Trouble writing large messages to MQ series queue from Mercator 6.7

[talltester]

We are running Mercator 6.7(306) on AIX v5.3 and we are having difficulty getting our system into production. The trouble is when we run small test files with just a few transactions, they process just fine with the message getting put on the MQ series queue. However, the problem is that when we try to run a "production size" file (with more than a few hundred transactions), we cannot get the message successfully placed on the MQ series queue. (We will normally be getting files with several thousand transactions per day from our trading partners.) The error we have been getting is:

<TargetReport card="1" adapter="IBM WebSphere MQ (C" bytes="19027038" adapterreturn="-1004">
<Message>Message could not be stored on the queue</Message>
<Settings>-QMN QUEUE.MANAGER -QN EDIInput -T</Settings>
</TargetReport>

Our configuration manager says that the Websphere channels are set as large as they will go, with no effect on our ability to get these "larger" messages on the queue. Any ideas anyone has would be oh so greatly appreciated.

 

[JEMIAZ]

What is the max message size,max message depth and current mesage depth of queue EDIInput?
If EdiInput queue is a remote queue then what is the max size and depth of the related transmit queue and channel.
What are the queue manager log file settings and what, if anything is recorded in the queue manager log?
What does your Mq trace show?

 

[talltester]

Max message size is 100 meg.
Max message depth is 5000.
Current message depth is 0.
Max size of the channel is 100 meg.

MQ Trace log shows:
<913544-22982>: TRACE ON
<913544-22982>: Adapter command: -QMN TWCC.QUEUE.MANAGER -QN EDIInput -T
<913544-22982>: [intm4mqsCheckAdapterCommand]
<913544-22982>: [intm4mqsCheckAdapterCommand] (rc = 0) OK
<913544-24007>: [m4mqsConnect]
<913544-24007>: | Connected successfully to the queue manager TWCC.QUEUE.MANAGER
<913544-24007>: [m4mqsConnect] (rc = 0) OK
<913544-24007>: [m4mqsOnNotify]
<913544-24007>: | OnNotify->PutStart
<913544-24007>: | [intm4mqsOpenQueue]
<913544-24007>: | | Obtaining handle for the queue EDIInput
<913544-24007>: | | Queue EDIInput is the first queue on the connection
<913544-24007>: | | Opening queue EDIInput
<913544-24007>: | | Queue EDIInput was opened successfully
<913544-24007>: | [intm4mqsOpenQueue] (rc = 0) OK
<913544-24007>: [m4mqsOnNotify] (rc = 0) OK
<913544-24007>: [m4mqsBeginTransaction]
<913544-24007>: | MQSeries transaction implicitly started
<913544-24007>: [m4mqsBeginTransaction] (rc = 0) OK
<913544-24007>: [m4mqsValidateConnection]
<913544-24007>: [m4mqsValidateConnection] (rc = 0) OK
<913544-24007>: [m4mqsPut]
<913544-24007>: | [intm4mqsObtainQueueNode]
<913544-24007>: | | Queue node found: EDIInput
<913544-24007>: | [intm4mqsObtainQueueNode] (rc = 0) OK
<913544-24007>: | [intm4mqsIsTransactionRequired]
<913544-24007>: | [intm4mqsIsTransactionRequired] (rc = 0) OK
<913544-24007>: | Syncpoint required
<913544-24007>: | Adapter error rc = -1004 [Message could not be stored on the queue]
<913544-24007>: | MQSeries error (ReasonCode = 2102)
<913544-24007>: [m4mqsPut] (rc = -1004) *** ERROR ***
<913544-24007>: [m4mqsEndTransaction]
<913544-24007>: | Rolling back unit of work
<913544-24007>: [m4mqsEndTransaction] (rc = 0) OK
<913544-24007>: [m4mqsOnNotify]
<913544-24007>: | OnNotify->PutStop
<913544-24007>: [m4mqsOnNotify] (rc = 0) OK
<913544-22982>: Destroying trace object in m4mqsDestroyAdapterInstance
<913544-24007>: [m4mqsDisconnect]
<913544-24007>: | Queue EDIInput was successfully closed
<913544-24007>: | Disconnected from the queue manager TWCC.QUEUE.MANAGER
<913544-24007>: [m4mqsDisconnect] (rc = 0) OK

 

[JEMIAZ]

The real error code is 2102
MQRC_RESOURCE_PROBLEM
(2102, X'836') Insufficient system resources available.

Not sure I've seen this one before. Is anything showing in the queue manager error log? Want is the queue manger max handles set to?