| Author |
Message
|
| mca |
 Posted: Thu Jun 21, 2012 7:23 am Post subject: issue creating Multi-Instance Queue Manager |
 |
|
Disciple
Joined: 09 Mar 2005
Posts: 196
|
I am trying to create a Multi-Instance Queue manager on one Linux guest and i have the NFS Server on another Linux Guest. On NFS Guest, i added /HA * to /etc/exports and mounted this directory onto MQ Guest.
On NFS Guest:
mqm@Guestnfs:~> cat /etc/passwd |grep mqm
mqm:x:1113:1406:mqm:/var/mqm:/bin/bash
mqm@Guestnfs:~> id mqm
uid=1113(mqm) gid=1406(mqm) groups=1406(mqm)
On MQ Guest
mqm@Guest1:~> cat /etc/passwd |grep mqm
mqm:x:1113:1406::/var/mqm:/bin/bash
mqm@guest1:~> id mqm
uid=1113(mqm) gid=1406(mqm) groups=1401(mqbrkrs),1012(tivapp),1406(mqm)
mqm@Guest1:~> amqmfsck /HA/WMQ/QMGR1/
AMQ6242: Incorrect ownership for '/HA/WMQ/QMGR1//amqmfsck.lck'.
Current(4294967294) Expected(1113)
The "mqm" UserID, including the numeric values are same on both the guests. Can someone help me figure out what else could be the issue here?
Thanks.
[/u] |
|
| Back to top |
|
 |
| exerk |
| Posted: Thu Jun 21, 2012 7:31 am Post subject: Re: issue creating Multi-Instance Queue Manager |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| mca wrote: |
mqm@Guest1:~> amqmfsck /HA/WMQ/QMGR1/
AMQ6242: Incorrect ownership for '/HA/WMQ/QMGR1//amqmfsck.lck'. |
Is the above a typo? For MI queue managers you need two servers and you haven't mentioned the other MI server - are the mqm group/user IDs the same as the other MI server?
_________________
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 |
|
|
| mca |
| Posted: Thu Jun 21, 2012 8:55 am Post subject: |
|
|
Disciple
Joined: 09 Mar 2005
Posts: 196
|
The third Linux Guest is hosting the standby instance of Queue manager. It only has MQ binaries installed as of now.
The MQM UID and GID are the exact same as other MQ Guest. |
|
| Back to top |
|
|
| mqjeff |
| Posted: Thu Jun 21, 2012 9:00 am Post subject: Re: issue creating Multi-Instance Queue Manager |
|
|
Grand Master
Joined: 25 Jun 2008
Posts: 17447
|
| mca wrote: |
mqm@Guest1:~> amqmfsck /HA/WMQ/QMGR1/
AMQ6242: Incorrect ownership for '/HA/WMQ/QMGR1//amqmfsck.lck'.
Current(4294967294) Expected(1113) |
Both the things marked in red are confusing... I suspect that the first one can be fixed with
| Code: |
| amqmfsck /HA/WMQ/QMGR1 |
The second one indicates that something's wrong with the UID/GID lookup between the two machines.
You also haven't specified that you've taken steps to ensure that you are using a supported version of NFS, rather than an unsupported version of NFS. The ownership complaint suggests that you might be at an unsupported version of NFS.
Last edited by mqjeff on Thu Jun 21, 2012 9:24 am; edited 1 time in total |
|
| Back to top |
|
|
| mca |
| Posted: Thu Jun 21, 2012 9:13 am Post subject: |
|
|
Disciple
Joined: 09 Mar 2005
Posts: 196
|
| We are using a supported NFS version, NFS4 |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Thu Jun 21, 2012 11:39 am Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
Can you display the full ownership of '/HA/WMQ/QMGR1/amqmfsck.lck' on all nodes of the cluster?
Can you display the full ownership of the files in '/HA/WMQ/QMGR1/active' on all nodes of the cluster?
Find out why it does not match throughout the cluster... 
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| mca |
| Posted: Thu Jun 28, 2012 10:41 am Post subject: |
|
|
Disciple
Joined: 09 Mar 2005
Posts: 196
|
In one of the form, someone suggested to mount the data and logs seperately rather than one Home directory mount.
i.e.
instead of
mount -t nfs -o hard,intr 10.2.100.73:/ /HA
use
mount -t nfs -o hard,intr 10.2.100.73:/HA/WMQ/QMGR1/data /HA/WMQ/QMGR1/data
mount -t nfs -o hard,intr 10.2.100.73:/HA/WMQ/QMGR1/logs /HA/WMQ/QMGR1/logs
it worked for me. Now there are no "mqm" id mismatches. But i ran into new issue here. When i try to mount "data" and "logs" sub directories separately with above commands, the "data" on MQ Guest is being mounted to "data" on NFS Guest which is right. But the "logs" on MQ Guest is being mounted to "data" instead of "logs" on NFS Guest. We tried recreating the mount on NFS and different ways, but it still acts the same.
Not sure why the mount is creating the problem |
|
| Back to top |
|
|
| mca |
| Posted: Mon Jul 16, 2012 10:09 am Post subject: |
|
|
Disciple
Joined: 09 Mar 2005
Posts: 196
|
We fixed the mount point issue. the problem is that we are suing fsid=0 (File System ID) on both of them. Atleast one should be set to something else (1-255). We set fsid=1 on second mount point and it worked.
I was finally able to create the multi-instance Queue Managers on 2 Linux Machines. When i do dspmq on both one says "Running" and other says "Running elsewhere".
When i end the Queue where where it says "Running", the other also ends. Doesn't the other one supposed to pick up from here? |
|
| Back to top |
|
|
| exerk |
| Posted: Mon Jul 16, 2012 10:13 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| mca wrote: |
| When i end the Queue where where it says "Running", the other also ends. Doesn't the other one supposed to pick up from here? |
And the documentation says what about ending MI queue managers? (note I state ending, not failing over). Have you tried power-interrupting the running-instance server to see what happens?
_________________
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 |
|
|
| mca |
| Posted: Mon Jul 16, 2012 10:18 am Post subject: |
|
|
Disciple
Joined: 09 Mar 2005
Posts: 196
|
| I now tried shutting down the Linux machine and the QM instance on other Linux Guest also "Ended unexpected". I was expecting it to say "Running" from "Running elsewhere". |
|
| Back to top |
|
|
| exerk |
| Posted: Mon Jul 16, 2012 10:27 am Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
| mca wrote: |
| I now tried shutting down the Linux machine and the QM instance on other Linux Guest also "Ended unexpected". I was expecting it to say "Running" from "Running elsewhere". |
So you started both instances with strmqm -x and ended the running instance with endmqm -s and both shut down when you tried it controlled? And note, I didn't say 'shut-down' the other Linux machine, I said 'power-interrupt' - they're two different things.
_________________
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 |
|
|
| mca |
| Posted: Mon Jul 16, 2012 10:32 am Post subject: |
|
|
Disciple
Joined: 09 Mar 2005
Posts: 196
|
Earlier, i started MI QM's using "strmqm -x" option. But i ended normally and other time i shut down linux guest. Now i tried "endmqm -s" option and it says
"AMQ7276: WebSphere MQ queue manager cannot switch over." |
|
| Back to top |
|
|
| exerk |
| Posted: Mon Jul 16, 2012 12:13 pm Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
Then there is still a problem with your set up.
_________________
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 |
|
|
| mca |
| Posted: Tue Jul 24, 2012 12:36 pm Post subject: |
|
|
Disciple
Joined: 09 Mar 2005
Posts: 196
|
My Set Up is quite simple. We have three separate z/Linux Machines that holds NFS Server, Active MI Queue Manager and Standby MI Queue Manager. The Queue Manager Name is QMGR1.
On NFS Server
I. Create the required mount point and assign it 10GB space
/dev/mapper/vgmqha-lvmqha 9.9G 200M 9.2G 3% /HA/WMQ/QMGR1
II. Create the shared directories and their ownership
mkdir -p /HA/WMQ/QMGR1/data
mkdir -p /HA/WMQ/QMGR1/logs
chown -R mqm:mqm /HA
chmod -R ug+rwx /HA
III. Add the mount to /etc/exports to make it available to other Linux machines
Vi /etc/exports
/HA/WMQ/QMGR1 *(rw,fsid=0,no_wdelay,sync)
/etc/init.d/nfsserver stop
/etc/init.d/nfsserver start
/etc/init.d/nfsserver start
On Middleware Linux Machines
I. Make sure the mount is accessible, create directories and mount it (On both Linux Machines that host Active and Standby Queue Managers)
Showmount –e [NFS Server IP]
Export list for [NFS Server IP]:
/HA/WMQ/QMGR1 *
mkdir –p /HA/WMQ/QMGR1/data
mkdir –p /HA/WMQ/QMGR1/logs
chown –R mqm:mqm /HA
chmod -R ug+rwx /HA
mount -t nfs -o hard,intr [NFS Server IP]:/HA/WMQ/QMGR1 /HA/WMQ/QMGR1
II. Create Multi-Instance Queue Manager QMGR1 on one Linux Machine that holds Active MI Queue Manager
crtmqm -md /HA/WMQ/QMGR1/data -ld /HA/WMQ/QMGR1/logs QMGR1
III. Display the properties of QMGR1 on the same Machine
dspmqinf -o command QMGR1
The Output will be …
addmqinf -s QueueManager -v Name= QMGR1 -v Directory= QMGR1 -v Prefix=/var/mqm -v DataPath=/mqha/WMQ/QMGR1/data/QMGR1
IV. Copy the above output and create a reference of this Queue Manager on Second Linux Machine that holds Standby MI Queue Manager
addmqinf -s QueueManager -v Name= QMGR1 -v Directory= QMGR1 -v Prefix=/var/mqm -v DataPath=/mqha/WMQ/QMGR1/data/QMGR1
The Multi-Instance Queue Manager QMGR1 is created.
V. Display the Queue Managers on both Middleware Linux Machines and start them
dspmq
QMNAME(QMGR1) STATUS(Ended immediately)
strmqm -x QMGR1
On Active QM Linux Machine
runmqsc QMGR1
define listener(QMGR1.LISTENER) trptype(tcp) port(1414) control(qmgr)
START LISTENER(QMGR1.LISTENER)
dspmq
QMNAME(QMGR1) STATUS(Running)
On Standby QM Linux Machine
dspmq
QMNAME(QMGR1) STATUS(Running Elsewhere)
On Active QM Linux Machine
endmqm -x QMGR1
The result is
On Active QM Linux Machine
dspmq
QMNAME(QMGR1) STATUS(Ended Normally)
On Standby QM Linux Machine
dspmq
QMNAME(QMGR1) STATUS(Ended Unexpectedly)
I was expecting the standby Queue Manager to be started at this point, but it also ends abruptly. |
|
| Back to top |
|
|
| exerk |
| Posted: Tue Jul 24, 2012 1:20 pm Post subject: |
|
|
Jedi Council
Joined: 02 Nov 2006
Posts: 6339
|
And you most definitely issued the strmqm -x command on both servers? And there is nothing in the logs, both queue manager and general error logs?
_________________
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 |
|
|
|
|
