Date: Thu, 17 Jun 2010 03:42:28 -0400 (EDT) From: Charles Sprickman <spork@bway.net> To: Pawel Jakub Dawidek <pjd@FreeBSD.org> Cc: freebsd-fs@freebsd.org Subject: Re: zfs panic "evicting znode" Message-ID: <alpine.OSX.2.00.1006170338401.26091@hotlap.local> In-Reply-To: <20100616132504.GK1739@garage.freebsd.pl> References: <alpine.OSX.2.00.1006152225050.26091@hotlap.local> <20100616132504.GK1739@garage.freebsd.pl>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, 16 Jun 2010, Pawel Jakub Dawidek wrote: > On Tue, Jun 15, 2010 at 10:34:20PM -0400, Charles Sprickman wrote: >> Howdy, >> >> I have a box running 8.0-RELEASE that recently started panicing every few >> hours with the following message: >> >> panic: evicting znode 0xa1eafe80 > > Do you have at least backtrace from DDB? Nope. The box had been stable and then paniced at least two times that I'm aware of. On the last one, I drove to the datacenter to see what it dumped on the console. I just built and installed a kernel with ddb/kdb and will reboot next time I'm there. > What kind of workload do you have? The box is essentially a development machine, and after it started acting up, I moved the jails to another host, so for this panic and the one before it, it was essentially idle. > I was no able to find a way to trigger znode_evict_error() call (this is > what you're seeing). You could try changing '#if 1' to '#if 0' in > zfs_znode.c:znode_evict_error(), but I don't think this code path was > ever tested, because as I said, I was unable to trigger it. What is this area of the code doing? Is this part of some type of consistency check that zfs performs? Does it in any way indicate that there's bad metadata that a scrub cannot detect? Thanks, Charles > -- > Pawel Jakub Dawidek http://www.wheelsystems.com > pjd@FreeBSD.org http://www.FreeBSD.org > FreeBSD committer Am I Evil? Yes, I Am! >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?alpine.OSX.2.00.1006170338401.26091>