[Pacemaker] The order of resources changes when resources register themselves by a crm command.

[renayama19661014 at ybb.ne.jp]

Hi,

We tested a crm command.

We tested it by the combination of the next version

- corosync 1.1.2
- Pacemaker-1-0-66b7bfd467f3
- Reusable-Cluster-Components-0ba7119b29ad
- Cluster-Resource-Agents-6d652f7cf9d8

We made the cli file which I attached and registered a resource in the next procedure.

Step1) Erase cib.xml
Step2) Start corosync
Step3) crm configure
Step4) crm# load update test.cli
Step5) crm# commit force
Step6) Confirm registration by a crm_mon command

But, there is the case that the registered order of resources turns into when I repeat a procedure.

-------------The order of the first and the second--------------------
[root at rh54-1 ~]# crm_mon -1
============
Last updated: Fri Dec 11 10:24:10 2009
Stack: openais
Current DC: rh54-1 - partition with quorum
Version: 1.0.6-66b7bfd467f3de43b7921474711cd9158fbb2721
2 Nodes configured, 2 expected votes
4 Resources configured.
============

Online: [ rh54-1 rh54-2 ]

 Clone Set: clnDiskd1
     Started: [ rh54-1 rh54-2 ]
 Clone Set: clnPingd1
     Started: [ rh54-1 rh54-2 ]
 Clone Set: clnDiskd2
     Started: [ rh54-1 rh54-2 ]
 Resource Group: grpPostgreSQLDB
     prmExPostgreSQLDB  (ocf::heartbeat:Dummy): Started rh54-1
     prmFsPostgreSQLDB1 (ocf::heartbeat:Dummy): Started rh54-1
     prmFsPostgreSQLDB2 (ocf::heartbeat:Dummy): Started rh54-1
     prmFsPostgreSQLDB3 (ocf::heartbeat:Dummy): Started rh54-1
     prmIpPostgreSQLDB  (ocf::heartbeat:IPaddr):        Started rh54-1
     prmApPostgreSQLDB  (ocf::heartbeat:Dummy): Started rh54-1   


-------------The order of the third-----------------------------------
[root at rh54-1 ~]# crm_mon -1
============
Last updated: Fri Dec 11 10:37:15 2009
Stack: openais
Current DC: rh54-1 - partition with quorum
Version: 1.0.6-66b7bfd467f3de43b7921474711cd9158fbb2721
2 Nodes configured, 2 expected votes
4 Resources configured.
============

Online: [ rh54-1 rh54-2 ]

 Clone Set: clnDiskd1
     Started: [ rh54-1 rh54-2 ]
 Clone Set: clnDiskd2
     Started: [ rh54-1 rh54-2 ]
 Resource Group: grpPostgreSQLDB
     prmExPostgreSQLDB  (ocf::heartbeat:Dummy): Started rh54-1
     prmFsPostgreSQLDB1 (ocf::heartbeat:Dummy): Started rh54-1
     prmFsPostgreSQLDB2 (ocf::heartbeat:Dummy): Started rh54-1
     prmFsPostgreSQLDB3 (ocf::heartbeat:Dummy): Started rh54-1
     prmIpPostgreSQLDB  (ocf::heartbeat:IPaddr):        Started rh54-1
     prmApPostgreSQLDB  (ocf::heartbeat:Dummy): Started rh54-1
 Clone Set: clnPingd1
     Started: [ rh54-2 rh54-1 ]
----------------------------------------------------------------------

The order of clnPingd1 changes.
It is a problem that the order of resources changes to use the same cli file.

Is this a bug?
Or is it a problem of the problem of my cli file or my operation?

Best Regards,
Hideo Yamauchi.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: test.cli
Type: application/octet-stream
Size: 2682 bytes
Desc: 1877427586-test.cli
URL: <http://oss.clusterlabs.org/pipermail/pacemaker/attachments/20091211/76cfcdd0/attachment.dll>