After a power failure or reset...

Rasjid rasjidw at bigpond.com
Tue Oct 2 00:27:04 EST 2001


A couple of weeks ago I had a power failure while I was using my
computer here.  Needless to say it was not a clean unmounting.

Now, as far as I can tell, everything is functioning just fine. 
However, is there stuff I should check?  I know with earlier version of
RedHat (I'm currently running RH7.1) a reboot without a clean shutdown
used to cause all sorts of problems (services would fail to start etc).

All I want is either a HOWTO or similar doc, or to be told that if
everything looks okay, then it probably is.  :-)

On a similar note, for those of you that remember some postings of mine
in August - work is now using E-Smith (or the Mitel SME Server as it now
is called).  Other than some work in getting the APC Smart UPS to work
and sorting out some backup issues, it has all gone resonably smoothly.

Not long after install we upgraded to version 5.  This also upgraded the
quota package to version 3.0, which I believe requires the 2.4 kernel. 
However, rpm seemed to let this happen (I'm getting more tempted to look
at Debian every day).  Not long after trying to get quota version 3.0
working on a 2.2 kernel, the system started to behave very strangly.  It
became impossible to log in either on the console or via ssh, but the
system was still up and samba was still working.

Upon initiating a reboot via the E-Smith web interface, they system hung
when trying to stop the keytable service.  Forced to use the reset key,
LILO now failed to run (it displayed LI and then stopped).  The
emergency floppy was used to get the system up again - and that evening
I reinstalled lilo and returned the system to quota 2.0.  As far as I
can tell, everything is okay now.  I actually have no idea whether quota
was the cause of the problem, but running a version of quota that seems
to want a 2.4 kernel (the most recent rpm insisted on it, just not the
version that shipped with E-Smith V5) on a 2.2 kernel seems like a bad
idea to me.

Anyway, I have already asked `what should I check after a crash' above,
so I don't need to do that again.  However, the server at work is a HP
LH3000 with a NetRaid controler.  It uses the megaraid controler.  We
are using hardware Raid 5.  How do I check that the Raid array is all
still okay, or is that all handled at a hardware level and can I safley
ignore it?  HP has custom kernel modules for the NetRaid controler, but
supposedly E-Smith (Mitel) won't support a modified system (I think
their custom modules allow the use of their Raid tools).  The whole Raid
thing is way outside my realm of experience.  I'd just like to know how
to check to see that it is all okay, perferably using the default
megaraid kernel module, although I'm willing to consider installing the
HP module if it is the only way.

Thanks in advance for any feedback,

Rasjid.




More information about the linux mailing list