Re: Ancient kernel TLB IPI WAIT

From: Rick Stevens (rstevens@publichost.com)
Date: Wed May 24 2000 - 11:59:00 EST


Alan Cox wrote:
>
> > The last messages in the syslog are "stuck on TLB IPI WAIT (CPU#0),"
> > although sometimes it's CPU#1 that complains.
>
> It means one CPU got bored of waiting for the other. If it never
> recovers its normally a deadlock - could be code or hardware. Given you are
> running 2.2.5 I'd put my money firmly on it being kernel problems.

Thank you. I suspected that. It's an LDAP server that handles all
of our virtual email account usernames and passwords. Recently,
some of our customers have been sending large mailings to their
clients. Our mail filters require that either the sender or receiver
must be in our LDAP database or the mail doesn't leave. So the
system is getting whacked much more than it used to.

It could also be temperature-related...the box is a 2U rackmount in
a 50U rack with about 20 other servers (mail servers, Postgres,
etc.). It gets warm in there. My first shot, though, is a kernel
upgrade. I run 2.2.14-12 and later on several other boxes without
problems, including another LDAP server. That box, however, is more
of an experimental hamster and doesn't get hit as often.

Thanks again for the confirmation of my suspicions.

-- 
----------------------------------------------------------------------
- Rick Stevens, CTO, PublicHost, Inc.        rstevens@publichost.com -
- 949-743-2010 (Voice)                     http://www.publichost.com -
-                                                                    -
-                 Dyslexics of the world: UNTIE!                     -
----------------------------------------------------------------------

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Wed May 31 2000 - 21:00:11 EST