topic Re: Redhat ES4 server hangs. in Operating System - Linux

Redhat ES4 server hangs.

GnanaShekar — Tue, 06 Jun 2006 02:03:25 GMT

Hi,

We have a Redhat ES4 server. It is a HP rx4640 server. It has 4 Itanium 2 servers and 8GB RAM.

A user running weblogic tests on this complained a couple of times that the server is hung. Ping to this server used to timeout and I have always restarted.

Now I want to find the root cause and fix the problem.

I found the following. Please suggest.

# more /etc/redhat-release
Red Hat Enterprise Linux ES release 4 (Nahant Update 2)

# last -n 5
skhan pts/11 172.22.45.200 Sun Jun 4 15:04 - crash (14:36)
skhan pts/10 172.22.45.200 Sun Jun 4 14:05 - crash (15:36)
skhan pts/9 172.22.45.200 Sun Jun 4 14:04 - crash (15:36)
skhan pts/10 172.22.45.200 Sun Jun 4 13:42 - 14:04 (00:21)
skhan pts/9 172.22.45.200 Sun Jun 4 13:40 - 14:04 (00:23)

/var/log/messages
-----------------
Jun 4 14:36:18 bangpcia4 kernel: java(678): floating-point assist fault at ip 00000001028f51f1, isr 0000020000000008
Jun 4 14:36:18 bangpcia4 kernel: java(678): floating-point assist fault at ip 00000001028f5201, isr 0000020000000008
Jun 4 14:36:18 bangpcia4 kernel: java(678): floating-point assist fault at ip 00000001028f5241, isr 0000020000000008
Jun 4 14:36:18 bangpcia4 kernel: java(678): floating-point assist fault at ip 00000001028f5251, isr 0000020000000008

Please suggest.
Thanks & Regards.

Re: Redhat ES4 server hangs.

GnanaShekar — Tue, 06 Jun 2006 02:05:25 GMT

Sorry for the typo... It hs 4 Itanium 2 CPU's.

Re: Redhat ES4 server hangs.

Steven E. Protter — Tue, 06 Jun 2006 03:36:36 GMT

Shalom,

Looks like an application fault. Is the system properly patched? Is the application version supported on 64 bit Linux?

SEP

Re: Redhat ES4 server hangs.

GnanaShekar — Tue, 06 Jun 2006 04:13:39 GMT

Hi,

Also I found the following in the logs.

Jun 1 15:16:55 bangpcia4 redhat-support-check: WARNING: Red Hat Enterprise Linux ES release 4 (Nahant Update 2) requires no more than 2 CPUs to run as a supported configuration. (4 detected)

Also in the redhat site I found that the maximum number of CPU's supported by RHEL ES is 2.

So my question is if it supports a maximum of 2 CPU's, what would be the consequence of installing it on 4 CPU's. Will it fail.

Thanks & Regards.

Re: Redhat ES4 server hangs.

GnanaShekar — Tue, 06 Jun 2006 04:13:59 GMT

Hi,

Also I found the following in the logs.

Jun 1 15:16:55 bangpcia4 redhat-support-check: WARNING: Red Hat Enterprise Linux ES release 4 (Nahant Update 2) requires no more than 2 CPUs to run as a supported configuration. (4 detected)

Also in the redhat site I found that the maximum number of CPU's supported by RHEL ES is 2. https://www.redhat.com/rhel/compare/server/

So my question is if it supports a maximum of 2 CPU's, what would be the consequence of installing it on 4 CPU's. Will it fail.

Thanks & Regards.

Re: Redhat ES4 server hangs.

Alan_152 — Wed, 07 Jun 2006 10:02:38 GMT

Is CPU threading turned on or off?

Also, please send us a screenshot of "top" when the system appears to hang.

Re: Redhat ES4 server hangs.

Mike Stroyan — Wed, 07 Jun 2006 16:12:08 GMT

The "Floating-point assist fault" messages are normal for jrockit java. Java code sometimes works with Double.MIN_VALUE, which is a denormalized value that causes a float-point emulation trap and the kernel message that you saw in /var/log/messages.

The 2 CPU limit of ES is a licensing agreement rather than a limit in the kernel. That is noted in
http://www.redhat.com/archives/taroon-list/2004-May/msg00132.html
and
http://www.redhat.com/whitepapers/rhel/ASIlluminata.pdf

You could set up netdump or diskdump to see what is happening the next time the OS crashes. See
http://kbase.redhat.com/faq/FAQ_43_4220.shtm
and
http://people.redhat.com/anderson/crash_whitepaper/