Topic: Raid corruption, and a total mess, need urgent help

[Drifting]

Hello

Please bear with me I have been dumped with solving this and at this moment in time I have flu, and am not exactly coherent.

Problem, two disk software raid, corrupt install of SME, whole of var directory gone (no idea why) Sme will not fire up properly so cannot take a backup. The server had another disk added that was used for the occasional backup. I have run up Knoppix and can see that the backup disk is fine and I can read the data on it. However I cannot see the raid array at all.

So at this stage I fire up the SME 7.4 CD and do a rescue, it happily sees the raid but does not mount the backup disk and complains it has an invalid partition ?

Hopefully you can see my problem, I need to get the data off the raid set to enable me to do a clean install (Tried and upgrade, no go) My ideal would be to use the backup drive, but have no idea why the SME rescue says it has partition problems? and does not mount.

Any help would really be appreciated, bad enough that I feel like death, supposed to be going on holiday on tuesday!

Paul.

[CharlieBrady]

So at this stage I fire up the SME 7.4 CD and do a rescue, it happily sees the raid but does not mount the backup disk and complains it has an invalid partition ?

When exactly does it say that it has an invalid partition? What are you doing, and what exactly do you see?

The subject says "Raid corruption" but you haven't presented evidence of raid corruption. You've only said that /var is missing.

What is the state of the RAID pair?

If they are in sync, you can shut down, remove one of them, and boot.

[Drifting]

Thanks for the reply Charley

The corruption was due to a faulty UPS, lost a number of files and directories, however the raid pair are happy. The disk with the partition problem is the spare disk in the server we used for taking the odd Rsync when doing anything dangerous. That is the one I would like to copy the remaining data to from the raid pair.

Paul.

[Drifting]

Resolved, managed to restore missing files from earlier Rsync

However I have another problem, this is related to the e-smith directory which was manually restored (copied) back on to the newly rebuilt server. This was a desperate act as it was early morning, and system needed to be got back up.
Problem is that for some  reason we are now not able to access the external web interface, even though it has been allowed, I assume this is due to Iptables blocking it ? or some configuration files being corrupt? Is there a way to reset this ?  And yes I know it is dangerous to use that, but handy when someone else remotely can give me a hand creating users.

Paul.

[cactus]

However I have another problem, this is related to the e-smith directory which was manually restored (copied) back on to the newly rebuilt server. This was a desperate act as it was early morning, and system needed to be got back up.

I think it might be permission errors. Please try to reset file and directory permissions using the following:

Code: [Select]

rpm --setugids $(rpm -qa)
rpm --setperms $(rpm -qa)
signal-event post-upgrade ; signal-event reboot
And let it be a lesson that even when you are under pressure you need to take time to reflect on what you are doing, just copying the e-smith directory is one of the stupidest things (sorry to say so) you could have done as this holds the core of the SME Server system.