pre-emptively replace hard drives?

[/dev/null]

Hey guys,

I just looked at my vm server (with 11 73G raptors) and there are a bunch with > 50k power on hours, and one with > 60k power on hours.

These are in 1 raid 10 of 10 disks + hot spare.

Egg has 150G raptors for $60 and 300G raptors for $89.

I was thinking of possibly replacing them ahead of time just due to speed/age.

Thoughts?

I DO have nightly backups...(this is a vm server).

 

[wtourist]

Only you can decide cost vs work involved in an HDD failure.
On the other hand Crucial M500 980G SSDs go on sale for about $500.
Or 2 1TB or 750 GB Samsung 840 EVO.

 

[IAPro]

IMO far more important than hours is any reallocated and/or pending sectors.

 

[/dev/null]

Only you can decide cost vs work involved in an HDD failure.
On the other hand Crucial M500 980G SSDs go on sale for about $500.
Or 2 1TB or 750 GB Samsung 840 EVO.

No chance of spending $1k or more on a non-problem

 

[AgentQ]

On the other hand Crucial M500 980G SSDs go on sale for about $500.
Or 2 1TB or 750 GB Samsung 840 EVO.

Agreed. You'd be much better off across the board if you just ditched the mechanical drives completely. For your application, it just doesn't make sense to throw more money at mechanical hard drives right now.

 

[AgentQ]

No chance of spending $1k or more on a non-problem

You currently have ~365GB of usable space. You don't need to spend $1K or more to replicate that with SSDs. Two 500GB Samsung 840s would set you back about $600, and you'd end up with substantially better performance, fewer parts to fail, and you'd be down to managing two drives instead of 11.

With your current plan: Depending on the Raptors and RAID level you go with, you're looking at spending a substantial portion of that anyway on old technology. Not a good investment.

 

[gantux]

No chance of spending $1k or more on a non-problem

So I wouldn't think about spending any money for replacement drives if there is no problem with them. If the SMARTs are good there is no immediate reason to replace them just because of age.

Given the power-on-hours the server and the drives seem to be ~6 years old. How long do you intend to keep this machine running? It is rather elderly hardware and neither new spinning disks nor fancy SSDs are well invested money. I'd only replace a disk when it dies. Eventually you will replace the server with something new anyway, even if you have to keep it running for a couple more years.