| |
|
RSS Feed - WebSphere MQ Support
|
RSS Feed - Message Broker Support
|
  |
|
| Active-Active HACMP Cluster on AIX 5.3 |
« View previous topic :: View next topic » |
| Author |
Message
|
| mqwbi |
 Posted: Thu Dec 25, 2008 10:04 am Post subject: Active-Active HACMP Cluster on AIX 5.3 |
 |
|
Novice
Joined: 15 Dec 2008
Posts: 24
|
Hi,
I am doing an Active-Active HACMP setup for WMQ 6 on AIX 5.3. I am taking reference from supportpac MC91
I have two Nodes, NodeA and NodeB. Each Node has 1 Resource Group. Each resource group will have only 1 Queue Manager. The shared disks are on the SAN.
I have a few questions, can you please help?
>> Will there be two separate Shared Storages required for Active-Active configuration? one SAN storage for each Node.
>> Will there be different mount points of the Shared Disks from each Node?
>> if there are different mount points, how will the halinkmqm command work when run from the takeover node?
Say, for instance, I have created a Queue Manager ha.csq1 on NodeA, the filesystem for data on the SAN being /MQHA/ha.csq1/data
Now, i have to run the halinkmqm from the takeover Node, NodeB
I understand that the command will be something like this
$halinkmqm ha.csq1 ha!csq1 /MQHA/ha.csq1/data
here /MQHA/ha.csq1/data is the filesystem in the SAN for Queue Manager's data in the NodeA. How can we access this filesystem (mounted on NodeA) from NodeB if the mount points are different.
addtionally, What changes have to be done to the start and stop script (provided with MC91) in order to make it work for active-active set-up
Can you please help?
Thank You And Merry Christmas |
|
| Back to top |
|
 |
| exerk |
| Posted: Sat Dec 27, 2008 2:10 pm Post subject: Re: Active-Active HACMP Cluster on AIX 5.3 |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| mqwbi wrote: |
| >> Will there be two separate Shared Storages required for Active-Active configuration? one SAN storage for each Node. |
Yes...different volume groups seem the sensible option.
| mqwbi wrote: |
| >> Will there be different mount points of the Shared Disks from each Node? |
That question might better be directed at your AIX Sys Admins, but I would expect the 'same' mount points on each node.
| mqwbi wrote: |
| >> if there are different mount points, how will the halinkmqm command work when run from the takeover node? |
I'd have to look inside the script to be absolutely sure, but I think it re-links the IPC key subdirectories and creates the necessary sym-links, as well as updating the mqs.ini file on the 'passive' node.
I'm sure a UNIX guru will be along to explain it all better than I...
_________________
It's puzzling, I don't think I've ever seen anything quite like this before...and it's hard to soar like an eagle when you're surrounded by turkeys. |
|
| Back to top |
|
|
| mqwbi |
| Posted: Tue Dec 30, 2008 6:01 am Post subject: UNABLE TO ACCESS AN INI FILE |
|
|
Novice
Joined: 15 Dec 2008
Posts: 24
|
Hi
Merry Christmas
I have started with the HACMP cluster.
I have an active-active HACMP cluster for WMQ 6.0.2.5 on AIX 5.3
I have two Nodes, and each Node has a resource group
Each Resource group has a Queue Manager in it.
I have modified and run all the required scripts given in the Supportpac. currently, i have created only one queue manager in NODE 1. I haven't created the queue manager for NODE 2. I want to test the failover of the resource group in NODE 1, then i want to create queue manager for NODE 2.
When i start the cluster, my queue manager in NODE 1 starts as a result of execution of the start scripts in the application server. When i pull down the NODE 1, failover occurs. Now, the filesystems in NODE 1 are mounted on NODE 2. I am able to see the shared data and log directories on the mounted filesystems. The MQS.INI file has the entry for the queue manager in NODE 1. But the queue manager in NODE 1 doesn't start .
When i get in to NODE 2 and start the start up script (for queue manager in NODE 1) manually, i get an error which says "UNABLE TO ACCESS AN INI FILE". I debugged the hamqm_start script and found out that it is occuring at the point when we are trying to start the queue manager (of NODE 1) from NODE 2.
The mounted filesystems are owned by mqm, scripts are also owned by mqm, mqs.ini lies in mqm's home directory /var/mqm . mqm group and mqm user have same values for UID and GID in both nodes. both have same passwords. Also, I have same version of WMQ installed in both Nodes.
i am not sure, what is going wrong. can you please help?
Thank You |
|
| Back to top |
|
|
| exerk |
| Posted: Tue Dec 30, 2008 5:12 pm Post subject: Re: UNABLE TO ACCESS AN INI FILE |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| mqwbi wrote: |
| ...When i get in to NODE 2 and start the start up script (for queue manager in NODE 1) manually, i get an error which says "UNABLE TO ACCESS AN INI FILE"... |
Does it state which INI file it cannot access?
_________________
It's puzzling, I don't think I've ever seen anything quite like this before...and it's hard to soar like an eagle when you're surrounded by turkeys. |
|
| Back to top |
|
|
| Vitor |
| Posted: Wed Dec 31, 2008 12:43 am Post subject: |
|
|
Grand High Poobah
Joined: 11 Nov 2005
Posts: 26093
Location: Texas, USA
|
|
| Back to top |
|
|
| KramJ |
| Posted: Fri Jan 09, 2009 11:05 am Post subject: |
|
|
Voyager
Joined: 09 Jan 2006
Posts: 80
Location: Atlanta
|
It's either /var/mqm/mqs.ini or /MQHA/<QMGR>/data/qmgrs/<QMGR>/qm.ini.
/var/mqm/mqs.ini doesn't get failed over so verify that the files are the same on node 1 and node 2.
The qm.ini does get failed over so make sure that it is available on node 2 after failover. |
|
| Back to top |
|
|
| kaniyan |
| Posted: Thu Jul 23, 2009 4:40 am Post subject: same issue |
|
|
Newbie
Joined: 23 Jul 2009
Posts: 1
|
We also faced same issue, on Node 2 we filled manually Qmager details in qm.ini , then it started working.
Bye |
|
| Back to top |
|
|
| exerk |
| Posted: Thu Jul 23, 2009 5:39 am Post subject: Re: same issue |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| kaniyan wrote: |
| We also faced same issue, on Node 2 we filled manually Qmager details in qm.ini , then it started working... |
mqs.ini maybe, but definitely not qm.ini. In both cases here, I suspect a step was missed, which is to run a script on the passive node, to populate the mqs.ini file with the relevant information - assuming that you were using the MC91 SupportPac...
_________________
It's puzzling, I don't think I've ever seen anything quite like this before...and it's hard to soar like an eagle when you're surrounded by turkeys. |
|
| Back to top |
|
|
|
|
|
| |
|
Page 1 of 1 |
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
|
|
