Server has slow response then normal response

[Acxiom Unix Team]

Hi,

We have a vPar on an rp8440 running HPUX 11.31. This morning NIS decided to spawn over 500 duplicate processes of the ypserv daemon and filled up /var in the process.

After killing the NIS processes manually and clearing out space in /var a new problem has arisen. The box seemed to return to normal as far as being able to login, respone time to commands etc.

However, it seems to be going through phases where the response time is very, very slow (even an ls command can take 5 minutes) then return to normal and then go slow again.

I cannot see anything in the syslog, replies from vmstat, swapinfo and top do not indicate the box is over stressed with any processes. So i am wondering if anyone has ever come across a similar issue before?

[Hakki Aydin Ucar · Accepted solution]

lots of things can cause this situation, a process OR processes OR a backup process, do you have a glance tool to find root cause ? If yes check bottleneck of CPU , Disk , Network and process .
-the simple way to solve this issue to make a reboot if you can in most situations like this.

[Rita C Workman]

Isn't RPC used to talk to and from the ypserv?

Might check and see if you've got hung RPC processes...

Just a thought,
Rita

[Acxiom Unix Team]

Closing this now... We managed to get a reboot done to cure the problems.

FYI, when we had the problems TOP, vmstat and swapinfo were not reporting any obivous issues. So guessing as Rita said maybe an RPC process had hung or something.

Anyway, its all clear now. Thanks for the input.