Automatically clear messages from queue in IBM MQ - ibm-mq

I was wondering if there is a way for you to configure a queue to automatically clear messages? We are striving to partially implement a component of our architecture and want to be able to send to the queue, but have the queue automatically remove the messages that are being sent so that we don't have to run scripts, etc to perform the clean-up.
So far the only thing I have been able to find is to run CLEAR QLOCAL or set the messages to expire from the publishing application.

For you use case there are a few options in IBM MQ:
Create a QALIAS that points to a TOPIC object which has a topic string with no subscribers, messages put to the QA will just disappear.
Have the sending application set message expiry.
Use the IBM MQ CAPEXPRY feature to administratively force message expiry at the queue level.
Run a script to issue CLEAR QLOCAL against the queue. There can not be open handles on the queue for this to work.
Programmatically issue the equivalent PCF command to CLEAR QLOCAL against the queue. There cannot be open handles on the queue for this to work.
Run the IBM MQ dmpmqmsg utility against the queue to read and discard the messages.

Related

IBM MQ - Ability to queue messages during maintenance without stopping channel

Hoping someone with MQ knowledge will know if the following is possible ...
Background : we have one IBM MQ 7.5 server currently but happy to upgrade if that helps.
Scenario : We have a dedicated channel and queues for a processing application. I'd like a way, on a single MQ Server, to be able to have new messages delivered and for them to queue up in "Messages_IN" during a release then let them through afterwards. I cannot stop the channel that is used as the consuming application will error (lose MQ access) and not work. i.e. sending applications and the processing application currently use the same channel / queues.
Diagram linked below to help explain ...
MQ options?
Is there anyway to have a flow control on a queue that would allow us to control messages not being consumed by an application without stopping the channel? Any new queues / channels / config etc are fine but currently limited to one MQ Server.
Many thanks for anyone who's taken the time to read this and hopefully its technically possible.
Regards
David
Hi the comment of JoshMC is your answer. You ned to set the "GET" propiertie of the QUEUES "MESSAGES_IN" to "DISABLED". With that you prevent any aplication of get message from the QUEUE, whe you end your maintenance you need to set the propertie to "ENABLED" for enabled applicaiton to gets messages from the queue.

IBM MQ message history

Is it possible to keep a history of messages (with message content would be perfect) that have already been retrieved and are no longer on a queue?
In the application I can see when the sender attempts to put the message in the queue and when the receiver attempts to pick the messages up, but I'd like to see when the message really arrived into the queue and when the messages were really received.
Does MQ Explorer have this function? How would I use it?
What you are looking for is a message tracking/auditing software for IBM MQ. You can find a list of what is available here.
It is possible to use an API exit to make copies of messages in a queue or to audit both PUT and GET operations.
It is also possible to put messages to a topic, then create as many administrative subscriptions to destination queues as required. Something can then GET and log messages from one of those destination queues. The problem with this is that MQ changes the message ID between publication and consumption whereas in a queue it remains static.
There is no native MQ function to capture messages. It's possible to use linear logs and later scrape the logs but these do not necessarily capture all messages due to optimization. (A message PUT to a waiting getter outside of syncpoint for example.) However there is at least one commercial product to scrape linear transaction logs to audit message activity.
The philosophy of MQ in general is that it is the delivery mechanism and deals with envelope data to route and deliver but does not deal with payload data. WAS, IIB and other broker/transformation engines are where IBM has put all of the functions that deal with message payloads.

ibm WebSphere MQ pausing the queue through java api

I need to pause the receiver queue some time (for process the already received messages) in ibm websphere queue through the ibm java api.
If you want to stop Applications from being able to get messages from a Queue you can inhibit the Queue by setting the GET attribute on the Queue to DISABLED. This will still allow applications to put messages to the Queue but no applications will be able to get messages from the Queue. IF you want to stop messages from being able to be put to a queue you can do this by setting the PUT attribute to DISABLED.
When you're ready for the applications to be able to get messages from the Queue again you set the GET attribute to ENABLED and applications will be able to get messages from the Queue.
example MQSC command to set GET to DISABLED
ALTER QLOCAL('') GET(DISABLED)
Be warned: If you inhibit the put or get on a Queue, applications attempting to put or get messages on that Queue will receive an error return code of MQRC_PUT_INHIBITED or MQRC_GET_INHIBITED. If you don't anticipate this your applications could fail.

How to go about messages in Dead Letter Queue

We are using WebLogic 10.3.6.0 and IBM MQ 7.5.
Application design is to send messages to a dead letter queue (in WebLogic) on re-delivery. The re-delivery happens as the first delivery has failed due to some network issue or database data source failure.
My Client wants a way to browse the messages in the dead letter queue from the application GUI and pull them for processing when the network issue or data source issue has been resolved.
What is the best way to go about this?
I cam across QueueBrowser coupled with activemq or some other implementation. Is QueueBrowser possible with WebLogic? Please suggest on best ways to achieve this requirement.
Kindly pardon if my question is too naive. I am only a PL/SQL programmer.
Valerie is referring to the SYSTEM DLQ and application should never ever write to it. Application's should have there own DLQ.
i.e. If your application queue is called 'TEST.Q1' then your application DLQ should be called 'TEST.Q1.DLQ'.
There is a whole long list of MQ tools here to view messages and manage your MQ environment.
Is the application actually designed to write to the DLQ? If so, that is a very poor design. The DLQ is for the queue manager and MQ software to place messages which can not be delivered. The application should not be writing to the DLQ.
As for how to view messages on DLQ, that can be done with the MQ Explorer GUI. Or to write a script, use the DLQ handler (runmqdlq) with a rules table for processing messages.

How to monitor an existing queue from WebSphere MQ?

I have a .NET application that needs to monitor a queue in WebSphere MQ. I need to react to each message without impacting the current process. The client application can't explicity send me the same message.
Can I read a message without removing it from the queue? Can I be notified for each message? Can I configure the MQ to duplicate the current queue?
Is there another solution?
If you are using WMQ v7 then you can do this without any impact to the existing applications other than to change the queue name for one of them.
Currently the message producer and consumer use the same queue. In v7 of WMQ you can create an alias over a topic so that the message producer thinks it's a queue. Then you can create two administrative, durable subscriptions such that one points to the existing input queue and another points to a queue dedicated to your new application.
Of course you are already using v7 since v6 goes out of service next year, right? You can upgrade the QMgr to v7 which enables this behavior while still using v6 client code for the apps.
If you are using WMQ v6 then the MirrorQ program might work for you.
You could change from using a queue to a durable topic and have both your reader and your browser subscribe to it.
You could also create a distribution list on your queue manager. A distribution list is used to send a copy of the same message to multiple queues. You would then have a processing queue and a browsing/monitoring queue.

Resources