I want to leave the details to beekhof as he should be announcing to
upgrade soon. Just so much:
There recently was a change on how master colocation is expressed score
wise which causes this.
I think you hit this bug here that has been fixed just 2 days ago.
Please upgrade to the latest 0.6.4 version from the repository or wait
for beekhof to build new packages.
That version also includes an updated showscores version that will
display sane group scores for the fixed behaviour.
Regards
Dominik
Geoffroy ARNOUD wrote:
Hi all,
We are setting up a MySQL HA cluster, with Heart/Pacemaker and DRBD.
Heartbeat is configured has follows:
- DRBD as a master/slave resource
- MySQL as a resource-group with the following primitives:
* a Virtual IP address (IPAddr2)
* a File system
* MySQL
- 2 constraints between DRBD and the resource group (rsc_order and
rsc_colocation)
We have some troubles to figure out how scores are computed.
We have set the following resource_stickiness and failure stickiness:
- Mysql : 210 / -100
- File system : 60 / -25
- VIP : 25 / -10
When cluster is started, the showscores output is:
Resource Score Node Stickiness #Fail
Fail-Stickiness
myserver01-drbd:0 76 uamwpdb2 100 0 -100
myserver01-drbd:0 -INFINITY uamwpdb1 100 0 -100
myserver01-drbd:0_(master)75 uamwpdb2 100 0 -100
myserver01-drbd:1 0 uamwpdb2 100 0 -100
myserver01-drbd:1 76 uamwpdb1 100 0 -100
myserver01-drbd:1_(master)665 uamwpdb1 100 0 -100
myserver01-fs 270 uamwpdb1 60 0 -25
myserver01-fs -INFINITY uamwpdb2 60 0 -25
myserver01-mysql 210 uamwpdb1 210 0 -100
myserver01-mysql -INFINITY uamwpdb2 210 0 -100
myserver01-vip 371 uamwpdb1 25 0 -10
myserver01-vip -INFINITY uamwpdb2 25 0 -10
From heartbeat doc, resource-group score is : 210 + 60 + 25 = 295
Here : myserver01-vip score is 371, which is 295 + 76. As there is a
constraint between DRBD and resource-group, I understand that scores
are sumed.
My questions are :
- where are the normal score (76) and master score of slave (75) computed ?
- the Master score of Master is 665, which seems to be rg-score (295)
* 2 + normal score (75) - is that right ? Why is the resource-group
score counted twice ?
With this config, after 3 failures of the database, all the resources
are migrated to the slave node, because the master score of DRBD on
the master node goes under 75. Which makes heartbeat fail the DRBD
over.
With other values of stickiness, it is possible to fall in a case
where the lysql database has a negative score for the master node, but
the master score of DRBD is greater than the slave node. Therefore,
heartbeat refuses to restart the database, but won't migrate the
resources
Is it possible to configure heartbeat so when a resource of a
resource-group is not able to run on the node anymore (<0), the
resource-group and the master/salve are failed-over ?
My CIB.xml is attached.
Software releases are (taken from
http://download.opensuse.org/repositories/server:/ha-clustering for
heartbeat):
- heartbeat-resources-2.1.3-22.1
- heartbeat-common-2.1.3-22.1
- pacemaker-heartbeat-0.6.4-7.1
- heartbeat-2.1.3-22.1
- drbd-8.0.12-3
- drbd-km-2.6.18_8.el5-8.0.12-3
Thanks in advance for all answers.
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
