JSS Connection drops several times a day

[gspiese]

Using JSS v8.52 and Mac OS X Server v10.5.8

Using WedAdmin, Casper Remote, or Casper Admin stalls several times a day. This happens when trying to authenticate or if a connection has been established, whatever we try to do fails. When using Casper Admin.app it just hangs and we have to force quit it.

We have been just letting the Server sit, doing nothing, and then after 10 minutes or so, we can again connect to it, and use it just fine. I've done a traceroute to it when it's not connecting and when it is, and the route is the same, so I am accesssing the box either way.

Any clues on what might be causing this problem, and possible solutions? We are aware that our Server Version of 10.5.8 isn't what's recommended to run JSS 8.52, but am wondering if that's the cause of this particular problem.

[jarednichols]

I'd be interested in server-side logging during these dead times. Are you able to ssh or console into the server during these periods? It would help to find out if it's a JSS-specific issue or something bigger on the server.

[rockpapergoat]

i've seen similar behavior on installs where mysql pegs the cpu during heavy load, like if you have tons of clients all updating inventory at the same time.

use top on the server to check load. if mysql is sitting there churning away all day, that could be the problem.

[mm2270]

We are aware that our Server Version of 10.5.8 isn't what's recommended to run JSS 8.52, but am wondering if that's the cause of this particular problem.

I was about to post that I didn't think Casper Suite 8.5 was supported on 10.5.x Server, but I see you already know that. I have a feeling that may be at least contributing to your problem. I would consider moving up to 10.6 server if you can.
Have you checked in with your JAMF AM about this to see what they think?

[gspiese]

To test a possible AD issue we created a local account on JSS. During the next stall, I was successful at connecting to the Web Admin with the local account while the AD account stalled. I'm off to talk to the AD folks. As soon as they here it's a Mac, they blame the Mac... sigh. I'll let you know.

[tsd25108]

We had a problem previously with the Java module that runs AD auth in tomcat crashing. We never actually could figure out what caused it, but my theory was that it had something to do with one of the OS patches we applied since it only happened after we'd applied that patch and apple had recalled it and re-released it due to issues with the original.

[gspiese · Accepted solution]

The original configuration of LDAP in JSS was set incorrectly- pointing to an DC that was too many hops away, which wa causing our time out settings in JSS to be violated and stop waiting for an answer. Getting the correct and closest DC resolved the issue of the drips. Now I'm waiting for the creation of our casperadmin service account in AD to be created... It's only been three weeks since the request was submitted!