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.

I have a 5 disk Raid5 array of Western Digital Green (WD10EADS) 1TB. Raid5 is via Windows Server (software raid).

One disk has failed, and it looks as though I can't get the same model replacement. I've also read that the newer (consumer level) WD alternatives will misbehave in a Raid array (TLER issues etc).

The array has been powered up for 29,000 hours (3.2 years). Given that these are consumer drives, should I try to replace the dead disk or assume that the rest are going to die soon and replace the whole lot?

Edit: Just realised that rep doesn't transfer from stackoverflow, so I can't upvote just yet... sorry!

share|improve this question
2  
You might want to check the serial numbers and manufacture date of the remaining drives. Drives from the same batch tend to like failing with one another... –  Charles Aug 19 '12 at 16:55

2 Answers 2

First of all, make sure you have good backups (both in general and before doing any big changes).

Secondly. If you have enough (coorporate) money consider HW RAID, support contracts and SAS drives. After all if your RAID fails and hundreds of employees can not access their data things get expensive.

(If this is a home situation: Try moving this post to the superuser forum). I suspect that their answer will be to just replace the drive (possible with a WD red drive or equivalant) and to add some monitoring software.

Lastly, check this question about RAID levels. Specifically on the advantages and disadvantages of RAID5. RAID 5 is not bad and often quite useful, but make sure you know about the pitfalls. Especially if you replace the drives.

share|improve this answer
    
It's a SOHO setup running 4 dev servers and 3 workstations. Everything is backed up and verified nightly. I'm considering replacing with RAID6 + WD RE disks. The WD Reds look interesting, but they're new - I can't find any reliability data. Monitoring is a really important point... I'm sad to say this was a silent failure. I've already coded an SMS notification service to ensure this doesn't happen again! –  Satellite Aug 19 '12 at 17:05

3.2 years is quite long for something running 24x7 and consumer grade. Be prepared for the rest of them to fail.

Do replace your drives with enterprise class drives (with 5 year warranty) since the new consumer drives have reduced warranty as less as 1-2 years and some for 3 years-- http://www.techspot.com/news/46726-seagate-and-western-digital-announce-reduced-warranties-for-hard-drives.html

Also, have a hard drive or 2 as hot spares so that you don't risk losing a second hard drive while you wait to replace your failed drive. RAID 6 can 'tolerate' 2 HDD failures and this is you can consider if your controller supports it.

As always, having RAID doesn't mean guaranteed availability of data and hardware can fail in many ways. Consider having a backup on another server/NAS as a second set of backup.

share|improve this answer
    
All good points, thanks. I'm seriously considering the RE drives and Raid6. I like that software raid is controller agnostic, and that there are great low-level recovery tools available. But given that Windows Software Raid only does level 5, I'll have to compromise if I want the safety of Raid6. –  Satellite Aug 19 '12 at 17:10
    
Be prepared on how to recover manually from disk failure from failed RAID set. It can be very tricky sometimes. –  Chida Aug 19 '12 at 17:21

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.