ASG
IBM
Zystems
Cressida
Icon
Netflexity
 
  MQSeries.net
Search  Search       Tech Exchange      Education      Certifications      Library      Info Center      SupportPacs      LinkedIn  Search  Search                                                                   FAQ  FAQ   Usergroups  Usergroups
 
Register  ::  Log in Log in to check your private messages
 
RSS Feed - WebSphere MQ Support RSS Feed - Message Broker Support

MQSeries.net Forum Index » Clustering » [b]Clustering and Failover[/b]

Post new topic  Reply to topic
 [b]Clustering and Failover[/b] « View previous topic :: View next topic » 
Author Message
cstubbs
PostPosted: Mon Aug 19, 2002 10:20 am    Post subject: [b]Clustering and Failover[/b] Reply with quote

Newbie

Joined: 16 May 2002
Posts: 9
Location: San Antonio

A server sends messages to a clustered queue on the mainframe systems where there are three qmgrs in the cluster. Where there's a qmgr, there's a CICS system that will read the messages. If a CICS dies, will MQSeries know about it and stop sending messages to that particular qmgr?

Any help would be greatly appreciated.
Back to top
View user's profile Send private message
kolban
PostPosted: Mon Aug 19, 2002 10:52 am    Post subject: Reply with quote

Grand Master

Joined: 22 May 2001
Posts: 1072
Location: Fort Worth, TX, USA

The messages will still be sent to the clustered queues even if there is no receiver there to consume the messages. Queues are only excluded from the distribution if there no active or startable channel between the local queue manager and the target.
Back to top
View user's profile Send private message
jc_squire
PostPosted: Mon Aug 19, 2002 4:48 pm    Post subject: Reply with quote

Centurion

Joined: 14 Apr 2002
Posts: 105
Location: New Zealand

not sure how you would do this on a mainframe but if the CICS system dies you can disable the Q's PUT attribute. This change will replicate through the cluster and the sending qmgr will see that the queue is not available and messages will be routed to the 2 "available" qmgrs. The issue here is that if the replication is delayed the sending qmgr will not see this change immediately. Normally replication only takes a few seconds.

another solution is to stop the chl when CICS goes down which will also force messages to be routed the way you want but this affects all msg traffic - not just messages destined for CICS.

On unix/linux we run scripts in the crontab that check if the process (e.g. CICS) is active and if not then disable the PUT attribute.

Regards
_________________
J C Squire
IBM Certified Specialist - MQSeries
Back to top
View user's profile Send private message
oz1ccg
PostPosted: Mon Aug 19, 2002 10:22 pm    Post subject: Reply with quote

Yatiri

Joined: 10 Feb 2002
Posts: 628
Location: Denmark

Why not using a shared queue ?

It has all the advantages.... It's available to all qmgrs in the QSG and if a server dies, the rest just processes the data....

There are limits, the size of msgs, the size of the coupling facillity(CF) maybe more...

Just my $0.02
_________________
Regards, Jørgen
Home of BlockIP2, the last free MQ Security exit ver. 3.00
Cert. on WMQ, WBIMB, SWIFT.
Back to top
View user's profile Send private message Send e-mail Visit poster's website MSN Messenger
bmccarty
PostPosted: Wed Aug 21, 2002 6:50 pm    Post subject: Reply with quote

Apprentice

Joined: 18 Dec 2001
Posts: 43

If you can't do the shared queues because you need to upgrade the hardware before using the coupling-facility...

How could you tell if a CICS transaction failed...

and if you did how do you automatically/adminstratively stop the incoming channel? Do you use a PCF program?

Any detailed suggestions would be appreciated.

Thanks,

B
Back to top
View user's profile Send private message AIM Address MSN Messenger
oz1ccg
PostPosted: Thu Aug 22, 2002 2:23 am    Post subject: Reply with quote

Yatiri

Joined: 10 Feb 2002
Posts: 628
Location: Denmark

Well, Is far as I can see it, if you're trying to follow the disabling way... there are a lot of pitfalls on the road:

Let's say our automation systems recognize that the CICS-application is dead: set local cluster q put inhibit, ok.

Now when the QLOCAL is put inhibitted, it's time to move the waiting messages to the other queues in the cluster, this can be done in same manner as mqclcyp.ccp does it, and we're thru.
(mqclcpy.ccp uses the functionallity that it's allowed to open the same queue for both INPUT and OUTPUT, and when it's a cluster Q, you can put to it even if the local copy is put-inhibit, and the data is routed to the remote instance.)

When the CICS is back in bussiness, it's just remeove the putinhibit, and we're back to normal.

mqclcpy.ccp can be found at:
http://d1o111.dk.telia.net/~u149101068/tips_and_tricks.htm

Just my $0.02
_________________
Regards, Jørgen
Home of BlockIP2, the last free MQ Security exit ver. 3.00
Cert. on WMQ, WBIMB, SWIFT.
Back to top
View user's profile Send private message Send e-mail Visit poster's website MSN Messenger
Display posts from previous:   
Post new topic  Reply to topic Page 1 of 1

MQSeries.net Forum Index » Clustering » [b]Clustering and Failover[/b]
Jump to:  



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
Protected by Anti-Spam ACP
 
 


Theme by Dustin Baccetti
Powered by phpBB © 2001, 2002 phpBB Group

Copyright © MQSeries.net. All rights reserved.