[Date Prev][Date Next] [Chronological] [Thread] [Top]

RE: Solaris instability: maybe the monitor backend?

Might try ElectricFence or a similar malloc-consistency checker to see
if back-monitor is tripping anything up. Is this a hang or a crash?
A hang would imply a mutex problem, not a malloc problem.

  -- Howard Chu
  Chief Architect, Symas Corp.       Director, Highland Sun
  http://www.symas.com               http://highlandsun.com/hyc
  Symas: Premier OpenSource Development and Support

> -----Original Message-----
> From: owner-openldap-software@OpenLDAP.org
> [mailto:owner-openldap-software@OpenLDAP.org]On Behalf Of Alan Sparks
> Sent: Friday, August 09, 2002 1:18 PM
> To: openldap-software@OpenLDAP.org
> Subject: Solaris instability: maybe the monitor backend?
> More about my instability problems with OpenLDAP 2.1.x and Solaris7.
> I've since patched the servers with Sun's latest recommended patch set.
> The servers have run without fail since Tuesday.  My test scripts which
> update a value and verify the replication have run since then without
> fail...
> Til today.
> The coincidence: 'til today, I have not run my MRTG probing script.  This
> script fetches a few values from the monitor backend (current/start time,
> connections, total ops) to feed into MRTG.  But, within a few hours of
> starting that again (running every 10 minutes), the replication failed.
> I've shut that back down, to see if the problem persists.  Things are back
> to normal for the moment.  While this is coincidental, and I have not a
> bit of proof, I wonder if there's some possible odd locking problem or
> something in the monitor backend that might be hanging the server.
> Will keep watching it.  Will try to find out more.
> -Alan
> ===========
> Alan Sparks, UNIX/Linux Systems Administrator    <asparks@doublesparks.net>