Everything working fine, today my server decides to have multiple dis failures.

 

So I had a power issue a day or so ago, got the "unclean shutdown" error despite NUT server doing it's thing, all disks were asleep, the cache (SSD) was the only drive in use.

 

Did a parity rebuild going fine, then unRaid decided 2 disks were bad, so I started to rebuild them, then it decided another disk is bad, then another, then another.

 

Now I have an unbelievable 4 disks bad. Crazy.

 

Is this a bad HBA?

 

Any ideas? I think Disk 7 is genuinely bad, the rest are ok.

 

I just stopped the server, because I am fuming right now.

You only have 2 disabled and dual parity so you should be able to start the array.

 

Start the array in normal (not maintenance) mode, then get Diagnostics and attach to your NEXT post in this thread.

 

 

Thanks Trurl, I think it's a power issue. I was lazy when I expanded my array from 12 to 16 drives.

 

Just opened her up and I have splitter cables everywhere. I'm going to the local PC store tomorrow to buy some 8 pin to 2 x Molex, and a couple of 6 pin to 2x Molex. I always have issues at start up / parity sync, so only when all drives are fired up. I'll swap out the cables and report back. I am going to put less drives on a string, I just looked and I had 8 drives running off a splitter, problem drives 5, 6 & 7.

 

Lesson learned #1.

 

I need to have a proper backup. Time to set up the Zima Cube I bought.

 

Lesson learned #2.

 

Think about power supply to drives, especially when they all spin up together.

Don't rebuild if any disks are unmountable. Filesystem repair to make disks mountable should be done before rebuilding.

 

After you get cables fixed post diagnostics if you want more advice.

21 hours ago, trurl said:

Don't rebuild if any disks are unmountable. Filesystem repair to make disks mountable should be done before rebuilding.

 

After you get cables fixed post diagnostics if you want more advice.

How do I do file system repair.

 

I just replaced my power supply (a great 1000w PSU with 6x SATA lanes), the disks that went AWOL are now showing, but 3 x WD 12tb are now missing. Mind boggling. It's not cables, not power, just FS errors now I would say. I want to get these drives back again before doing anything.

 

I did a clean shut down, so no idea why these just decided to die, as you can note, I have 6 of these WD 12tb, 3 died.

 

Jesus this new PSU id noisy.

 

I think I'm inn the territory;

 

 

4 hours ago, CasaP said:

How do I do file system repair.

We don't know if you need that, or if you can.

On 3/31/2025 at 9:00 PM, trurl said:

After you get cables fixed post diagnostics if you want more advice.

 

Well that was exciting. So what happened was that when replacing the power supply 3 WD drives had their fuses blown (apparently quite a common thing). I brought them into my friendly Polish electronics repair shop and they brought them back to life by replacing the barely visible SMD fuses.

 

Popped them back into the server, they mounted without issue.

 

The next problem was the two drives I mistakenly started to rebuild (before I realised the power supply was dying), the repair finished and they were both marked "unmountable". I followed the Disk Repair Procedure using the GUI and they both had a "zero log" error. They refused to mount, so I just cleared the log with the GUI and ran another file check, which both drives passed.

 

All back up and running. I'm going to swap those WD's out for 3 new Toshiba 18Tb drives in the next few days and finally power up the Zimacube Pro backup box I bought, which I am going to install unRaid onto.

 

unRaid really is a sublime piece of software that really keep data safe no matter what we throw at it. I learned a big lesson, relax and analyse the problem before pushing buttons.

 

Thanks @trurl for your attention.

Just now, CasaP said:

when replacing the power supply 3 WD drives had their fuses blown (apparently quite a common thing)

The common reason is trying to reuse modular PSU cables from a different PSU. There is no standard pinout for modular PSU cables.

 

Did you end up with a lost+found user share from filesystem repair?

4 hours ago, trurl said:

The common reason is trying to reuse modular PSU cables from a different PSU. There is no standard pinout for modular PSU cables.

 

Did you end up with a lost+found user share from filesystem repair?

That was exactly the problem, the new PSU was extremely picky about what cable was used. Crazy stuff. Even the main PSU to CPU cable wouldn't work until I changed it of the one that came in the box.

 

Also, as you guessed I am dealing with the "lost+found" issue now, manually moving things back in place. Is there a faster way to do this? I am using Krusader synchroniser to get things back to normal, it's taking forever.

5 minutes ago, CasaP said:

PSU was extremely picky about what cable was used. Crazy stuff. Even the main PSU to CPU cable wouldn't work until I changed it of the one that came in the box.

4 hours ago, trurl said:

There is no standard pinout for modular PSU cables.

You should assume all modular PSU are picky and only work with the cable that came in the box, or one marketed specifically for that model. Don't make that mistake again.

 

On 3/31/2025 at 9:00 PM, trurl said:

After you get cables fixed post diagnostics if you want more advice.