Hello, A couple of times lately I have had files clobbered on me, on reiser fs's, with a most mysterious error. After editing and saving lilo.conf on one occasion, I ran lilo -v, which gave some very peculiar error. On vi'ing the file, I got an error, which sadly I seem to have not kept (I know, not much use then) but it was something about a "pre-fetch" failure. And lo, lilo.conf did not exist anymore. This has happened on 2 subsequent occasions, on two different filesystems. I ran a 6 hour disk exerciser test overnight on the disk in question and not a single error. Fortunately my backup/restore procedures work :) Any clues anyone ? I have been using reiser for a while without problem. This is mantel 2.4.12 kernel. Strangely the errors did not get logged anywhere, not in messages, or warn or.. -- Regards Cliff
On Mon, 5 Nov 2001, Cliff Sarginson wrote:
A couple of times lately I have had files clobbered on me, on reiser fs's, with a most mysterious error.
After editing and saving lilo.conf on one occasion, I ran lilo -v, which gave some very peculiar error. On vi'ing the file, I got an error, which sadly I seem to have not kept (I know, not much use then) but it was something about a "pre-fetch" failure. And lo, lilo.conf did not exist anymore.
This has happened on 2 subsequent occasions, on two different filesystems.
I ran a 6 hour disk exerciser test overnight on the disk in question and not a single error.
Fortunately my backup/restore procedures work :)
Any clues anyone ?
I have been using reiser for a while without problem. This is mantel 2.4.12 kernel.
Please note, that these kernels are experimental/under development - please be cautious when running them on productive systems! Chances are high, that there is a problem with file system corruption here - IIRC, we already received some reports about this.
Strangely the errors did not get logged anywhere, not in messages, or warn or..
I would advise you to stay with the updated 2.4.10 for the time being - or do you need a special feature that is only available in later kernels? Bye, LenZ -- ------------------------------------------------------------------ Lenz Grimmer SuSE GmbH mailto:grimmer@suse.de Schanzaeckerstr. 10 http://www.suse.de/~grimmer/ 90443 Nuernberg, Germany You tell 'em Skyscraper, you have more than one story.
On Mon, Nov 05, 2001 at 03:08:55PM +0100, Lenz Grimmer wrote:
On Mon, 5 Nov 2001, Cliff Sarginson wrote:
A couple of times lately I have had files clobbered on me, on reiser fs's, with a most mysterious error.
After editing and saving lilo.conf on one occasion, I ran lilo -v, which gave some very peculiar error. On vi'ing the file, I got an error, which sadly I seem to have not kept (I know, not much use then) but it was something about a "pre-fetch" failure. And lo, lilo.conf did not exist anymore.
This has happened on 2 subsequent occasions, on two different filesystems.
I ran a 6 hour disk exerciser test overnight on the disk in question and not a single error.
Fortunately my backup/restore procedures work :)
Any clues anyone ?
I have been using reiser for a while without problem. This is mantel 2.4.12 kernel.
Please note, that these kernels are experimental/under development - please be cautious when running them on productive systems!
I realise that Lenz, I would have thought that the resierfs portion is now pretty stable however.
Chances are high, that there is a problem with file system corruption here - IIRC, we already received some reports about this.
Strangely the errors did not get logged anywhere, not in messages, or warn or..
I would advise you to stay with the updated 2.4.10 for the time being - or do you need a special feature that is only available in later kernels?
There is a bit of a conundrum here. Of course SuSE is quite right with this advice, the conservative approach here is clearly the safest one. On the other hand the Linux kernel (let us say the vanilla one) is in a constant state of update, and yes there are reasons I track it. I understand the risk, the fiasco of 2.4.11 should warn us all. I am reporting the risk to the list, which I think is the correct thing to do. And if it gets from you to the SuSE kernel developers as a reported problem, then so much the better. Thanks for your reply, as always Lenz, your presence on this list is appreciated. I have kept track of SuSE 2.4.10 as well, so I do have a fallback (my network is not yet publicly accessible)
Bye, LenZ -- ------------------------------------------------------------------ Lenz Grimmer SuSE GmbH mailto:grimmer@suse.de Schanzaeckerstr. 10 http://www.suse.de/~grimmer/ 90443 Nuernberg, Germany You tell 'em Skyscraper, you have more than one story.
-- To unsubscribe send e-mail to suse-linux-e-unsubscribe@suse.com For additional commands send e-mail to suse-linux-e-help@suse.com Also check the FAQ at http://www.suse.com/support/faq and the archives at http://lists.suse.com
-- Regards Cliff
participants (2)
-
Cliff Sarginson
-
Lenz Grimmer