ePO communicating with agents sporadically - "server is too busy" and "failed to connect"

[pierce]

Hey All,

Since Friday my ePO (5.1.0) has been throwing strange errors, nothing was changed to this server to my knowledge.

The server.log shows the follow sorts of errors repeatedly

20151006093336	E	#05260	NAIMSERV	naSPIPE.cpp(338): Failed to connect to "internal IP address":8081, network error was 10060
20151006093336	E	#05260	NAIMSERV	Server failed to wake up machinename.domain.com, detailed error = -999
20151006093337	E	#06244	MOD_EPO	Server is too busy (245 connections) to process request
20151006093337	E	#03120	MOD_EPO	Server is too busy (245 connections) to process request

Also checking agent handler page i keep seeing my agent handler in DMZ and on ePo going into inactive mode, restarting the apache process brings it back out of inactive mode but this is every twenty mins or so.

McAfee told me to add more ram as its a resource issue(server has 8GB, SQL is on another box and its currently using 4GB).

[tucker84]

I believe this was resolved in 5.1.1 or 5.1.2 waiting for the KB to load now to confirm. Hotfix 1 also resolves a fair amount of issues. Sorry the KB will not load so I'm not able to verify the fixes. Do you have any hotfixes installed? I'll try to add more once the articles load.

[pierce]

I thought I had a hotfix or two installed, ill go off and look. Seems strange that server has been running fine for a year and now suddenly this!

And I do have a plan to upgrade, unfortunately the last 5 attempts have all failed miserably, even with McAfee support.

[tucker84]

Were the failed updates to 5.1.1 or a different version? If you tried to update and had to roll back it's possible the hotfixes were removed.

[pierce]

Yeah failed going to 5.1.1, however we had snapshotted the server so just restored to the known good state with latest hotfixes.

Ill get back on phone to mcafee and see what they say, something must have changed to cause this just need to work out what!

[pierce · Accepted solution]

Found the issue, buried in the logs were one event mentioning LDAP failure.

Sure enough one of our LDAP connections was failing as someone changed the service account password, updated the password and tested. then restarted ePO and Agent handler and everything is working well again.