SMTP Monitor sometimes failed...

Hi all,

 

Currently I run in a very strange issue. We are using LTM v11.6.0 HF5 for loadbalancing of multiple mail servers (postfix).

 

This week I checked the /var/log/ltm logs and realized that from time to time the standard SMTP monitor (tcp/25) will fail for a very short period of time (5-15 sec max.) so you see log messages like "no pool members available". After that the monitor will recover quickly. The funny thing is that the same monitor will be used for multiple pools. But only two pools are having this issue, so it think the monitor in general should be fine.

 

I already adjusted the SMTP monitor intervals to 30sec and timeout value to 121sec. It's much better now but it could still happen (5 times a day) that the SMTP monitor will fail. I also have a default node monitor with ICMP for all nodes active without any issues, so it seems that the network between the LTM and the servers is fine.

 

Any ideas or suggestions? I will try to enable SMTP monitor debug tomorrow. Maybe I will use some additional TCP or TCP half-open monitors on TCP port 25 for testing as well.

 

Thanks a lot for feedback Regards Thrillseeker