Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 26 Oct 2005 13:35:50 -0700
From:      Frank Mayhar <frank@exit.com>
To:        current@freebsd.org
Subject:   Re: "panic: initiate_write_inodeblock_ufs2: already started" on 6.0-RC1 with Intel SRCU42L RAID.
Message-ID:  <1130358950.47465.11.camel@realtime.exit.com>
In-Reply-To: <1130048351.3442.8.camel@realtime.exit.com>
References:  <1130048351.3442.8.camel@realtime.exit.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, 2005-10-22 at 23:19 -0700, Frank Mayhar wrote:
> I ran into this panic this evening; PR entered as kern/87861.  The
> filesystem that gets this is on an Intel SRCU42L RAID5 array and that
> seems to be the important characteristic.  This also happens in
> 5.4-stable, so it's not something special about 6.0.  I can reproduce
> this at will so it will be easy for me to help diagnose it.

Following up my own email to add a little bit of information.  I built
the kernel with INVARIANTS and managed to catch a KASSERT() at the
beginning of bundirty():
        panic: bundirty: buffer 0xd6d6cc00 still on queue 1

Queue 1 is the QUEUE_CLEAN queue, for (from the commentary)
"non-B_DELWRI buffers."

Somewhat more interesting is this error message from just before the
panic.  There are 200+ of these messages, but this one appears to be
associated with this buffer, since the offset and length match:

g_vfs_done():da6s1e[WRITE(offset=63781224448, length=16384)]error = 16

Error 16 is EBUSY.

Unfortunately the stack is messed up (apparently because of a panic
cascade) so I can't extract a backtrace for this one.

A dump of the buf structure is in the PR here:
	http://www.freebsd.org/cgi/query-pr.cgi?pr=87861
-- 
Frank Mayhar frank@exit.com     http://www.exit.com/
Exit Consulting                 http://www.gpsclock.com/
                                http://www.exit.com/blog/frank/



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1130358950.47465.11.camel>