Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 22 Dec 2009 14:49:53 +0100
From:      Florian Smeets <flo@smeets.im>
To:        Pete French <petefrench@ticketswitch.com>
Cc:        freebsd-stable@FreeBSD.org
Subject:   Re: Disc lock up on 8.0-STABLE
Message-ID:  <4B30CE81.7030303@smeets.im>
In-Reply-To: <E1NN3fP-00048P-Tb@dilbert.ticketswitch.com>
References:  <E1NN3fP-00048P-Tb@dilbert.ticketswitch.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On 12/22/09 1:16 PM, Pete French wrote:
> I've been gradually testing 8.0 on several machines propr to deploying it
> live, but I currently have a machine which appears to lock-up at 3am
> every day. The symptoms are that the machine is still pingable, but doing
> anything which requires access to the disc just freezes (so you cant login for
> example). I've seen simiilar behaviour behore on machines when the disc
> syste has locked up for some reason, so am ttentatively guessing that
> this is the cause.
> 
> The machine is an HP DL360 G5 with a ciss0 controller for the drives.
> I have upgraded to the latest STABLE but the freeze still happens.
> Am including a dmesg below, and will compile it with KDB, DDB to
> see what happens.
> 
> The machine is booting from a UFS partition, but is using ZFS for everything
> else. The fcat it deadlocks at 3am makes me thing this is something to
> do with scheduled jobs maybe ? Then again, I have an almost identical DL360
> which is running 8.0 and is rock solid.
> 

Hi Pete,

i'm trying to track down the same problem. The box in question has
everything on UFS (mirrored ataraid) and a backup disk with ZFS on it
attached to USB.

The freeze happens at 3am too, i have these log messages:
Dec 20 03:00:00 XXX newsyslog[2810]: logfile turned over due to size>100K
Dec 20 03:03:21 XXX kernel: Approaching the limit on PV entries,
consider increasing either the vm.pmap.shpgperproc or the
vm.pmap.pv_entry_max sysctl.
Dec 20 03:03:22 XXX kernel: maxproc limit exceeded by uid 0, please see
tuning(7) and login.conf(5).
Dec 20 03:03:53 XXX last message repeated 31 times

I had increased vm.pmap.pv_entry_max and vm.pmap.shpgperproc then only
the maxproc limit exceeded message remained.

As the box is remote with only ssh access, it's a little difficult to
debug this. During the weekend i waited till 3 o'clock with a top
running, and saw that hundreds/thousands of /bin/sh processes were
started. After that i commented out periodic daily in /etc/crontab, that
"solved" the problem for me.

I was not able to debug this any further yet, i have one other box with
all UFS and a ZFS backup disc also running latest 8-STABLE but it does
not exhibit the problem.

Cheers,
Florian



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4B30CE81.7030303>