| Author |
Message
|
| pintrader |
 Posted: Mon Apr 07, 2014 4:55 am Post subject: MSCS clustering with MQ 7.5.0.3 |
 |
|
Disciple
Joined: 22 Jan 2014
Posts: 164
|
hi,
am trying to set up HA with Microsoft CLustering on Windows 2012 R2 Vmware VMs, with MQ version 7.5.0.3(fixpack). I followed this infocenter page: http://pic.dhe.ibm.com/infocenter/wmqv7/v7r5/topic/com.ibm.mq.con.doc/q018010_.htm, but it was only updated to Windows 2008.
Has anyone have experienced in setting up MSCS with Windows 2012 R2 VMs? Nevertheless I tried the steps outlined for windows 2008, but i have problems with the failover.
I have this error:
| Quote: |
| Locking is not possible in the directory D...:\eclipse\configuration\org.eclipse.osgi. A common reason is that the file system or Runtime Environment does not support file locking for that location. Please choose a different location, or disable file locking passing "-Dosgi.locking=none" as a VM argument. D:...\eclipse\configuration\org.eclipse.osgi.manager.fileTableLock (Access is denied) |
I have 2 identical VM setups, but I only created one Q manager on the first VM (the 2nd VM I did not create the Q Manager, which is correct right?).
Then i use hamvmqm command to move its data/log path to the shared directory.
Of course I also use haregtype /r on both VMs.
any ideas?
thanks |
|
| Back to top |
|
 |
| exerk |
| Posted: Mon Apr 07, 2014 5:01 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
Does not this:
| Quote: |
| ...or disable file locking passing "-Dosgi.locking=none" as a VM argument... |
Suggest it may be VM-related? Have you tried restarting your VMs with that argument in place?
_________________
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 |
|
|
| smdavies99 |
| Posted: Mon Apr 07, 2014 5:08 am Post subject: |
|
|
Jedi Council
Joined: 10 Feb 2003
Posts: 6076
Location: Somewhere over the Rainbow this side of Never-never land.
|
This
| Code: |
D...:\eclipse\configuration\org.eclipse.osgi.
|
would seem to indicate that you were trying to run an eclipse based program. Perhaps MQExplorer?
did you
1) install wmq on a NON shared drive in each cluster member eg C:\IBM\WMQ etc
2) create the queue manager using the -a option? This is needed when you are using Server 2012. If you are using all server 2008 then you don't need it.
3) create the queue manager with a default location for the queue and logfile directories
4) move the queue manager data and logs to the shared drive with the hamvmqm command?
Why not capture the steps you took to create the QMGR and move it and post them here.
DO NOT add the qmgr as a clustered resource until you have failed over the cluster and manually started the qmgr on the other side.
OR if you do make sure the QMGR resource is offline before failing over the cluster.
_________________
WMQ User since 1999
MQSI/WBI/WMB/'Thingy' User since 2002
Linux user since 1995
Every time you reinvent the wheel the more square it gets (anon). If in doubt think and investigate before you ask silly questions. |
|
| Back to top |
|
|
| pintrader |
| Posted: Mon Apr 07, 2014 5:50 am Post subject: |
|
|
Disciple
Joined: 22 Jan 2014
Posts: 164
|
| smdavies99 wrote: |
This
| Code: |
D...:\eclipse\configuration\org.eclipse.osgi.
|
would seem to indicate that you were trying to run an eclipse based program. Perhaps MQExplorer?
|
yes, I was trying to start the Explorer and it gives me the file locking error.
| smdavies99 wrote: |
1) install wmq on a NON shared drive in each cluster member eg C:\IBM\WMQ etc
|
I suppose NON shared drive means the local drive. Yes, i installed 2 identical MQ installation on the 2 VMs. By default its installed to C:\Program Files(x86)\IBM\Websphere MQ\......
I only have the qmgrs and log directory moved to the shared drive, which is at e: , using the command havmvmqm /m Q.Mgr /dd e:\Websphere /ld e:\Websphere\log.
| smdavies99 wrote: |
2) create the queue manager using the -a option? This is needed when you are using Server 2012. If you are using all server 2008 then you don't need it.
|
I created without the -a option. I guess that's one mistake i made. Thanks for highlighting. I checked the documentation (as i am not really well versed in security stuff. I guess for MSCS i must create a global group? What should the group name be? Or i can name it any name I like? I already have a Domain mqm group created using the steps in Creating and Setting up Domain accounts. I also created a WMQ domain user , called "mqservicer" to run the "IBM WebSphere MQ (Installation1)" service at startup. mqservicer is under the Domain mqm group. Is that correct?
| smdavies99 wrote: |
3) create the queue manager with a default location for the queue and logfile directories
|
yes, before the MSCS cluster is up and running, I created Q.mgr using default data path and log path. (After the MSCS cluster is up, I then move the data and log path using hamvmqm command)
| smdavies99 wrote: |
4) move the queue manager data and logs to the shared drive with the hamvmqm command?
|
yes. as stated in 3)
| smdavies99 wrote: |
Why not capture the steps you took to create the QMGR and move it and post them here.
|
I created the Q.mgr using "crtmqm Q.mgr" , but without -a. Then started it using strmqm Q.mgr. All follow iBM docs. I also did the haregtyp /r command on both VMs , then my VM guys did the Putting Queue manager under MSCS control page (But its only till 2008, there's no update for 2012). After that, i did the hamvmqm command. I can see that the e: drive contains qmgrs and log directory after that.
thanks again for the help |
|
| Back to top |
|
|
| pintrader |
| Posted: Mon Apr 07, 2014 5:59 am Post subject: |
|
|
Disciple
Joined: 22 Jan 2014
Posts: 164
|
| exerk wrote: |
Does not this:
| Quote: |
| ...or disable file locking passing "-Dosgi.locking=none" as a VM argument... |
Suggest it may be VM-related? Have you tried restarting your VMs with that argument in place? |
According to here, it could be few possibilities, some pertaining to eclipse ini and some say its UAC. I will have to try them out. |
|
| Back to top |
|
|
| pintrader |
| Posted: Wed Apr 09, 2014 10:12 pm Post subject: |
|
|
Disciple
Joined: 22 Jan 2014
Posts: 164
|
We manage to get things up to work, except for a minor problem. We failover Active node (A) to Passive node (P), with the IBM services running (IBM MQSeries and IBM Websphere MQ installation1) as well as the Queue manager is up. But when we fail back from the (P) to (A), all ibm services are up but not the Queue manager. Anyone has this experience?
thanks |
|
| Back to top |
|
|
| smdavies99 |
| Posted: Wed Apr 09, 2014 10:27 pm Post subject: |
|
|
Jedi Council
Joined: 10 Feb 2003
Posts: 6076
Location: Somewhere over the Rainbow this side of Never-never land.
|
Have you made sure that the directories on the shared drive are setup for the mqservice account to have FULL access? On both nodes?
using the -a "mqservice" may well work here.
You may be better off deleting the QMGR on both sides (don't forget the passive node rmvmqing <qmgr> )
deleting all QMGR specific files on the shared drive and doing a clean build.
1) create the qmgr and verify that it runs.
2) move to shared drive
3) verify it runs again
4) stop QMGR, failover and start qmgr check it out.
5) Stop, failover and add to cluster making sure you set
the dependencies -- shared drive, ipaddr, service name
run with a separate resource monitor
6) start using Cluster manager.
7) stop again
 failover and start service using cluster manager.
finally,
9) failover using cluster manager.
DONT set the qmgr to Autostart.
If the qmgr does not start then there should be a reason why in the event log.
_________________
WMQ User since 1999
MQSI/WBI/WMB/'Thingy' User since 2002
Linux user since 1995
Every time you reinvent the wheel the more square it gets (anon). If in doubt think and investigate before you ask silly questions. |
|
| Back to top |
|
|
| pintrader |
| Posted: Wed Apr 09, 2014 10:55 pm Post subject: |
|
|
Disciple
Joined: 22 Jan 2014
Posts: 164
|
| smdavies99 wrote: |
Have you made sure that the directories on the shared drive are setup for the mqservice account to have FULL access? On both nodes?
|
hi, thanks for replying. am confused about this part. The shared drive is provisioned as a SAN disk. If mqservice has no full access then i won't be able to failover from (A) to (P) ? What do you mean by both nodes? because the shared disk is seen as "one disk" by both nodes.
| smdavies99 wrote: |
using the -a "mqservice" may well work here.
|
sorry, this option is from which command? thanks
| smdavies99 wrote: |
You may be better off deleting the QMGR on both sides (don't forget the passive node rmvmqing <qmgr> )
deleting all QMGR specific files on the shared drive and doing a clean build.
1) create the qmgr and verify that it runs.
2) move to shared drive
3) verify it runs again
4) stop QMGR, failover and start qmgr check it out.
5) Stop, failover and add to cluster making sure you set
the dependencies -- shared drive, ipaddr, service name
run with a separate resource monitor
6) start using Cluster manager.
7) stop again
8) failover and start service using cluster manager.
finally,
9) failover using cluster manager.
DONT set the qmgr to Autostart.
If the qmgr does not start then there should be a reason why in the event log. |
yes, i will try re-creating everything and see. thanks |
|
| Back to top |
|
|
| exerk |
| Posted: Wed Apr 09, 2014 11:38 pm Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| pintrader wrote: |
| ...am confused about this part. The shared drive is provisioned as a SAN disk. If mqservice has no full access then i won't be able to failover from (A) to (P) ? What do you mean by both nodes? because the shared disk is seen as "one disk" by both nodes... |
In MCSC terms, a 'shared' disk is one that can be mounted on one, or other, nodes in the MCSC Cluster - but NOT concurrently, i.e. at the same time; it will either be mounted on Node A, or mounted on Node B. If you can see it mounted on both Nodes, and access it from both Nodes, it is not an MSCS Cluster resource and what you're doing won't work - are you sure you're trying to achieve MCSC Cluster control of the queue manager, or Multi-Instance control of the queue manager?
_________________
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 |
|
|
| smdavies99 |
| Posted: Wed Apr 09, 2014 11:46 pm Post subject: |
|
|
Jedi Council
Joined: 10 Feb 2003
Posts: 6076
Location: Somewhere over the Rainbow this side of Never-never land.
|
| pintrader wrote: |
| smdavies99 wrote: |
using the -a "mqservice" may well work here.
|
sorry, this option is from which command? thanks
|
this is on the crtmqm command
_________________
WMQ User since 1999
MQSI/WBI/WMB/'Thingy' User since 2002
Linux user since 1995
Every time you reinvent the wheel the more square it gets (anon). If in doubt think and investigate before you ask silly questions. |
|
| Back to top |
|
|
| pintrader |
| Posted: Wed Apr 09, 2014 11:52 pm Post subject: |
|
|
Disciple
Joined: 22 Jan 2014
Posts: 164
|
| exerk wrote: |
In MCSC terms, a 'shared' disk is one that can be mounted on one, or other, nodes in the MCSC Cluster - but NOT concurrently, i.e. at the same time; it will either be mounted on Node A, or mounted on Node B. If you can see it mounted on both Nodes, and access it from both Nodes, it is not an MSCS Cluster resource and what you're doing won't work - are you sure you're trying to achieve MCSC Cluster control of the queue manager, or Multi-Instance control of the queue manager? |
yes, my e drive is a shared disk. When Node A is active, Node B won't be abe to see the drive. my e drive contains only the qmgr and log directories. |
|
| Back to top |
|
|
|
|
