MMP Performance Issue After Upgrading to 220.127.116.11/18.104.22.168 with Single CPU Server
(Doc ID 2358403.1)
Last updated on MARCH 12, 2018
Applies to:Oracle Communications Messaging Server - Version 8.0.2 and later
Information in this document applies to any platform.
We are in the process of upgrading from Messaging Server 7.0.5 to 8.0.2 and are seeing a significant performance reduction on our test 8.0.2 MMP system.
- This problem only manifests on messaging server with a single CPU.
- Strace on the AService process resolves the issue.
- Changing the message store to 2 CPUs also resolves the issue.
We suspect the issue has something to do with threading on a single CPU and messaging server 8.0.2.x on Linux driving this behavior.
After changing these options, we do observe improved pop3 performance:
Interestingly, when running vmstat the "r" column under "kthr" would never go above 1 when running in epoll() mode. Now with poll() the number can be >1. This also serves to illustrate a race condition or bottleneck in epoll() implementation.
To view full details, sign in with your My Oracle Support account.
Don't have a My Oracle Support account? Click to get started!