Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 16 Mar 1998 23:09:30 -0800
From:      Manfred Antar <mantar@netcom.com>
To:        Michael Hancock <michaelh@cet.co.jp>, Karl Denninger <karl@mcs.net>
Cc:        current@FreeBSD.ORG, dyson@FreeBSD.ORG
Subject:   Re: WARNING WILL ROBINSON! Risk of severe filesystem damage suspected
Message-ID:  <199803170709.XAA23541@mantar.slip.netcom.com>
In-Reply-To: <Pine.SV4.3.95.980317145358.8814B-100000@parkplace.cet.co.j p>
References:  <19980316234921.07586@mcs.net>

next in thread | previous in thread | raw e-mail | index | archive | help
At 02:55 PM 3/17/98 +0900, Michael Hancock wrote:
>
>Is it possible that the previous kernel quietly trashed your disks and it
>didn't manifest itself until after the last vm fix commit?
>
>Regards,
>
>Mike Hancock
>
>On Mon, 16 Mar 1998, Karl Denninger wrote:
>
>> ************ URGENT **************
>> 
>> I strongly advise that you *DO NOT* use -CURRENT as it sits at the 
>> present time.
>> 
>> The kernel you build may do SEVERE damage to your filesystems.  
>> 
>> I am working on finding the commit which is causing this and will post to 
>> the list as soon as I have identified it.  
>> 
>> This is not a trivial matter.  It is a very good thing indeed that I use a
>> scratch disk for my /usr/obj area, or I would be really rather upset. :-)
>> 
>> The symptoms are random core dumps and other problems - but the underlying
>> reason is that the filesystem has been trashed beyond belief!  
>> 
>> I unmounted the disk in question CLEANLY and found both partially
truncated 
>> and duplicate inodes all over it, along with severe internal file damage.
>> 
>> I suspect the commit today in kern/vfs_cluster.c, as that's the only
>> filesystem or VM related commit that my first-level examination has found.
>> I am backing that out locally right now and will advise if it makes the 
>> problem go away.
>> 
>> Note that kernels built LAST NIGHT (3/15) at about 23:00 *do not* exhibit 
>> the problem.
>> 
>> --
Same thing happened to me. My disk was hosed. Luckily I had just swapped
disks last Thursday. I put the old one back in and 
it runs fine. 
fstab:

/dev/sd0s1a  on /
/dev/sd0s1e on /var
/dev/sd0s1f on /usr
/dev/sd0s1g on /usr/obj =====softupdates enabled.
most of the damage was on /usr . although after a make -k -j4 world last
night lots of / was trashed
/sbin/init among others. I don't know if softupdates had any thing to do
with it though. I disabled it
and rebuilt tools libs sbin bin and a new kernel and the same thing
happened. no NFS just the one SCSI drive.
Manfred
|=======================|
|         mantar@netcom.com     |
|         Ph. (415) 681-6235         |
|=======================| 

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199803170709.XAA23541>