[NBLUG/talk] Older Server

S. Saunders sms at sonic.net
Fri Jan 18 16:58:20 PST 2008


On Thu, January 17, 2008 11:13, Walter Hansen wrote:
> I have an older server running RH 7.3 ... It has an issue booting
> and I think I know why.  It had a single directory with 15,000+
> files in it under ext2 or ext3 (I think ext3).  I used knoppix and
> copied the important stuff to a usb drive and then used
> e2fsck to repair the file system after deleting the 15,000+ files. There
> were a lot of problems mostly in the upper blocks. I'm running e2fsck -c
> to check for bad blocks right now, but I don't think it's a drive issue.
>
> It's still not booting and I'll post where it grinds down after this, but
> my question is if there is anything I should try to revive it or if I
> should just start over and re-install everything on it? I should have had
> a good complete backup, but this was really a low priority test server
> that we had just gotten to do something useful (no excuse).  Also
> does it sound like my guestimate is correct and the 15,000+ files
> creates problems for ext3?
>
> RH 7.3 (2.4.18-3)
> from screen:
>
> INIT: version 2.84 booting
> INIT: can not execute "etc/rc.d/rc"
> INIT: open(/dev/console): input/output error
> INIT: Id "1" respawning too fast: disabled for 5 minutes
> INIT: Id "2" respawning too fast: disabled for 5 minutes
> INIT: Id "6" respawning too fast: disabled for 5 minutes
> INIT: Id "3" respawning too fast: disabled for 5 minutes
> INIT: Id "4" respawning too fast: disabled for 5 minutes
> INIT: no more processes left in this runlevel
>
> If left going it produces more of the respawning too fast
> messages every so often.

Hrm.  I'll take a stab at this one... while I do know that some FS's
have problems with too many files in a single directory, I can neither
confirm nor deny that ext2 or ext3 are among those.  Have you
Google'd "too many files"+<ext2|ext3> ?

However, the errors above don't seem related to that issue
(unless p'raps your 15Kilofile directory was /etc/rc.d or maybe /etc).
As for the errors above...

Does /etc/rc.d/rc exist?  Is it on the root FS (it appears to be, from
the above)?  Is it executable?

What does "ls -l" say about /etc/console?  Compare it to other /dev's.
Is is possible you need to re-create it (I dunno the RH command(s) to
do so, but presumably someone will be along shortly who either
knows the commands, or knows that something else is appropriate).

As for the respawns... PID's [1,2,3,4,6] are REALLY low-level items;
I wouldn't expect J.Random Directory to force (m)any of them to
respawn, unless some boot-time scripts are forcing I/O to/from that
questionable directory.

At present, I would be *dis*inclined to think the 15Kilofile directory
was the only (or even the primary) problem, though I wouldn't
eliminate it.  Also, it could be just a "contributory" issue on a system
that has OTHER problems...


My .02, FWIW...

- Steve





More information about the talk mailing list