[Last Call] Learn how to a build a cloud-first strategyRegister Now

x
  • Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 1925
  • Last Modified:

Hot Swap HP RAID 1 Drive...

Hey guys,
Probably a retarded question.
We have an HP ML350 G5 server, 6x disks, 3x RAID 1 arrays.
SMART has predicted the failure of a disk on the C Volume. It's flashing away.
The drives are hot swap.
So am I 100% ok to just pull this disk out and put the next one in, and the server will automatically start mirroring the drives.
Nothing to worry about.
it's literally that easy...
Can someone confirm.
RAID 1 and it's the OS Volume (SBS 2003).
Thanks!
0
optimus_nz
Asked:
optimus_nz
  • 7
  • 6
  • 4
  • +3
1 Solution
 
kinecsysCommented:
Hi there,

100% GO

Hot Swap means you pull it out while the server is running and you plug the new one right in and it starts rebuilding automatically. I have done this on HP servers many times for RAID 1 and RAID 5 arrays.

JUST MAKE SURE YOU PULL OUT THE BAD ONE, OTHERWISE YOU'LL MESS THINGS UP.
0
 
BelushiLomaxCommented:
I totally disagree. If it is in predictive failure, you need to knock it offline (preferrably during POST or using HP ACU but offline is preferred) then possibly rescan adapter, then add. All easy in the GUI (although Really old school) I've had the same situation that blue screened my DC even though tech support told me to go. It *is* your system drive after all...take a few minutes to make sure nothing gets messed up. If it had failed, yanking it out isnt a bad idea, you just need to be SURE the drives-adapter are hot swappable.
0
 
optimus_nzAuthor Commented:
A battle brews... anyone else?
0
Transaction-level recovery for Oracle database

Veeam Explore for Oracle delivers low RTOs and RPOs with agentless transaction log backup and transaction-level recovery of Oracle databases. You can restore the database to a precise point in time, even to a specific transaction.

 
kinecsysCommented:
Well, like I said, I have done it without taking the disc offline, but it is so easy to take it offline that I don't see why you can't do it just in case. Just go to your HP Manager in Windows, expand the virtual drive in question, select the physical disc that's in predictive failure and change its status to offline and make sure to execute the action. Then pop it out and replace it with the new one. However, by doing it this way, you may need to go back in the HP  manager and add the new disc to the degraded RAID and rebuild it.
0
 
noxchoCommented:
It must work in Hot Swap operation as it is DESIGNED to do so. But there are always unpredicted issues. So I would first take backup and only then replace the drive.
0
 
optimus_nzAuthor Commented:
Noxcho - "it must work in hot swap operation"
are you saying that if I take the drive offline and then replace it won't work?

I intend to image the drive using shadowprotect as a precaution first.
0
 
livianCommented:
Just identify the drive from system management homepage  and  replace the failed drive with a new spare!
Good luck!
0
 
noxchoCommented:
Noxcho - "it must work in hot swap operation"
are you saying that if I take the drive offline and then replace it won't work?
No. I was referring to your plan. If you take it offline it works for sure. But if the RAID has Hot Swap feature designed then it should would also without taking it offline.
0
 
andyalderSaggar makers bottom knockerCommented:
There is no option to take a drive offline in the ACU, you just unplug it and replace. It should be done hot since the controller can get confused if you replace it cold if there is data on the replacement.
0
 
optimus_nzAuthor Commented:
OK. So plan : Image the drive (for safety) and swap it hot.
Are we all agreed...
Quite interesting all the varying thoughts on something that is supposed to be straight forward!
0
 
livianCommented:
Or the best answer is to make a call to HP , so you will have an official response :)
0
 
noxchoCommented:
optimus_nz, agreed. Image and replace.
0
 
BelushiLomaxCommented:
I just have been burned by ones that were "Definitely Hot Swappable" so I tend to err on the side of caution. It is a simple thing that is being "overkilled", but better to do it here than watch a server crash and burn, right?
0
 
optimus_nzAuthor Commented:
So what's your step by step suggestion BelushiLomax?
0
 
BelushiLomaxCommented:
I prefer to down the server, use the Array Config during boot, knock it offline, remove, rescan adapter, add drive, rescan and allow it to build while you let the server complete the boot process.
0
 
kinecsysCommented:
Well, I think shutting down the server was not an option in the original question. If you want the official HP answer it is to make a backup, then simply pull the BAD drive out and replace it with the new one. This can only fail if something unpredictable occurs, but unpredictable events can occur in all procedures so all you can do to prepare for those is have a good backup (which you have said you will have). The present status of the disk is no object in whether you can pull it out or not. All you have to worry about is the disk (or disks, depending on the RAID type) that are staying plugged in. In your case, looking at the specific pair of disks that are part of the mirror array in question, as long as the disk staying plugged in is in good shape, the OS won't even notice when the bad disk comes out because the RAID controller is ready to continue using whichever one is left when one of them fails. The RAID controller is ready for a sudden disk failure, connection failure, etc. and it can handle all of those types of "removal" in the same way: by letting the remaining disk continue to work like nothing has happened. Pulling a disk out manually from the front panel of the server is no different to the RAID controller than having the SATA/SAS cable go bad at an unexpected time. That is what it's designed to deal with. Shutting down the server definitely defeats the purpose of Hot-Swap technology altogether.
0
 
optimus_nzAuthor Commented:
Hi Guys,
Thanks for all the information. We're scheduling to do an image of the server, then will hot swap and monitor.
I'll post results.
0
 
kinecsysCommented:
So how did it go?
0
 
optimus_nzAuthor Commented:
it's happening right now... 2 minutes to go of the imaging then i'll reboot and hot swap.
0
 
kinecsysCommented:
Daddy, daddy... are we there yet? are we there yet? are we there yet? :)
0
 
andyalderSaggar makers bottom knockerCommented:
Never, in the course of human conflict and machine failures, has it taken so long to replace so small a part.
0
 
optimus_nzAuthor Commented:
Haha. Rebooted the machine to Shadowprotect CD, imaged the server offline without a problem, rebooted to Windows, logged in and allowed everything to start up. Pulled the drive and installed the new one, fired up the CLI and watched what was going on and watched as it found the drive and started the rebuild. Completed in no time, all systems go!
Thanks guys!
0
 
kinecsysCommented:
Sweet! It did exactly what it is designed to do.
0
 
BelushiLomaxCommented:
I love it when that happens...it's the times that it doesnt that cause threads like this...
0

Featured Post

Free recovery tool for Microsoft Active Directory

Veeam Explorer for Microsoft Active Directory provides fast and reliable object-level recovery for Active Directory from a single-pass, agentless backup or storage snapshot — without the need to restore an entire virtual machine or use third-party tools.

  • 7
  • 6
  • 4
  • +3
Tackle projects and never again get stuck behind a technical roadblock.
Join Now