It sounds like a Zookeeper issue, which I guess you already suspect.  

Maybe Zookeeper is not functioning properly on Server 5.  Maybe Server 4 cannot talk to Server 5.

https://www.novell.com/documentation/zenworks-2020/zen_sys_servers/data/t4f1y9ouwt65.html

https://www.novell.com/documentation/zenworks-2020/zen_sys_servers/data/t4f1y9ov00nk.html

I Craig,

Thank for your reply

I thought about it at first but when I did the “zman srzlc”.

the “zman srzuc” command was run just in case but we still have the problem.

What I don't understand is that the problem only occurs when servers are switched off.

My degraded mode is designed to allow part of the servers to be online to allow services to continue despite the loss or damage of servers.

Kind Regards,

For reasons I don't know....I suspect Zookeeper is not functioning properly when those servers are off, even though in theory it should be.

Zookeeper handles locks...If locks are not being handled....it could spike the DB.

You may need an Service Request.....They would likely want to extra log debugging as shown in my second link.

It could be something as simple as a FireWall issues between 4 and 5.....or more complex...

I might also want to test what happens if you only leave 1 and 2 running.....

That would help isolate if it was an issue with losing 2 of 3 Zookeeper nodes or if it was something about Zookeeper on 5.

We've checked with support in the database and on the server configurations and it's fine from that point of view.

Servers 1, 2 and 3 are on one server network and located on one physical site, and servers 4 and 5 are on another physical site and on another server network. During normal operation, the servers communicate correctly with each other.

I don't think the problem is the network.

From my point of view, it's the servers remaining up in the zone that saturate the connections with the database and block the release of the ZENworks Client management service, but I don't understand why.

Is it possible that the last remaining server is saturated because the 2 switched-off servers with zookeeper can't communicate?
If so, is there any way to free up resources or work around the problem?

Really it would be necessary to work with Support and Analyze the logs.

Also the DBA could look at what is going on with the server.

--

Logically, your configuration should work.....The only concern would be if the connection between Primaries 4 and 5 and the DB was insufficient.  Presumably, Servers 4 and 5 are at a "DR" SIte.....but in such configurations, there is usually a backup DB as well there since the main DB would also be down....though it would take some manual work to get those servers over.

--

If your case has not been escalated to a backline engineer, I might suggest doing that.  They can work with development to figure out where the issue lies.  

Really it would be necessary to work with Support and Analyze the logs.

Also the DBA could look at what is going on with the server.

--

Logically, your configuration should work.....The only concern would be if the connection between Primaries 4 and 5 and the DB was insufficient.  Presumably, Servers 4 and 5 are at a "DR" SIte.....but in such configurations, there is usually a backup DB as well there since the main DB would also be down....though it would take some manual work to get those servers over.

--

If your case has not been escalated to a backline engineer, I might suggest doing that.  They can work with development to figure out where the issue lies.