Re: Filesystem completely hosed
Suspicious filesystem problems have come up more than once from
different people. I don't have a test case to validate and prove any
problem, but I don't think that means that a problem doesn't exist. I
certainly don't trust my Hurd system to keep it's own filesystems
"whole" at this point. I have personal anecdotal experience and I've
listened to some others who have mentioned it this same suspicion.
Until we have live machines used for production server use or full-time
developer machines used regularly (with high disk usage), the community
may be unable to definitively track down these kinds of sporadic
problems.
--
-- Grant Bowman <grantbow@grantbow.com>
* Johannes Rohr <j.rohr@gmx.de> [030307 10:49]:
> Hello,
>
> some time ago I reported frequent file system corruptions after
> extensive write operations (installing large packages).
>
> I thought that something might be wrong with the file system, so I
> re-formatted it and then restored my GNU/Hurd installation (running
> native-install etc.)
>
> Today I tried out if anything had changed - I tried reinstalling some
> large packages (which had been damaged by previous crashes): xfree-*,
> tetex-bin, tetex-doc, tetex-base.
>
> The result was truely disastrous. At some point, while initex was
> running the machine froze. (No error message anywhere, probably because
> I was using the new virtual consoles while the errors get displayed to
> the old mach console?) Immediately before the machine froze, I
> observered that ext2fs.static consumed 34% CPU time. Is this normal?
>
> I could only pull the plug (no second computer available to telnet to my
> box.). On reboot (to GNU/Hurd), fsck complained about an unexpected
> inconsistency, restarted several times from the beginning and then the
> machine suddenly rebooted. On trying to boot GNU/Hurd, grub informed me
> that there was no such thing as /boot/gnumach.gz. So I booted GNU/Linux
> and fscked again. I have a typescript of this available but even in
> bzipped format it is still 49K large (uncompressed 400K).
>
> Well, in the end, my / directory was _empty_, except for lost+found,
> from where I have recovered some of the installation. (I don't have
> GNU/Hurd CDs here.)
>
> Here's the output of tune2fs -l /dev/hdc7
>
> tune2fs 1.32 (09-Nov-2002)
> Filesystem volume name: <none>
> Last mounted on: <not available>
> Filesystem UUID: 2ce47e2f-64b2-4c0c-b3c1-c14714edd6ad
> Filesystem magic number: 0xEF53
> Filesystem revision #: 1 (dynamic)
> Filesystem features: sparse_super
> Default mount options: (none)
> Filesystem state: not clean
> Errors behavior: Continue
> Filesystem OS type: GNU/Hurd
> Inode count: 120576
> Block count: 240967
> Reserved block count: 12048
> Free blocks: 61975
> Free inodes: 81958
> First block: 0
> Block size: 4096
> Fragment size: 4096
> Blocks per group: 32768
> Fragments per group: 32768
> Inodes per group: 15072
> Inode blocks per group: 471
> Filesystem created: Fri Feb 28 15:20:19 2003
> Last mount time: Fri Mar 7 17:58:40 2003
> Last write time: Fri Mar 7 18:22:29 2003
> Mount count: 1
> Maximum mount count: 31
> Last checked: Fri Mar 7 17:58:30 2003
> Check interval: 15552000 (6 months)
> Next check after: Wed Sep 3 18:58:30 2003
> Reserved blocks uid: 0 (user root)
> Reserved blocks gid: 0 (group root)
> First inode: 11
> Inode size: 128
>
> Any hints someone?
>
> Thanks,
>
> Johannes
>
>
> --
> To UNSUBSCRIBE, email to debian-hurd-request@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
>
--
-- Grant Bowman <grantbow@grantbow.com>
Reply to: