How do I know if a hard drive fails in RAID 5?

[Mar]

I got three 1 tebibyte drives with a hardware RAID controller set up on RAID 5 in a little Acer 100 home server unit I got from Memory Express, special order. I put Ubuntu 13.10 server on it and regularly SSH into it to run various tasks but how do I know if one of the hard drives fails? Will it alert me over SSH somehow when I log in? Will it just not boot after an upgrade? I'd like to know prior to trying to reboot it after any security patches since it's off site.

[firebane]

Being in that RAID 5 allows for 1 drive failure to ensure data is not compromised if you do indeed have a failed drive the software you are using or something in the OS should warn you of a drive failure.

In Windows you'd have Intel RST that would let you know of a failure of a drive allowing you to remove it and re-install a new one and have it rebuild the array onto the drive again.

You would have to research how Ubuntu would notify you of a failed drive.

[Zhariak]

You mention that you're using a hardware RAID controller...

This means it's transparent to the operating system, unless of course you actually identify the hardware RAID controller and the install the controllers management software.

Without the RAID management software, the system just thinks of hardware controllers RAID volumes as a normal disk volume... This could be dangerous in case you have a failure.

If you install it, you'll more than likely be able to setup alerts to let you know of failures.

[Alterac]

^ What he said :D


Depending on the controller there is software you can install to monitor the raid array.

If it was MDADM (Software Linux Raid) you would set that up to alert you.


you can also setup the Smartmontools daemon to monitor the drives smart status and alert you there.

[toor]

Originally posted by Alterac
Depending on the controller there is software you can install to monitor the raid array.

If it was MDADM (Software Linux Raid) you would set that up to alert you.

you can also setup the Smartmontools daemon to monitor the drives smart status and alert you there.

^ what they all said :P

...then configure it to email you and make damn sure the mailer works!

Another example, if you have a 3ware card you'd want 3dm2 web interface and/or tw_cli.

I've never used smartmontools to monitor overall 'array' status (not sure it's designed for that) but getting alerts for individual drives (assuming your controller is supported/passthru) that were deteriorating has saved my ass a few times.

Next time, just get an HBA and go software RAID. So much easier, cheaper and more flexible...