HDD Lifespan - RAID5/6

[rashdanml]

I've been debating a cloud server setup, possibly with multiple 3TB HDDs in a RAID 5 or 6 setup. Discovered FlexRAID through some searching, seems like a good option.

 

However, just out of curiosity; I know a few people here have run a bunch of RAID setups.

 

1) How long have you been running the RAID setup for? Hardware or Software RAID?

2) Has any of the drives failed in that time, and if so, after how long?

 

I haven't really found any ballpark figure (maybe it's one of those things where each individual has vastly different experiences due to a lot of factors, like frequent writes, HDD manufacturer, etc), but I've been curious to know a rough estimate anyway. Considering the frequent shuffling of data (depending on how frequent), I'd expect HDDs to not last very long, but would be nice to put a rough (however unrealistic) number on it.

 

Cheers.

 

[h264]

Raid 6 of 3tb Seagate drives, 12 total

1 drive failed in the first 24 hours

None since. I have two hot spares though.

[h264]

Raid 6 of 3tb Seagate drives, 12 total since about a year ago.

1 drive failed in the first 24 hours

None since. I have two hot spares though.

[Andi]

you can't put a number on it, any drive can die any time :D I have a 40GB drive, over 10 years old and runs still fine. I lost other drives after 6 month.

 

If you want to build a raid setup it's important that you don't buy all drives at once from the same store. if there is a manufacturing failure/rough shipping you could lose all drive close together.

[rashdanml]

If you want to build a raid setup it's important that you don't buy all drives at once from the same store. if there is be a manufacturing failure/rough shipping you could lose all drive close together.

 Interesting point. Thanks for that. I'll bear that in mind.

 

Still curious though. I read a topic on stress-testing HDDs to get it past the initial infant mortality failure period; that probably explains most short-term failures.

[Vitalius]

 Interesting point. Thanks for that. I'll bear that in mind.

 

Still curious though. I read a topic on stress-testing HDDs to get it past the initial infant mortality failure period; that probably explains most short-term failures.

If you have a brick and mortar store you can buy large amounts of HDD's from, I would stress test one drive for a while (1 week tops) and if it's good, buy more of that same drive model at the same store. That way you can be fairly certain you get ones of the same quality and condition.

That might sound unrealistic, but I live near a Micro Center and I can do that if I wish.

Then again, RMAing isn't so bad these days. 

[rashdanml]

I live near 3-4 NCIX stores in Vancouver; think they opened up a large retail store recently too. They usually have plenty of drives in stock.

[xXxYOLOxSWAGxXx_420BlazeIt]

Look at the MTBF of the drives you are buying.

[adhisakti]

My old server runs for about 3 years with 2xSeagate Barracuda 160GB now i'm replace it with new server that has been running about 1 year with 2xSeagate Barracuda 2TB plus 1xSeagate Barracuda 40GB for boot drive, all with Software RAID 1 and nothing problem until now.