Dell PERC Question (Server Down)

BRRABill

Limited, sure. But limited beyond what you really need to do?

It's hard to say at this point.

Most of what I have had to do is deal with foreign config on the array/disk, but ONLY after a cold boot. So, it's never been something I could look into.

scottalanmiller

@BRRABill said in Dell PERC Question (Server Down):

@scottalanmiller

Got it.

So we're really just trying to avoid rebooting.

Exactly, because if the server is still functional in any way (rebuilding) then rebooting puts the data at risk.

Dashrender

Once you move to booting not off the array, you should have to cold boot much amnymore, unless the drives require it to come back online

Dashrender

@scottalanmiller said in Dell PERC Question (Server Down):

@BRRABill said in Dell PERC Question (Server Down):

@scottalanmiller

Got it.

So we're really just trying to avoid rebooting.

Exactly, because if the server is still functional in any way (rebuilding) then rebooting puts the data at risk.

Thanks, so I wasn't crazy for saying that last night.

scottalanmiller

@Dashrender said in Dell PERC Question (Server Down):

@scottalanmiller said in Dell PERC Question (Server Down):

@BRRABill said in Dell PERC Question (Server Down):

@scottalanmiller

Got it.

So we're really just trying to avoid rebooting.

Exactly, because if the server is still functional in any way (rebuilding) then rebooting puts the data at risk.

Thanks, so I wasn't crazy for saying that last night.

You aren't crazy for saying that.

BRRABill

@Dashrender said

Thanks, so I wasn't crazy for saying that last night.

Did I call you crazy? If so, I apologize. I know I was GOING crazy.

Not really dealing with this stuff a lot, I was trusting DELL tech support to know what they were doing.

Their take is to let is build to like 5%, and then it is perfectly safe to reboot.

Dashrender

@scottalanmiller said in Dell PERC Question (Server Down):

@Dashrender said in Dell PERC Question (Server Down):

@scottalanmiller said in Dell PERC Question (Server Down):

@BRRABill said in Dell PERC Question (Server Down):

@scottalanmiller

Got it.

So we're really just trying to avoid rebooting.

Exactly, because if the server is still functional in any way (rebuilding) then rebooting puts the data at risk.

Thanks, so I wasn't crazy for saying that last night.

You aren't crazy for saying that.

Never said I wasn't crazy in general.

BRRABill

I mean, in theory, the array has instructions for rebuilding, right?

It has a plan of what to d, block by block. It just pauses on reboot.

Now, if you unplugged it, yeah I could see an issue there.

But isn't it possible it is smart enough to just pick back up?

Dashrender

Sure it is, but it is still a risk. And I've luckily never had to reboot while replacing a drive, different problem from you, so this has never come up fort me. My systems just stay up and running.

BRRABill

I did a few Google search. It seems like other places (including Adaptec) also say rebooting a rebuilding array is perfectly safe.

Not arguing with anyone, just questioning why the people making the RAID hardware seem to think it is OK, but ML does not.

http://blog.adaptec.com/2015/05/13/a-technical-issue-regarding-raid-buildrebuild/

scottalanmiller

@BRRABill said in Dell PERC Question (Server Down):

I did a few Google search. It seems like other places (including Adaptec) also say rebooting a rebuilding array is perfectly safe.

Of course they do, you are reading their marketing.

scottalanmiller

VMware advocates moving running workloads around, too. But you should fire anyone who did it in production based on that recommendation. I've seen people do that thinking that they would not get caught because "Vmware said it was safe" and brought down a bank because of it (hint: it wasn't safe.)

JaredBusch

@scottalanmiller said in Dell PERC Question (Server Down):

@BRRABill said in Dell PERC Question (Server Down):

@scottalanmiller said

That means that the PERC has failed. That's a different issue.

Are we back to that?

I mean, it's definitely on the table. Just not sure why all of a sudden you think that's so likely.

Because you said that it was rebuilding. The PERC should remain up and viable even without disks attached to it. You can continue to manage it online through the iDRAC or a VM that isn't using those drives.

You do not manage Dell PERC through iDRAC or a VM.

JaredBusch

@scottalanmiller said in Dell PERC Question (Server Down):

@BRRABill said in Dell PERC Question (Server Down):

I did a few Google search. It seems like other places (including Adaptec) also say rebooting a rebuilding array is perfectly safe.

Of course they do, you are reading their marketing.

I've done it many times over the years and never once had a failure. Yes it is a risk, but it is also adding downtime waiting for a rebuild to happen.