Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 26 Oct 2005 15:56:42 +0200
From:      Jeremie Le Hen <jeremie@le-hen.org>
To:        current@freebsd.org
Subject:   Re: panic: ffs_valloc: dup alloc
Message-ID:  <20051026135641.GV14063@obiwan.tataz.chchile.org>
In-Reply-To: <20051021163151.GG985@galgenberg.net>
References:  <20051021163151.GG985@galgenberg.net>

next in thread | previous in thread | raw e-mail | index | archive | help
Hi Ulrich,

> updated my server to 6.0RC1 (needed GELI, so I went straight to 6.0) and
> got this panic while moving 20GB of data to a GELI+GMIRROR backed
> volume.
> 
> mode = 015115, inum = 444208, fs = /vol2
> panic: ffs_valloc: dup alloc
> Uptime: 7h55m14s

It would be far more useful to provide a backtrace.  You can achieve
this by asking FreeBSD to frop to DDB upon panic and then use the
"trace" command.

> No WITNESS, no INVARIANTS, no DDB. Sorry. Should I retry with these
> settings?

Yes, INVARIANTS are useful to have, and I think DDB is a must :-).

> Due to performance reasons, I went with a geli-sectorsize of 8192 for
> this particular volume, resulting in a bs/fs of 16384/8192.
> 
> Is this an incredibly stupid thing to do? How else could I speed up the
> operation of GELI? (No, sadly GBDE is even slower).

I don't know, but the kernel should not panic anyway.

Regards,
-- 
Jeremie Le Hen
< jeremie at le-hen dot org >< ttz at chchile dot org >



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20051026135641.GV14063>