Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

System: Centos 5.5 64bit

RAID: Linux Soft RAID with 2 Disks Mirror

Smartd complains about the disk failure in system log:

[root@xxx ~]# tail /var/log/messages
Jul  3 10:22:45 w11 smartd[6043]: Device: /dev/sda, FAILED SMART self-check. BACK UP DATA NOW! 
Jul  3 10:52:45 w11 smartd[6043]: Device: /dev/sda, FAILED SMART self-check. BACK UP DATA NOW! 
Jul  3 11:22:45 w11 smartd[6043]: Device: /dev/sda, FAILED SMART self-check. BACK UP DATA NOW! 
Jul  3 11:52:45 w11 smartd[6043]: Device: /dev/sda, FAILED SMART self-check. BACK UP DATA NOW!

Sorry, but what do I do now?

share|improve this question
6  
I'd start by backing up your data NOW! –  SpacemanSpiff Jul 3 '12 at 5:10
    
Thanks mate. I got daily backup, but good time to test restoring the backup. –  starchx Jul 3 '12 at 12:20

2 Answers 2

up vote 4 down vote accepted

In all seriousness... SMART is a predicated failure, there are tools you can find which will dive a little deeper and let you know what kind of failures it is detecting, there are quite a number of them. If this is a critical system, don't mess around. Start verifying/taking backups and procuring replacement physical disks.

share|improve this answer
4  
While SMART is a predictive failure system, its thresholds for saying that the disk is "failed" are set rather high. It tends to miss failures more than it overpredicts them. In my experience, by the time SMART gets around to saying that the disk is "failed", it's usually almost too late to get the data off. There are other warnings that come sooner that you need to watch for. If there is any data on that system that you want to keep, back up the drive. Also, if you are in a RAID, take backups of the entire thing and then fail out and replace the drive before all your data goes poof! –  Scrivener Jul 3 '12 at 5:49
    
Thanks for the detailed comment. Looks like I need to action ASAP.Thanks! –  starchx Jul 3 '12 at 12:21

It's a RAID 1 mirror, right? Just unplug the bad drive, plug in a new one, and rebuild the RAID. No need to hyperventilate about data loss - as you would if you had no RAID.

Obviously you should be backing up as a matter of course, even a free online backup of just your most important stuff, but that disk could go down in flames and you'd still be OK working off of a single one for a while. It's smart to replace early than take your chances on a double-failure though.

share|improve this answer
    
Thanks mate. Will get a new disk and replace the faulty one. –  starchx Jul 3 '12 at 12:19

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.