JMS IBM MQ-限制浏览时返回消息的大小

Question

In my scenario the queue contains huge (32MB+) messages, and the data is structured such that the information that identifies the type of message, is contained in the first few bytes of the message.

When using the low-level Websphere MQ API it is possible to provide a buffer of a specific (limited) size, then issue the MQGET call with options that include the MQGMO_ACCEPT_TRUNCATED_MSG flag. The reason code returned by the MQGET would then probably be MQRC_TRUNCATED_MSG_ACCEPTED (which is to be expected).

This enables one to write a very fast "browse all" kind of logic that returns the first (lets say) 100 bytes of each message instead of the full buffer for each.

Is it possible to do something similar using java JMS? I have not been able to find anything that allows me to specify a buffer size on either the session.createBrowser(queue...) or the browser.getEnumeration() calls.

Thanks.

Answer 1

Instead of selecting a pattern then defining your data, you have data and want to force it to fit a pattern. That's a little backwards.

Questions:

(1) Are you putting different types of data into the same queue? ie multiple and different type of senders?

(2) Is the sender (publisher) a JMS application?

(3) Are you trying to create an application that will move/filter messages like what IIB (Message Broker) would do? If not, then what is the real purpose of your program?

---

Each unique application should have its own queue. Queues should be handed out like candle to kids on Halloween.

ie Lets say HR has the following 5 applications:

• Recruiting
• Payroll
• Benefits
• Training
• Employees

Do not create a queue called 'HR.Q' that will process all of the HR messages. Bad design, bad to support.

Rather each unique application should have its own queue. ie

• 'HR.RECRUITING.Q'
• 'HR.PAYROLL.Q'
• 'HR.BENEFITS.Q'
• 'HR.TRAINING.Q'
• 'HR.EMPLOYEES.Q'