lkml.org 
[lkml]   [1999]   [Oct]   [3]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
Subject2.2.13pre14 file corruption (semi-documented)
This is probably related to two current threads, 2.2.13pre14 crash
(documented) by Barrett G. Lyon and File corruption in 2.2.12* by
Nicholas R LeRoy.

I have an Alpha system (AS200 4/233, 128 MB RAM, TGA card) which has until
recently been running 2.2.9 with no problems. A few days ago I decided to test
2.2.13pre14 and all seemed fine.

Today, after approximately 10 hours of uptime spent in an X session and
building varying amounts of source, I decided to build the latest release of
XFree86 3.9. The entire build went without a hitch, at times I was also using
an X session, Netscape and so on, so the box had a reasonably high load.

After the build completed, I decided to do a 'make install'. After running for
a couple of minutes, I started getting ext2fs errors on /dev/sda6. After a
few screenfulls, these stopped. This coincided with 'make install' producing a
'Virtual Memory Exhausted' error, also stopping.

Experimentally I tried a few commands, some of which (I don't remember the
exact commands) produced lots of shell errors, suggesting that the executable
had been corrupted and was being treated as a shell script, others (mount) ran
fine.

My immediate course of action (not wishing to jeopardise too much data, since
I had last run a backup on this machine a while back), was to press the reset
button. As I write this, e2fsck is busy checking /usr. From what I can see so
far aside from the expected errors after a hard reset (Deleted inode XXX has
no dtime), the _only_ errors occuring (lots of them) are Duplicate/Bad blocks
in inode XXX, in what seems to be spatially co-located blocks. It remains to
be seen what the result of e2fsck will be.

This seems to point to problems in the VFS/buffer cache layer, as the two
threads I mention above also suggest. Any ideas?

Martin

(please Cc: replies to me as I am not subscribed to the list)
--
Martin Lucina http://www.kotelna.sk/mato/ Wellington, New Zealand
I've always been mad I know I've been mad like the most of us are
Pretty hard to explain why you're a madman even if you're not mad

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:54    [W:0.051 / U:2.028 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site