Topic: Good thing it was just a test server

[Michael Smith]

Mitel 5.5, all updates, test box, pressed into service as a dialup router after Hurricane Lili took out a client's cable Internet.  Cable restored, instead of taking the time to halt the machine gracefully I just turned it off.  (Running w/o monitor or keyboard; I would've had to've found a workstation not in use, didn't feel like getting my laptop out of the car, etc. etc. etc. = lazy!)

Turned it on for another task in another place ... oops!  Corrupt filesystem.  "Fsck /" proceeds to completion OK but gosh this thing is acting flaky ... won't boot up all the way, no services .... hmmm better reinstall.

Except it won't.  Reinstall, that is.  To make a rather long story shorter, I ended up having to use the hard drive maker's utility to write zeroes to the drive, destroying ALL data including partition table.  After that, all's well with a reinstall.

The moral to this story:  don't crash the box unless absolutely necessary, but if you do, and don't need the data, wipe that hard drive & start completely over.

[darren]

why didn't you just fix the corrupt files
when it normally crashes like that it allows you to drop in to some sort of command line like a SINGLE boot or somthing
just type  fsck /dev/hda6
Well that's where mine is anyway just replace the hda6 with wherever your drive is.

[Michael Smith]

Actually, Darren ...

a)  I did say I was making a long story short
b)  I fscked just about everything that would stand still for it
c)  Fsck eventually fixed the fileSYSTEMS but it became obvious that there were still corrupt FILES (and lots of them!)
d)  Most importantly ... it's a *test server* ... it exists to be destroyed in interesting ways and remade.

[Lazo]

I have a sme Server 5.0, and there has been five time this machine has turned off (not shutdown), an lucky for me, It just boots fine without any error!! It depends on luck after all!

[Lazo]

I have a sme Server 5.0, and there has been five time this machine has turned off (not shutdown), an lucky for me, It just boots fine without any error!! It depends on luck after all!

[Gerard Roberts]

Hi

I'm new to this list and new to Linux.
Darren's comments helped me  because I also had an SME v 5.5 "Test Server" that was going okay but ran into a hard disk problem which the manufacturer's utility could not fix.

I wanted to have a look at Red Hat 8.0 and tried to replace the SME Server (which had no problems at the time) with Redhat 8.

BTW I am only using gui interface at this stage.

Chose "automatic partitioning" and received following warning:
-The partition on device hda was unreadable. To create new partitions it must be initialised, causing the loss of ALL DATA on this drive.
-Would you like to initialise this drive?
-Yes - NO

I chose YES. Received error message:
-Input/output during read on /dev/hda
-RETRY-IGNORE-CANCEL

-RETRY or IGNORE returned same error message as above.
CANCEL returned the message:
-No Drives found
-An error has occurred - no valid devices were found on which to create new file systems. Please check your hardware for the cause of this problem.

Same results trying "Disk Druid" and "Fdisk."

So I tried to boot up the server and got:
-Pri Master HDD Error
-Run SETUP
-Press F1 to resume

So I press F1.
-Insert BOOT disk in A

I inserted the SME Server 5.5 emergency boot disk:
- Could not find kernel image: linux

Using a Quantum Fireball ST3.2A 3Gig disk. The diagnostic tool is "Powermax" (http://www.maxtor.com/en/support/downloads/powermax.htm).

DIAGNOSIS: " the drive should be replaced."

So my question is: Does the SME Server installation do something to the hard drive that makes it impossible to wipe it and start all over? Or is there a problem with the partitioning tools supplied with the redhat 8?

Gerard Roberts
gr@ffwpu.asn.au

Michael Smith wrote:
>
> Actually, Darren ...
>
> a)  I did say I was making a long story short
> b)  I fscked just about everything that would stand still for
> it
> c)  Fsck eventually fixed the fileSYSTEMS but it became
> obvious that there were still corrupt FILES (and lots of them!)
> d)  Most importantly ... it's a *test server* ... it exists
> to be destroyed in interesting ways and remade.

[Gerard Roberts]

Oops.
I should have said that Michael's experience (not Darren's) helped me.

Gerard Roberts
gr@ffwpu.asn.au