On Wed, 29 Jul 1998, Alan Thew wrote:
>After upgrading to NIS+ with a large number of patches to 2.5.1, syslog
>now hangs about once a week. It _seems_ to be related to log rotatation
>but that's the best I have. Everything was fine before the upgrade
>(2.5.1 and NIS).
>
>Has anyone else seen this?
>
I got 2 replies:
From: Karl Vogel <vogelke@c17mis.region2.wpafb.af.mil>
From: Chris Marble <cmarble@orion.ac.hmc.edu>
From: Chris Marble <cmarble@orion.ac.hmc.edu>
Alan Thew wrote:
>
> After upgrading to NIS+ with a large number of patches to 2.5.1, syslog
> now hangs about once a week. It _seems_ to be related to log rotatation
> but that's the best I have. Everything was fine before the upgrade
> (2.5.1 and NIS).
I've found that syslogd dies completely when it receives too many
kill -HUP commands. I have to use syslogd from an unpatched 2.5.1
on all my 2.6 boxes to have one that doesn't die every night when
we rotate all our logs.
-- Chris_Marble@hmc.edu - HMC UNIX & Information Systems Manager My opinions are my own and probably don't represent anything anyway.From: Karl Vogel <vogelke@c17mis.region2.wpafb.af.mil>
>> On Wed, 29 Jul 1998 10:42:34 +0100 (BST), >> Alan Thew <Alan.Thew@liverpool.ac.uk> said:
A> After upgrading to NIS+ with a large number of patches to 2.5.1, syslog A> now hangs about once a week. It _seems_ to be related to log rotatation A> but that's the best I have. Everything was fine before the upgrade A> (2.5.1 and NIS).
Use "truss -p" as root to attach to the running syslog process and see what's happening. If you don't get any output, send something to syslog:
/usr/ucb/logger "wake up"
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:12:45 CDT