On Monday 16 December 2002 23:44, zentara wrote:
On Mon, 16 Dec 2002 16:36:46 +0100
Cees van de Griend <cees-list@griend.xs4all.nl> wrote:
The crash happened during a power failure but I'd like a software answer that we can avoid problems in future.
Uhh. Probably software isn't the anser for an hardware problem. Think about UPS-es. A journalling filesystem is a usefull tool, not a garanty, during a power failure.
It isn't normal to get that kind of filesystem corruption just because the power went out.
I can't comment, I don't have enough data. For example what was de server doing just before the crash, which processes ran, which processes where writing to the disk, etc...
I've pulled the plug on my machine, and I may lose recent information, but my filesystem is still there.
I can't comment, I don't have enough data. I can say that on my home machine I've done the same, but for a server the load is very light and there is normally little writing to disks.
Do you think it's possible for reiserfs to go down that badly? Maybe his hard drive or controller is getting flaky?
I think you have the wrong idea about journaling filesystems. Normally you are right and normally a journaling filesystem should recover quickly after an crash. But there are no garanties. Journaling is for a quick recover after an crash, so you have a 'more or less' working system. (I don't know if you have ever waited for a fsck on an 60 GB ext2 file system - I have and it's a very long time - but if you have you will apriciated journaling filesystems.) Of course, a journaling filesystem tries to get the old data back and normaly it does a very good job, but there are senarios in which this doesn't work. Regards, Cees.