Andrew Beekhof" <andrew at beekhof.net>:
>>>  On 12 Nov 2013, at 4:42 pm, Andrey Groshev <greenx at yandex.ru> wrote:
>>>>   11.11.2013, 03:44, "Andrew Beekhof" <andrew at beekhof.net>:
>>>>>   On 8 Nov 2013, at 7:49 am, Andrey Groshev <greenx at yandex.ru> wrote:
>>>>>>    Hi, PPL!
>>>>>>    I need help. I do not understand... Why has stopped working.
>>>>>>    This configuration work on other cluster, but on corosync1.
>>>>>>
>>>>>>    So... cluster postgres with master/slave.
>>>>>>    Classic config as in wiki.
>>>>>>    I build cluster, start, he is working.
>>>>>>    Next I kill postgres on Master with 6 signal, as if "disk space left"
>>>>>>
>>>>>>    # pkill -6 postgres
>>>>>>    # ps axuww|grep postgres
>>>>>>    root      9032  0.0  0.1 103236   860 pts/0    S+   00:37   0:00 grep postgres
>>>>>>
>>>>>>    PostgreSQL die, But crm_mon shows that the master is still running.
>>>>>>
>>>>>>    Last updated: Fri Nov  8 00:42:08 2013
>>>>>>    Last change: Fri Nov  8 00:37:05 2013 via crm_attribute on dev-cluster2-node4
>>>>>>    Stack: corosync
>>>>>>    Current DC: dev-cluster2-node4 (172793107) - partition with quorum
>>>>>>    Version: 1.1.10-1.el6-368c726
>>>>>>    3 Nodes configured
>>>>>>    7 Resources configured
>>>>>>
>>>>>>    Node dev-cluster2-node2 (172793105): online
>>>>>>           pingCheck       (ocf::pacemaker:ping):  Started
>>>>>>           pgsql   (ocf::heartbeat:pgsql): Started
>>>>>>    Node dev-cluster2-node3 (172793106): online
>>>>>>           pingCheck       (ocf::pacemaker:ping):  Started
>>>>>>           pgsql   (ocf::heartbeat:pgsql): Started
>>>>>>    Node dev-cluster2-node4 (172793107): online
>>>>>>           pgsql   (ocf::heartbeat:pgsql): Master
>>>>>>           pingCheck       (ocf::pacemaker:ping):  Started
>>>>>>           VirtualIP       (ocf::heartbeat:IPaddr2):       Started
>>>>>>
>>>>>>    Node Attributes:
>>>>>>    * Node dev-cluster2-node2:
>>>>>>       + default_ping_set                  : 100
>>>>>>       + master-pgsql                      : -INFINITY
>>>>>>       + pgsql-data-status                 : STREAMING|ASYNC
>>>>>>       + pgsql-status                      : HS:async
>>>>>>    * Node dev-cluster2-node3:
>>>>>>       + default_ping_set                  : 100
>>>>>>       + master-pgsql                      : -INFINITY
>>>>>>       + pgsql-data-status                 : STREAMING|ASYNC
>>>>>>       + pgsql-status                      : HS:async
>>>>>>    * Node dev-cluster2-node4:
>>>>>>       + default_ping_set                  : 100
>>>>>>       + master-pgsql                      : 1000
>>>>>>       + pgsql-data-status                 : LATEST
>>>>>>       + pgsql-master-baseline             : 0000000002000078
>>>>>>       + pgsql-status                      : PRI
>>>>>>
>>>>>>    Migration summary:
>>>>>>    * Node dev-cluster2-node4:
>>>>>>    * Node dev-cluster2-node2:
>>>>>>    * Node dev-cluster2-node3:
>>>>>>
>>>>>>    Tickets:
>>>>>>
>>>>>>    CONFIG:
>>>>>>    node $id="172793105" dev-cluster2-node2. \
>>>>>>           attributes pgsql-data-status="STREAMING|ASYNC" standby="false"
>>>>>>    node $id="172793106" dev-cluster2-node3. \
>>>>>>           attributes pgsql-data-status="STREAMING|ASYNC" standby="false"
>>>>>>    node $id="172793107" dev-cluster2-node4. \
>>>>>>           attributes pgsql-data-status="LATEST"
>>>>>>    primitive VirtualIP ocf:heartbeat:IPaddr2 \
>>>>>>           params ip="10.76.157.194" \
>>>>>>           op start interval="0" timeout="60s" on-fail="stop" \
>>>>>>           op monitor interval="10s" timeout="60s" on-fail="restart" \
>>>>>>           op stop interval="0" timeout="60s" on-fail="block"
>>>>>>    primitive pgsql ocf:heartbeat:pgsql \
>>>>>>           params pgctl="/usr/pgsql-9.1/bin/pg_ctl" psql="/usr/pgsql-9.1/bin/psql" pgdata="/var/lib/pgsql/9.1/data" tmpdir="/tmp/pg" start_opt="-p 5432" logfile="/var/lib/pgsql/9.1//pgstartup.log" rep_mode="async" node_list=" dev-cluster2-node2. dev-cluster2-node3. dev-cluster2-node4. " restore_command="gzip -cd /var/backup/pitr/dev-cluster2-master#5432/xlog/%f.gz > %p" primary_conninfo_opt="keepalives_idle=60 keepalives_interval=5 keepalives_count=5" master_ip="10.76.157.194" \
>>>>>>           op start interval="0" timeout="60s" on-fail="restart" \
>>>>>>           op monitor interval="5s" timeout="61s" on-fail="restart" \
>>>>>>           op monitor interval="1s" role="Master" timeout="62s" on-fail="restart" \
>>>>>>           op promote interval="0" timeout="63s" on-fail="restart" \
>>>>>>           op demote interval="0" timeout="64s" on-fail="stop" \
>>>>>>           op stop interval="0" timeout="65s" on-fail="block" \
>>>>>>           op notify interval="0" timeout="66s"
>>>>>>    primitive pingCheck ocf:pacemaker:ping \
>>>>>>           params name="default_ping_set" host_list="10.76.156.1" multiplier="100" \
>>>>>>           op start interval="0" timeout="60s" on-fail="restart" \
>>>>>>           op monitor interval="10s" timeout="60s" on-fail="restart" \
>>>>>>           op stop interval="0" timeout="60s" on-fail="ignore"
>>>>>>    ms msPostgresql pgsql \
>>>>>>           meta master-max="1" master-node-max="1" clone-node-max="1" notify="true" target-role="Master" clone-max="3"
>>>>>>    clone clnPingCheck pingCheck \
>>>>>>           meta clone-max="3"
>>>>>>    location l0_DontRunPgIfNotPingGW msPostgresql \
>>>>>>           rule $id="l0_DontRunPgIfNotPingGW-rule" -inf: not_defined default_ping_set or default_ping_set lt 100
>>>>>>    colocation r0_StartPgIfPingGW inf: msPostgresql clnPingCheck
>>>>>>    colocation r1_MastersGroup inf: VirtualIP msPostgresql:Master
>>>>>>    order rsc_order-1 0: clnPingCheck msPostgresql
>>>>>>    order rsc_order-2 0: msPostgresql:promote VirtualIP:start symmetrical=false
>>>>>>    order rsc_order-3 0: msPostgresql:demote VirtualIP:stop symmetrical=false
>>>>>>    property $id="cib-bootstrap-options" \
>>>>>>           dc-version="1.1.10-1.el6-368c726" \
>>>>>>           cluster-infrastructure="corosync" \
>>>>>>           stonith-enabled="false" \
>>>>>>           no-quorum-policy="stop"
>>>>>>    rsc_defaults $id="rsc-options" \
>>>>>>           resource-stickiness="INFINITY" \
>>>>>>           migration-threshold="1"
>>>>>>
>>>>>>    Tell me where to look - why does pacemaker not work?
>>>>>   You might want to follow some of the steps at:
>>>>>
>>>>>      http://blog.clusterlabs.org/blog/2013/debugging-pacemaker/
>>>>>
>>>>>   under the heading "Resource-level failures".
>>>>   Yes. Thank you.
>>>>   I've seen this article and now I study it in more detail.
>>>>   A lot of information in the logs, so it is difficult to determine where the error is, and where the consequence of error.
>>>>   Now I'm trying to figure it out.
>>>>
>>>>   BUT...
>>>>   While I can say with certainty that the RA with monitor in the MS(pgsql) is called ONLY on the node on which the last was launched PACEMAKER.
>>>  It looks like you're hitting https://github.com/beekhof/pacemaker/commit/58962338
>>>  Since you appear to be on rhel6 (or a clone of rhel6), can I suggest you use the 1.1.10 packages that come with 6.4?
>>>  They include the above patch.
>>  I already use (builded from source two weeks ago)
>
> Upstream 1.1.10 does not include the above patch.

Strangely, in my source code, these lines exist.
Maybe I do not collect build.
I have the source - "master", and build RPM - 1.1.10.

>
>>  * pacemaker 1.1.10
>>  * resource-agents 3.9.5
>>  * corosync 2.3.2
>>  * libqb 0.16
>>  & CentOS 6.4
>>
>>  The same config work on pacemaker 1.1.9/corosync 1.4.5
>>  Not ideal, but no such problem.
>>
>>  At first idea - I thought I should move the target-role=Master from MS to primitive pgsql.
>>  And so even working.
>>  But after a crash killing the main PostgreSQL process - started the same.
>>  Today's experiments showed that this behavior starts after I add in MS "notify=true".
>>  But primitive pgsql not properly work without "notify" messages.
>>  While I in frustration :(
>>>  Also, just to be sure. Are you expecting monitor operations to detect when you started a resource manually?
>>>  If so, you'll need a monitor operation with role=Stopped. We don't do that by default.
>>  I expect that the resource monitoring on all the time, otherwise how to control them?
>
> When a node joins the cluster, we check to see if it had any resources running.
> If no-one has the resource running, we pick a node and start it there.
>
> If a malicious admin then starts the resource somewhere else, manually, we would not normally detect this.
> It is assumed that someone trusted with root privileges would not do this on purpose.
>
> However, if you do not trust your admins, as explained above, you can configure pacemaker to periodically re-check the node to detect and recover from this situation.

This individual servers allocated for development.
Their only one malicious admin - It's I.
But, I trust myself yet. :) 

And even if I run something "hands" , then each time the utility deployment of cluster:
* Certainly removes all the packages that were installed prior to this (referring to the cluster (libqb, corosync1 / 2 , cman, pacemaker, resource-agents, pcs, crmsh, cluster-glue));
* Deletes all files that might remain from a previous installation ( and configs , and temp) ;
* Install my RPM's;
* Turns off and disables all of the services that may interfere with the operation of the cluster ;
* Cleared the old configuration, calculates new configs corosync and pacemaker;
* Writes and runs their cluster alternately every synchronizing each node.

In this situation it is difficult to survive something forgotten.
I reinstalled the OS would be if I had access to a level above :)

>
>>  Or I do not quite understand the question.
>>>>>   'crm_mon -o' might be a good source of information too.
>>>>   Therefore, I see that my resources allegedly functioning normally.
>>>>
>>>>   # crm_mon -o1
>>>>   Last updated: Tue Nov 12 09:27:16 2013
>>>>   Last change: Tue Nov 12 00:08:35 2013 via crm_attribute on dev-cluster2-node2
>>>>   Stack: corosync
>>>>   Current DC: dev-cluster2-node2 (172793105) - partition with quorum
>>>>   Version: 1.1.10-1.el6-368c726
>>>>   3 Nodes configured
>>>>   337 Resources configured
>>>>
>>>>   Online: [ dev-cluster2-node2 dev-cluster2-node3 dev-cluster2-node4 ]
>>>>
>>>>   Clone Set: clonePing [pingCheck]
>>>>       Started: [ dev-cluster2-node2 dev-cluster2-node3 dev-cluster2-node4 ]
>>>>   Master/Slave Set: msPgsql [pgsql]
>>>>       Masters: [ dev-cluster2-node2 ]
>>>>       Slaves: [ dev-cluster2-node3 dev-cluster2-node4 ]
>>>>   VirtualIP      (ocf::heartbeat:IPaddr2):       Started dev-cluster2-node2
>>>>
>>>>   Operations:
>>>>   * Node dev-cluster2-node2:
>>>>     pingCheck: migration-threshold=1
>>>>      + (20) start: rc=0 (ok)
>>>>      + (23) monitor: interval=10000ms rc=0 (ok)
>>>>     pgsql: migration-threshold=1
>>>>      + (41) promote: rc=0 (ok)
>>>>      + (87) monitor: interval=1000ms rc=8 (master)
>>>>     VirtualIP: migration-threshold=1
>>>>      + (49) start: rc=0 (ok)
>>>>      + (52) monitor: interval=10000ms rc=0 (ok)
>>>>   * Node dev-cluster2-node3:
>>>>     pingCheck: migration-threshold=1
>>>>      + (20) start: rc=0 (ok)
>>>>      + (23) monitor: interval=10000ms rc=0 (ok)
>>>>     pgsql: migration-threshold=1
>>>>      + (26) start: rc=0 (ok)
>>>>      + (32) monitor: interval=10000ms rc=0 (ok)
>>>>   * Node dev-cluster2-node4:
>>>>     pingCheck: migration-threshold=1
>>>>      + (20) start: rc=0 (ok)
>>>>      + (23) monitor: interval=10000ms rc=0 (ok)
>>>>     pgsql: migration-threshold=1
>>>>      + (26) start: rc=0 (ok)
>>>>      + (32) monitor: interval=10000ms rc=0 (ok)
>>>>
>>>>   In reality now killed (signal 4|6) the PG master and the penultimate slave PG.
>>>>   IMHO, even if I have something configured incorrectly, the inability to monitor the resource must cause a fatal error.
>>>>   Or is there a reason not to do so?
>>>>
>>>>   _______________________________________________
>>>>   Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
>>>>   http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>>>>
>>>>   Project Home: http://www.clusterlabs.org
>>>>   Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
>>>>   Bugs: http://bugs.clusterlabs.org
>>>  ,
>>>  _______________________________________________
>>>  Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
>>>  http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>>>
>>>  Project Home: http://www.clusterlabs.org
>>>  Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
>>>  Bugs: http://bugs.clusterlabs.org
>>  _______________________________________________
>>  Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
>>  http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>>
>>  Project Home: http://www.clusterlabs.org
>>  Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
>>  Bugs: http://bugs.clusterlabs.org
>
> ,
> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org




More information about the Pacemaker mailing list