| Author |
Message
|
| mattynorm |
 Posted: Wed Feb 13, 2008 11:46 am Post subject: Automated channel failover - replacing a scripted process |
 |
|
Acolyte
Joined: 06 Jun 2003
Posts: 52
|
Apologies if this has been asked before
I've been asked to look at a piece of work for a client. Their current MQ set up (version 5.3 CSD 5, Windows) is that they have an application putting to an alias queue (where the target queue is a remote queue) on a (non-clustered) queue manager. This queue manager connects to a remote queue manager, owned by a third party. They currently have a script which polls the channel (using a loop, which is chewing a fair bit of cpu) for it's current status, and if they think the connection may have failed, then pings the channel to confirm this. If the connection is indeed unavailable, they then change the target queue in the alias to another remote queue, pointing to a different queue manager (owned I think by the same third party).
Although this seems to work for them, they have been informed by another consultant that under version 6 this can all be automated (part of the work is to move this queue manager and 1 other to version 6). Is this possible? I've had a look in the documentation, but I can't see anything that would provide any automated channel failover of the kid they are requesting.
Thanks in advance for any help. |
|
| Back to top |
|
 |
| PeterPotkay |
| Posted: Wed Feb 13, 2008 12:29 pm Post subject: |
|
|
Poobah
Joined: 15 May 2001
Posts: 7722
|
|
| Back to top |
|
|
| jefflowrey |
| Posted: Wed Feb 13, 2008 12:58 pm Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
Channel Events?
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| mattynorm |
| Posted: Thu Feb 14, 2008 1:20 am Post subject: |
|
|
Acolyte
Joined: 06 Jun 2003
Posts: 52
|
I don't think any clustering solution will work, because the destination queue managers will be outside the cluster, and thus all messages will go to the remote queue \queue manager with the highest priority, and will sit on the transmission queue to the destination queue manager. At least that's how it's worked when I've tested it this morning.
I'm looking towards the channel events, though to be honest I don't know a lot about them. An IBM contact has alos said there is a fixpack containing a channel exit program with the capabilitity to stop a channel and start another one. Thus 2 channels with the same XMITQ and this fixpack running may be the best solution I can offer. |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Thu Feb 14, 2008 1:40 am Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
| mattynorm wrote: |
I don't think any clustering solution will work, because the destination queue managers will be outside the cluster, and thus all messages will go to the remote queue \queue manager with the highest priority, and will sit on the transmission queue to the destination queue manager. At least that's how it's worked when I've tested it this morning.
I'm looking towards the channel events, though to be honest I don't know a lot about them. An IBM contact has alos said there is a fixpack containing a channel exit program with the capabilitity to stop a channel and start another one. Thus 2 channels with the same XMITQ and this fixpack running may be the best solution I can offer. |
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Thu Feb 14, 2008 2:16 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
| mattynorm wrote: |
| An IBM contact has alos said there is a fixpack containing a channel exit program with the capabilitity to stop a channel and start another one. Thus 2 channels with the same XMITQ and this fixpack running may be the best solution I can offer. |
It's likely not a "fixpack" but a "supportPac". Although I can't say I know which one, off the top of my head.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| mattynorm |
| Posted: Thu Feb 14, 2008 2:53 am Post subject: |
|
|
Acolyte
Joined: 06 Jun 2003
Posts: 52
|
| Quote: |
| It's likely not a "fixpack" but a "supportPac". |
Yes, sorry, it is of course a support pac. I think the pac in question is MR01, which states that it is for "Creating a HA Link between WebSphere MQ and a Service Integration Bus", and says that a prerequisite is WAS. The description of it though is
| Quote: |
This SupportPac provides a simple channel message exit (referred to as 'the Exit' for the rest of this document), which will cycle through a list of possible CONNAME values, copying each one into the channel definition in turn until a connection between the queue manager and the target ME is made. Each time the connection fails the channel will retry and the Exit will select a new potential CONNAME value. The Exit is written in the C (for distributed platforms) and Assembler (for the z/OS platform) programming languages and is provided in source code format only and can be modified (depending on requirements) and installed on the desired platform.
|
So I don't really see why it wouldn't work between queue managers only. |
|
| Back to top |
|
|
|
|
