Thursday, 9 April 2015

MQOPEN ended with reason code 2085 - WebSphere MQ and Clustered Queues

I hit this problem earlier today: -

/opt/ibm/mqm/usr/mqm/samp/bin/amqsput ESB.TO.BPM.CLQ ESB1.DEV1

Sample AMQSPUT0 start
target queue is ESB.TO.BPM.CLQ
MQOPEN ended with reason code 2085
unable to open queue for output
Sample AMQSPUT0 end


I could see the queue on the Queue Manager local to where I was running the command ( my ESB box, running IIB 9 and MQ 8 ).

However, when I looked more deeply: -

display queue(ESB*)

     1 : display queue(ESB*)

AMQ8409: Display Queue details.

   QUEUE(ESB.TO.BPM.CLQ)               TYPE(QLOCAL)

display queue(ESB.TO.BPM.CLQ)

     2 : display queue(ESB.TO.BPM.CLQ)

AMQ8409: Display Queue details.

   QUEUE(ESB.TO.BPM.CLQ)                   TYPE(QLOCAL)
   ACCTQ(QMGR)                             ALTDATE(2015-04-07)
   ALTTIME(13.28.28)                       BOQNAME( )
   BOTHRESH(0)                             CLUSNL( )
   CLUSTER(UAT1)                           CLCHNAME( )
   CLWLPRTY(0)                             CLWLRANK(0)
   CLWLUSEQ(ANY)                           CRDATE(2015-04-07)
   CRTIME(13.28.28)                        CURDEPTH(0)
   CUSTOM( )                               DEFBIND(NOTFIXED)
   DEFPRTY(0)                              DEFPSIST(YES)
   DEFPRESP(SYNC)                          DEFREADA(NO)
….
 
I could immediately see the problem ….

The cluster to which I was expecting the Queue to belong was called DEV1 and yet the CLUSTER() value above was set to UAT1.

I verified this: -

display CLUSQMGR(*)

     5 : display CLUSQMGR(*)

AMQ8441: Display Cluster Queue Manager details.

   CLUSQMGR(LRPO1.DEV1)                  CHANNEL(DEV1.LRPO1.DEV1)
   CLUSTER(DEV1)

AMQ8441: Display Cluster Queue Manager details.

   CLUSQMGR(REP1.DEV1)                   CHANNEL(DEV1.REP1.DEV1)
   CLUSTER(DEV1)

AMQ8441: Display Cluster Queue Manager details.

   CLUSQMGR(REP2.DEV1)                   CHANNEL(DEV1.REP2.DEV1)
   CLUSTER(DEV1)

which confirmed that the cluster name should indeed be DEV1.

I changed the Queue definition: -
 
alter qlocal(ESB.TO.BPM.CLQ) cluster(DEV1)

and re-tested: -

/opt/ibm/mqm/usr/mqm/samp/bin/amqsput ESB.TO.BPM.CLQ ESB1.DEV1

it responded as I'd expected: -

Sample AMQSPUT0 start
target queue is ESB.TO.BPM.CLQ

Hello BPM from ESB
Sample AMQSPUT0 end

which is nice.

Thanks to this: -


for making me think about the Cluster name within the Queue definition: -

Slide 44 says: -

In order to make a queue visible to the other members of the cluster, it is necessary to ALTER the queue and specify the cluster name:

ALTER QLOCAL(CQ1) CLUSTER('CLUSTER1')

You do a remote PUT into a clustered queue. However, you cannot do a remote GET from a clustered queue.

The GET can ONLY be done from the local queue.


No comments:

Note to self - use kubectl to query images in a pod or deployment

In both cases, we use JSON ... For a deployment, we can do this: - kubectl get deployment foobar --namespace snafu --output jsonpath="{...