Zenoss ZenTech Community

I am running a 5.1.2 Core System set up as a 3-host Zookeeper Cluster, and I just recently started having issues with Zope.

Normally, the number of TCP connections ranges between 100-120, and there's no issue at all. Then suddenly there is a spike in the number of TCP connections, sometimes over 500, that occurs within a few minutes and crashes the system. Restarting the Zope service resolvesthe issue, but there's nothing really in any of the logs that I see causing this to occur.

Has anyone had this happen before Looking for some guidance on how to fix the issue.

Subject:	In addition to Zope's sudden
Author:	[Not Specified]
Posted:	2016-08-04 11:39

In addition to Zope's sudden increase in connections, zenhub, zenpython and both metric shippers also have corresponding connection increases at the same time, according to the control center graphs.

Subject:	I am still having this issue,
Author:	[Not Specified]
Posted:	2016-09-16 08:46

I am still having this issue, if anyone is able to help.

Subject:	When it crashes, can you take
Author:	Jay Stanley
Posted:	2016-09-19 09:51

When it crashes, can you take a look at one of the event.log's on a zope container

Is there a spike of events as well

Subject:	The event.log shows a lot of
Author:	[Not Specified]
Posted:	2016-09-19 10:52

The event.log shows a lot of lines like the following:

2016-09-19T15:31:33 WARNING urllib3.connectionpool HttpConnectionPool is full, discarding connection: 127.0.0.1
------
Other than that, nothing out of the ordinary.

Also, there isn't any spike in events, as far as I've noticed.

Subject:	No, ERRORs when zope crashes?
Author:	Jay Stanley
Posted:	2016-09-19 13:43

No, ERRORs when zope crashes Can you see what is causing all the connections in Z2.log

Wonder if there is anything useful in /var/log/messages

Subject:	Zope Crashing
Author:	[Not Specified]
Posted:	2016-08-03 13:33

Zope Crashing