For anyone who didn't see it, IBM put out an interesting piece on a fix for an older issue around intermittent retry loops occurring in certain circumstances in IBM MQ channel instances.
If you've encountered intermittent retry loops when establishing new channel instances between multi-instance queue managers, you're not alone. IBM MQ users have reported channels entering retry mode intermittently, especially in multi-instance environments. These issues can manifest as errors such as "AMQ9524E: Remote queue manager unavailable" and "AMQ9202E: Remote host not available, retry later."
This behavior arises due to a logic error in handling comma-separated multiple connames in channel definitions. When one conname is unreachable and the system attempts to reconnect using another conname, the remote side may not immediately retry the connection. Consequently, users may experience delays and interruptions in their messaging systems.
IBM has identified and addressed this issue with a targeted fix in the following PTFs:
Version 9.2 LTS: 9.2.0.25
Version 9.3 LTS: 9.3.0.20
Version 9.x CD: 9.3.6
The fix corrects the inbound channel logic to properly communicate errors to the remote peer, ensuring prompt retries in case of failures. For users affected by this issue, obtaining the latest maintenance from IBM MQ Recommended Fixes is recommended.
By addressing these intermittent retry loops, IBM MQ aims to enhance the reliability and performance of channel instances in multi-instance environments. Stay updated with the latest fixes and maintenance releases to optimize your messaging infrastructure's stability and efficiency.
------------------------------
Scott Treggiari
Avada Software
------------------------------