[AMQ-5529] Deadlock using JMX & copyMatchingMessagesTo

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Incomplete
• Affects Version/s: 5.10.0
• Component/s: JMX

Environment

Linux app1.syd.acx 2.6.39-200.24.1.el6uek.x86_64 #1 SMP Sat Jun 23 02:39:07 EDT 2012 x86_64 x86_64 x86_64 GNU/Linux

java version "1.7.0_60"
Java(TM) SE Runtime Environment (build 1.7.0_60-b19)
Java HotSpot(TM) 64-Bit Server VM (build 24.60-b09, mixed mode)

• Assignee: Unassigned
• Reporter: Paul Smith
• Created: 20/Jan/15 09:08
• Updated: 19/Apr/16 18:04
• Resolved: 19/Apr/16 18:04

Description

Today we had an issue relating to a consumer that was somehow not properly ack'ing messages and letting the queue grow.

We had the idea to use JMX operations to move messages we knew had been delivered and processed, but before hand tried to copy a few of the messages to a new destination to confirm using a selector of "JMSTimestamp<[timestampAsLongValueOfYesterdayMorning]" (obviously a long-based timestamp I've since forgotten to write down.

I tested the Selector with a browse(String selector) operation, and that successfully returned the small handful easily.

I invoked the copyMatchingMessagesTo Operation on the Queue JMX operation, and the attached Thread Deadlock occurred. At the time there were around 20,000 messages in the queue, all very small in size.(roughly 1k each as reported by the JMX averageMessageSize property)

This doesn't give us much confidence in the ability to use Operational JMX to recover from strange scenarios. The total number of messages being copied was a small handful, 5-20 at most.