Replace Faulty Disk

[perotine]

It did not happened yet. But I like to know how to do it - I did not found any info on Wiki yet.

Let say disk fail (I've disconnected one of 4 drives to try it).]

I restarted server without 1 disc connected and It did not restarted. It gave me message
something like : fsck.ext4 fatal error - cant resolve it, bla bla bla. and disk ID (lots of digit and letters), Ctrl-D to restart or password to fix it.

Any one have any guide how to act in this case? I dont like to stuck on this if one of discs really fails.

Please plain language (for NOOBs) if possible.
Thanks.

[lrevxl]

Disclamer: This question is not a Greyhole question, it's a general Amahi/Linux question.

Here's a quick answer anyway:

This happens because your drive is specified in your system's fstab file. (You can google fstab, but put simply it's a file with a list of your system drives/partitions -- either by name or uuid -- and the locations in your file system they will be mounted to.) If one of the drives listed in the fstab file are missing, the system will not start, it's that simple. Why this is the behavior, I don't know. It'd be much better to ignore non-essential missing drives. (i.e. any drive that's not a system or swap volume) but I'm not a Linux developer, and I don't know the reasoning / rationale behind it.

That said, to recover, you need to use a live cd of some kind, mount your system's file system, and edit your fstab to remove the offending drive. (I've had to do this on several occasions.) This link gives a brief description on how to do it: http://ubuntuforums.org/showthread.php?t=208588 (And FYI, you may be better off commenting out all the storage drives -- and I mean ONLY the storage volumes, it'd be very bad to comment out swap or system partitions -- and reboot until you can isolate which one is actually missing upon getting back into your server.)

[perotine]

Thanks.
As I told - i'm kinda new here - to Amahi and to Fedora, etc.
Was using Bill Gates products for last 20 years.
But thanks for answer - will add it to my bookmarks in case one of drives fail - it happened few times before on other systems and I happy to know how to deal with it now.
Thanks!

[jritchie]

So, for Greyhole, do the drives have to be mounted by fstab or could you use a seperate script to mount the storage volumes at boot time? This just happened to me yesterday. This may be a better way to setup Greyhole as I don't want to have to track down an Ubuntu live CD every time a disk fails.

[jritchie]

I guess I'll be the first to try it then since no one has responded. I will report back my findings.

[lrevxl]

You'd probably have better luck getting an answer if you asked in a new thread. As to your question, yes, it's possible and common. For instance you can put mount statements in your /etc/rc.local file which will mount your drives by UUID and would not prevent your server from booting if missing a drive since it would not be in your fstab. Depending on the order in which things are run, though, this might lead to conflicts with Greyhole. (I.e. if Greyhole starts up before the drives are mounted it will think the drives are missing and begin an fsck which will wipe out the symlinks -- meaning the files won't be accessible again until an additional fsck is done post mount.) So you'd probably need to trigger the Greyhole start from the script as well.