A HD in my WD PR4100 32TB server took a crap - need advice

[Louie Detroit]

Hey all, seeking any advice from the many fine minded IT mavens here at the Barn.

I just noticed a red light is on the status indicator of drive number one (of four) slot of my server and was wondering what you would do to best alleviate a data disaster. A quick search showed a similar matching HD may be on back order from Chy-na and not available for a while. The chinks are probably pissed off at the busting of their balloon.

Should I just unplug the server and wait until I can install a replacement drive? I can live without it's availability short term.

I run it on a $130 or so APC UPS and see no problems reported there, no power issues lately AFAIK.

Here are some screen shots related to the issue, what do you think?

[DJ_Critterus]

If the spinner drive is bad, you should be able to recover data. It's all done using parity, so you should be able to get the data back without losing any or much at all being that you used RAID 5. Even if you do lose it, you'll still have bits of the data stored across the other drives.

[Louie Detroit]

Originally Posted by DJ_Critterus

If the spinner drive is bad, you should be able to recover data. It's all done using parity, so you should be able to get the data back without losing any or much at all being that you used RAID 5. Even if you do lose it, you'll still have bits of the data stored across the other drives.

Well, my understanding is with the current hardware/software config it is still 100% good/salvageable. I'm just worried about the implications of a second drive taking a shit.

I give zero ****s about trying to repair the bad drive and wish to replace it most rikki-tik.

[DJ_Critterus]

Originally Posted by Louie Detroit

Well, my understanding is with the current hardware/software config it is still 100% good/salvageable. I'm just worried about the implications of a second drive taking a shit.

I give zero ****s about trying to repair the bad drive and wish to replace it most rikki-tikki.

Do so and then back the data back up. How old are the drives? Do they see a lot pof power spikes or outages? I wouldn't worry too much about the other drives at this point, but do replace the failed one and get your data back to how it was originally stored.

[ZipZap]

Depending on the RAID setup/rules, the drive is already mirrored on one of the other three. I know there's a process to go through to replace one, but if you had it set up that way, you shouldn't need to do any physical recovery.

[Stevedore]

If I recall correctly from my working years, any 3 of your RAID5 drives have all the info needed to rebuild a failed 4th drive. If your hardware is hot swappable, you just unplug the bad drive & plug in the replacement. If not, power it down, replace the drive, & turn it all back on. You might be prompted to recover the data onto the new drive. It takes a while, but the new drive will be rebuilt as a participating member of the array.

From your pics, it looks like a WD drive system. Might be good to look at their documentation for drive replacement procedures.

[CurtP]

Should only have to match the specs for a replacement drive. 8TB, 5400 RPM, 6 Gb/s, 256MB cache according to a quick search of wd80efax-68knbn0

https://www.amazon.com/Western-Digit.../dp/B07D3MWMNZ

Be warned, OEM drives usually don't come with any warranty: https://www.amazon.com/OEM-Drive-dri.../dp/B00V5I65NW

[Louie Detroit]

Originally Posted by CurtP

Should only have to match the specs for a replacement drive. 8TB, 5400 RPM, 6 Gb/s, 256MB cache according to a quick search of wd80efax-68knbn0

https://www.amazon.com/Western-Digit.../dp/B07D3MWMNZ

Be warned, OEM drives usually don't come with any warranty: https://www.amazon.com/OEM-Drive-dri.../dp/B00V5I65NW

Thanks for the info Curt!

[CurtP]

I have the same PR4100 and mine is loaded with WD80EFAX-68LHPN0 drives.

[Frankie the Fink]

Originally Posted by ZipZap

Depending on the RAID setup/rules, the drive is already mirrored on one of the other three. I know there's a process to go through to replace one, but if you had it set up that way, you shouldn't need to do any physical recovery.

Yes - I run an antique D-Link RAID system and recovery should be easy but I would replace that drive ASAP as if another goes down you're SOL...

[Louie Detroit]

Well, I ordered a replacement HD from Amazon and it should be here in a week to ten days. I believe/hope it will be hot swappable, I'll know more in a couple of days when I read up on the replacement process. Thanks for the advice guys!

[CurtP]

Originally Posted by Louie Detroit

Well, I ordered a replacement HD from Amazon and it should be here in a week to ten days. I believe/hope it will be hot swappable, I'll know more in a couple of days when I read up on the replacement process. Thanks for the advice guys!

Not hot-swappable. Power down and replace. Settings > Utilities > Shutdown. Replace drive, then follow the instructions for RAID 5:

https://support-en.wd.com/app/answer...%20replacement

[dvarapala]

Not only can you replace the drive and rebuild your RAID, you can even upgrade your storage space. Throw an 8TB WD Red in place of the bad one, wait for the RAID to rebuild itself, and then repeat the process for the other 3 drives.

Easy Peasey!

[CurtP]

Originally Posted by dvarapala

Not only can you replace the drive and rebuild your RAID, you can even upgrade your storage space. Throw an 8TB WD Red in place of the bad one, wait for the RAID to rebuild itself, and then repeat the process for the other 3 drives.

Easy Peasey!

He has 8TB Reds already. Depending on revision, it will support up to 18TB drives, but if I was going to use drives of that size, I'd want a larger array. I think the memory modules have to be upgraded too.

[ZipZap]

Originally Posted by Louie Detroit

Well, I ordered a replacement HD from Amazon and it should be here in a week to ten days. I believe/hope it will be hot swappable, I'll know more in a couple of days when I read up on the replacement process. Thanks for the advice guys!

Why hot? You have something so time dependent that you can't power down? Just shut down and replace. It should rebuild fine. I chuckle a bit at folks saying to upgrade to more storage. What the hell are they storing? Upgrading to better drive technology, if your drive age dictates, is a good thing.

[CurtP]

Originally Posted by ZipZap

Why hot? You have something so time dependent that you can't power down? Just shut down and replace. It should rebuild fine. I chuckle a bit at folks saying to upgrade to more storage. What the hell are they storing? Upgrading to better drive technology, if your drive age dictates, is a good thing.

It's usually easier to replace hot than during a reboot and is the method I prefer. On some units, if a drive is failed and you try to bring it up, it'll go through a whole diagnostic process, then do a data check which can take ages. I have an old Buffalo TerraStation Pro 4-drive unit that can take days to come up with a failing/failed drive and it's only a 8TB unit. With hot-swap, it's just a matter of replace, config and let it start rebuilding, which may take several days just doing that. I haven't had to replace a drive in the PR4100 yet, so I looked up the process so I'd be familiar when the time comes. My Pegasus R6 is hot-swap and nearly seamless when replacing a hard drive.

I'm always needing more storage. I wish it was as simple as replacing hard drives with bigger ones in the arrays, but not all drives are supported, and then there are drive size limitations. If there wasn't, that R6 would be 132TB instead of 24.

[Louie Detroit]

Originally Posted by CurtP

Not hot-swappable. Power down and replace. Settings > Utilities > Shutdown. Replace drive, then follow the instructions for RAID 5:

https://support-en.wd.com/app/answer...%20replacement

Unfortunately mine is the newer version of the server that doesn’t have >shutdown, it has >hibernate instead. I guess for some reason they don’t want people turning it off on a regular basis.

So I put it into hibernate, and it’s still in the process of doing its thing with a msg something like drive is currently shutting down displaying on the device itself. I assume once it’s entered hibernation I can simply unplug it from the power, but not 100% sure. Looks like I may have to call tech support, I know this hibernate mode appears to really irritate people.

[Louie Detroit]

Originally Posted by ZipZap

Why hot? You have something so time dependent that you can't power down? Just shut down and replace. It should rebuild fine. I chuckle a bit at folks saying to upgrade to more storage. What the hell are they storing? Upgrading to better drive technology, if your drive age dictates, is a good thing.

I was planning on ripping some DVDs which eat up on the storage, but haven’t gotten around to it yet. Out of 23.9TB available I’m using about 4TB of space so yea it’s a bit of overkill.

[ZipZap]

Originally Posted by CurtP

It's usually easier to replace hot than during a reboot and is the method I prefer. On some units, if a drive is failed and you try to bring it up, it'll go through a whole diagnostic process, then do a data check which can take ages. I have an old Buffalo TerraStation Pro 4-drive unit that can take days to come up with a failing/failed drive and it's only a 8TB unit. With hot-swap, it's just a matter of replace, config and let it start rebuilding, which may take several days just doing that. I haven't had to replace a drive in the PR4100 yet, so I looked up the process so I'd be familiar when the time comes. My Pegasus R6 is hot-swap and nearly seamless when replacing a hard drive.

I'm always needing more storage. I wish it was as simple as replacing hard drives with bigger ones in the arrays, but not all drives are supported, and then there are drive size limitations. If there wasn't, that R6 would be 132TB instead of 24.

If I ever got to that size, I'd be outsourcing...

Is that all static storage, like dvd's and such? Or, are you doing something capturing dynamic data that requires such extensive size?

[CurtP]

Originally Posted by Louie Detroit

Unfortunately mine is the newer version of the server that doesn’t have >shutdown, it has >hibernate instead. I guess for some reason they don’t want people turning it off on a regular basis.

So I put it into hibernate, and it’s still in the process of doing its thing with a msg something like drive is currently shutting down displaying on the device itself. I assume once it’s entered hibernation I can simply unplug it from the power, but not 100% sure. Looks like I may have to call tech support, I know this hibernate mode appears to really irritate people.

I think mine changed to say hibernation during the last update too, but I usually hold the power button until it says "shutting down" on the display. As far as I know it's the same thing, so I don't know why it bothers so many. I have no real complaints about mine other than it's slow, even with LACP turned on, but every NAS I've owned has been slow. The R6 is direct-connect to a 2012 Mac mini through Thunderbolt 1 and the network shares to it are extremely fast. I bought that R6 around the same time I bought the Mac mini.

I was just contemplating running Plex directly on the PR4100 last night. It has hardware transcoding, an intel processor and supposedly was designed for Plex. But there are so many complaints about how poorly it performs that I think I'll just keep it on one of the minis.