Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 15 Apr 2000 21:20:54 +0800
From:      Adrian Chadd <adrian@freebsd.org>
To:        Matthew West <mwest@uct.ac.za>
Cc:        Rahul Dhesi <dhesi@rahul.net>, Gregory Bond <gnb@itga.com.au>, freebsd-bugs@FreeBSD.ORG
Subject:   Re: kern/17122: crash due to: softdep_disk_write_com
Message-ID:  <20000415212053.A52908@ewok.creative.net.au>
In-Reply-To: <20000414192543.A94912@apotheosis.org.za>; from Matthew West on Fri, Apr 14, 2000 at 07:25:43PM %2B0200
References:  <200004140623.QAA04267@lightning.itga.com.au> <Pine.BSF.4.21.0004132332280.55512-100000@yellow.rahul.net> <20000414192543.A94912@apotheosis.org.za>

next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Apr 14, 2000, Matthew West wrote:
> On Thu, Apr 13, 2000 at 11:38:45PM -0700, Rahul Dhesi wrote:
> > In my case the machine does many things:
> >    squid proxy cache
> >    incoming and outgoing mail
> >    NFS service for exported mail file system
> >    general-purpose shell access for users
> >    recursive name service for PPP users
> >    authoritative name service for one busy DNS zone
> 
> We have several FreeBSD machines running quite smoothly with softupdates
> enabled, under varying loads.  The only time softupdates proved to cause
> instability was on our (heavily loaded) NFS server.  Disabling softupdates
> made the crashes go away.  The problem existed in 3.4-RELEASE and is still
> present in 4.0-RELEASE.
> 

Squid has some interesting UFS access patterns. How loaded is your
squid? Pull up the cachemgr script, and email me the output of
the general runtime statistics. If your squid is cycling through
objects at a high rate it is probably confusing softupdates.



Adrian



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-bugs" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20000415212053.A52908>