[Pacemaker] [PACEMAKER] Why cant't migrate group resource with collation in drbd resource
Kashif Jawed Siddiqui
kashifjs at huawei.com
Wed Jan 23 09:15:40 EST 2013
There is a Pacemaker bug which cannot be replaced due to legacy tracking and backward compatibility
colocation FS_WITH_DRBD inf: IP-AND-FS ms_drbd:Master
order DRBD_BEF_FS inf: IP-AND-FS:start ms_drbd:promote
if colocation and order is specified between 2 Resources then it means the 2nd one is the first and 1st one is the next
For example,
colocation FS_WITH_DRBD inf: IP-AND-FS ms_drbd:Master
//Actually means first start IP-AND-FS then make ms_drbd as Master
// But in behavior, it is first make ms_drbd as Master and then start IP-AND-FS
Also for order, it is the same behavior...
order DRBD_BEF_FS inf: IP-AND-FS:start ms_drbd:promote
//Actually means first start IP-AND-FS then promote ms_drbd
// But in behavior, it is first promote ms_drbd and then start IP-AND-FS
But when you define order and colocation between 3 or more resources, then the interpretation is fine...
Therefore in your scenario, you must configure (as I said)
colocation FS_WITH_DRBD inf: IP-AND-FS ms_drbd:Master
order DRBD_BEF_FS inf: IP-AND-FS:start ms_drbd:promote
Though it means opposite, but the behavior is as you expect.(owing to Pacemaker backward compatibility)
Regards,
Kashif Jawed Siddiqui
***************************************************************************************
This e-mail and attachments contain confidential information from HUAWEI, which is intended only for the person or entity whose address is listed above. Any use of the information contained herein in any way (including, but not limited to, total or partial disclosure, reproduction, or dissemination) by persons other than the intended recipient's) is prohibited. If you receive this e-mail in error, please notify the sender by phone or email immediately and delete it!
________________________________
From: and k [not4mad at gmail.com]
Sent: Wednesday, January 23, 2013 7:31 PM
To: The Pacemaker cluster resource manager
Subject: Re: [Pacemaker] [PACEMAKER] Why cant't migrate group resource with collation in drbd resource
2013/1/23 emmanuel segura <emi2fast at gmail.com<mailto:emi2fast at gmail.com>>
Ummm
Fist the IP-AND-FS? but what happen if the FS is on drbd?
Emmanuel, you are right, filesystem is on top of drbd device, so i cant run that group before drbd promote.
I also noticed that, I can migrate it by force with use:
crm_resource -M -f -r fs_r1 -H drbd02
master is promoted correctly, and group is succesfully migrated.
But i wonder why it doesn't work without force (-f) parameter ?? Any idea ?
Thanks
2013/1/23 Kashif Jawed Siddiqui <kashifjs at huawei.com<mailto:kashifjs at huawei.com>>
You must change the order
#order DRBD_BEF_FS inf: ms_drbd:promote IP-AND-FS:start
order DRBD_BEF_FS inf: IP-AND-FS:start ms_drbd:promote
//First start IP-AND-FS, only then promote ms_drbd
Regards,
Kashif Jawed Siddiqui
***************************************************************************************
This e-mail and attachments contain confidential information from HUAWEI, which is intended only for the person or entity whose address is listed above. Any use of the information contained herein in any way (including, but not limited to, total or partial disclosure, reproduction, or dissemination) by persons other than the intended recipient's) is prohibited. If you receive this e-mail in error, please notify the sender by phone or email immediately and delete it!
________________________________
From: and k [not4mad at gmail.com<mailto:not4mad at gmail.com>]
Sent: Wednesday, January 23, 2013 4:34 PM
To: pacemaker at oss.clusterlabs.org<mailto:pacemaker at oss.clusterlabs.org>
Subject: [Pacemaker] [PACEMAKER] Why cant't migrate group resource with collation in drbd resource
Hello Everybody,
I've got a problem (but I am not quite sure if it is not a feature in pacemaker ) that's why I decided to write on that mailing list.
It comes about migrate resource with collation in drbd resource.
I've got group including virtual IP and filesystem which is collated with ms resource drbd in mster-slave configuration.
============
Last updated: Wed Jan 23 03:40:55 2013
Stack: openais
Current DC: drbd01 - partition with quorum
Version: 1.0.9-74392a28b7f31d7ddc86689598bd23114f58978b
============
Online: [ drbd01 drbd02 ]
Master/Slave Set: ms_drbd
Masters: [ drbd01 ]
Slaves: [ drbd02 ]
Resource Group: IP-AND-FS
fs_r1 (ocf::heartbeat:Filesystem): Started drbd01
VIRTUAL-IP (ocf::heartbeat:IPaddr): Started drbd01
I would like to migrate that group manually to another node which is slave. So i type in: crm resource migrate IP-AND-FS drbd02
after that configuration include additional line:
location cli-prefer-IP-AND-FS IP-AND-FS \
rule $id="cli-prefer-rule-IP-AND-FS" inf: #uname eq drbd02
and in logs i see:
Jan 23 11:30:12 drbd02 cibadmin: [1126]: info: Invoked: cibadmin -Ql -o resources
Jan 23 11:30:12 drbd02 cibadmin: [1129]: info: Invoked: cibadmin -Ql -o nodes
Jan 23 11:30:12 drbd02 cibadmin: [1131]: info: Invoked: cibadmin -Ql -o resources
Jan 23 11:30:12 drbd02 cibadmin: [1133]: info: Invoked: cibadmin -Ql -o nodes
Jan 23 11:30:12 drbd02 cibadmin: [1135]: info: Invoked: cibadmin -Ql -o resources
Jan 23 11:30:12 drbd02 cibadmin: [1137]: info: Invoked: cibadmin -Ql -o nodes
Jan 23 11:30:14 drbd02 cibadmin: [1166]: info: Invoked: cibadmin -Ql -o resources
Jan 23 11:30:14 drbd02 cibadmin: [1168]: info: Invoked: cibadmin -Ql -o nodes
Jan 23 11:30:14 drbd02 cibadmin: [1170]: info: Invoked: cibadmin -Ql -o resources
Jan 23 11:30:14 drbd02 cibadmin: [1172]: info: Invoked: cibadmin -Ql -o nodes
Jan 23 11:30:16 drbd02 cibadmin: [1174]: info: Invoked: cibadmin -Ql -o resources
Jan 23 11:30:16 drbd02 cibadmin: [1176]: info: Invoked: cibadmin -Ql -o nodes
Jan 23 11:30:16 drbd02 cibadmin: [1178]: info: Invoked: cibadmin -Ql -o resources
Jan 23 11:30:16 drbd02 cibadmin: [1180]: info: Invoked: cibadmin -Ql -o nodes
Jan 23 11:30:40 drbd02 cibadmin: [1211]: info: Invoked: cibadmin -Ql -o nodes
Jan 23 11:30:40 drbd02 crm_resource: [1213]: info: Invoked: crm_resource -M -r IP-AND-FS --node=drbd02
Jan 23 11:30:40 drbd02 cib: [1214]: info: write_cib_contents: Archived previous version as /var/lib/heartbeat/crm/cib-73.raw
Jan 23 11:30:40 drbd02 cib: [1214]: info: write_cib_contents: Wrote version 0.225.0 of the CIB to disk (digest: 166251193cbe1e0b9314ab07358accca)
Jan 23 11:30:40 drbd02 cib: [1214]: info: retrieveCib: Reading cluster configuration from: /var/lib/heartbeat/crm/cib.tk72Ft (digest: /var/lib/heartbeat/crm/cib.hF2UsS)
Jan 23 11:30:44 drbd02 cib: [30098]: info: cib_stats: Processed 153 operations (1437.00us average, 0% utilization) in the last 10min
but nothing happened resource group is still active on drbd01 node, as well as there was no new master promotion.
Shouldn't pacemaker automatically promote second node to master and move my resource group ?
Below is my test configuration, i will be appreciate for help:
crm(live)# configure show
node drbd01 \
attributes standby="off"
node drbd02 \
attributes standby="off"
primitive VIRTUAL-IP ocf:heartbeat:IPaddr \
params ip="10.11.11.111"
primitive drbd ocf:linbit:drbd \
params drbd_resource="r1" \
op start interval="0" timeout="240" \
op stop interval="0" timeout="100" \
op monitor interval="59s" role="Master" timeout="30s" \
op monitor interval="60s" role="Slave" timeout="30s"
primitive fs_r1 ocf:heartbeat:Filesystem \
params device="/dev/drbd1" directory="/mnt" fstype="ext3" \
op start interval="0" timeout="60" \
op stop interval="0" timeout="120" \
meta allow-migrate="true"
group IP-AND-FS fs_r1 VIRTUAL-IP \
meta target-role="Started"
ms ms_drbd drbd \
meta master-node-max="1" clone-max="2" clone-node-max="1" globally-unique="false" notify="true" target-role="Master"
location cli-prefer-IP-AND-FS IP-AND-FS \
rule $id="cli-prefer-rule-IP-AND-FS" inf: #uname eq drbd02
colocation FS_WITH_DRBD inf: IP-AND-FS ms_drbd:Master
order DRBD_BEF_FS inf: ms_drbd:promote IP-AND-FS:start
property $id="cib-bootstrap-options" \
dc-version="1.0.9-74392a28b7f31d7ddc86689598bd23114f58978b" \
cluster-infrastructure="openais" \
expected-quorum-votes="2" \
stonith-enabled="false" \
no-quorum-policy="ignore" \
last-lrm-refresh="1358868655" \
default-resource-stickiness="1"
Regards
Andrew
_______________________________________________
Pacemaker mailing list: Pacemaker at oss.clusterlabs.org<mailto: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
--
esta es mi vida e me la vivo hasta que dios quiera
_______________________________________________
Pacemaker mailing list: Pacemaker at oss.clusterlabs.org<mailto: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
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.clusterlabs.org/pipermail/pacemaker/attachments/20130123/c9d1c4a9/attachment-0003.html>
More information about the Pacemaker
mailing list