xfs
[Top] [All Lists]

Re: xfs bug ?

To: Simon Matter <simon.matter@xxxxxxxxxxxxxxxx>
Subject: Re: xfs bug ?
From: Steve Lord <lord@xxxxxxx>
Date: Mon, 23 Jul 2001 10:37:10 -0500
Cc: matt@xxxxxx, linux-xfs@xxxxxxxxxxx
In-reply-to: Message from Simon Matter <simon.matter@xxxxxxxxxxxxxxxx> of "Mon, 23 Jul 2001 17:29:34 +0200." <3B5C42DE.607AF347@xxxxxxxxxxxxxxxx>
Sender: owner-linux-xfs@xxxxxxxxxxx
> "Mathew A. Chrystal" schrieb:
> > 
> > Hi:
> > 
> > I have been installing redhat 7.1 with the sgi-installer iso. After
> > installation is complete and the machine boots for the first time
> > everything is fine. When you reboot the second time thes messages start
> > appearing:
> > ...
> > checking hardware
> > touch creating '/var/lock/subsys/kudzu': unknown error 990
> > ...
> > ..
> > bringing up interface eth0
> > touch creating '/var/lock/subsys/network': unknown error 990
> > starting system logger..
> > there seems to be something wrong with the /var/lock/subsys directory.
> > 
> > if you go into the subsys directory and try to make the files you get
> > the same error 990 message.
> > 
> > i will attach the dmesg and output from an ls on the subsys directory.
> > 
> > ps. I am using a kickstart install and /var has its own partition.
> > also if i umount the /var partition and run xfs_repair it fixes the
> > problem.
> > 
> > Matt Chrystal
> > University of Notre Dame
> > 
> >   ------------------------------------------------------------------------
> > 
> >    probName: prob
> >        Type: Plain Text (text/plain)
> > 
> >    mprobName: mprob
> >         Type: Plain Text (text/plain)
> 
> Hi,
> I was having filesystem corruption on DELL Precision WS when using
> more than one disk, no matter whether in RAID configuration or
> just mounting different partitions on both disks.
> Can you try two things:
> 
> 1) What does rpm -Va say. Do you see lot of corrupt files?
> 2) Can you remove the second drive and install on just the
> first drive?
> 
> -Simon
> 

I should add that error 990 is EFSCORRUPTED which usually means XFS has
detected a metadata problem on the disk and has shut it down. There should
have been a console message when this started, I did not see one in the
output Mathew sent.

The xfs_create looping error should not be happening either, this is another
inconsistency between metadata in different locations - a lookup on a
directory is failing to find an entry in a directory, but the create code
says it is there.

At this point I would distrust the hardware, Simon's suggestion would be
a good one.

Steve




<Prev in Thread] Current Thread [Next in Thread>