Large Disk Failure Help

[trurl]

Just make sure you don't agree to format anything.

[tential]

4 minutes ago, trurl said:

Just make sure you don't agree to format anything.

I'll make sure to check back in here before formatting anything.  I've got 3 new drives, so I'm sure I have to format them at some point to add them in.  That's how this next round of issues started.  I tried to bring the 3 new drives online (They were always plugged in with no sata cables attached).

And so far so good, no errors, looks like it was me going above the amps.

On the HDD thread it says my HDD can do 34 amps.

30 amps from 15 HDDs

5 amps from mobo

Then fans/controller cards as I added more and more drives.  Looks like I hit my power limit.  Next time I'll remember to do a quick sanity check first and ask.

I think 1300W should be enough this time around, but I'll still make sure since I usually go by watts, and if I go by watts in that thread I get a different result and that's not the proper way to do it I guess.  Going to try to upgrade to a Dual Xeon/Nvidia GPU setup by the end of the year for gaming/mining.

[trurl]

1 hour ago, tential said:

I've got 3 new drives, so I'm sure I have to format them at some point to add them in.

If you are adding them to a new data slot then they will have to be formatted after adding them to the array.

 

If you use them to replace and rebuild an existing data disk then you absolutely must not format them. Format is never part of a rebuild.

[tential]

4 hours ago, johnnie.black said:

They might have failed because of power issues, but it could also be bad luck, impossible to say for sure, W4J1LSZ0 is currently unassigned (it was disk10 IIRC) and disk9 is also failing, hopefully with the new power supply the rebuild will now go OK.

 

It's telling me Disk 9 has errors

Here is my diagnostic report:

tower-diagnostics-20180228-1532.zip

 

What happens now?

tower-diagnostics-20180228-1532.zip

[JorgeB]

That's expected since disk9 is failing, hopefully there are only a few bad sectors so the rebuild of disk10 can be mostly OK, but there will always be some corrupt files.

[tential]

2 hours ago, johnnie.black said:

That's expected since disk9 is failing, hopefully there are only a few bad sectors so the rebuild of disk10 can be mostly OK, but there will always be some corrupt files.

Ok cool, that's what I was expecting.  The email feature was nice, I got it working now so I knew right away

So far so good.  Time keeps extending a bit, looks like it will finish tonight while I'm asleep.

[tential]

So here is my latest Diagnostic report now that the assembling disk 10 is finished.tower-diagnostics-20180301-0101.zip

What next?

[JorgeB]

Replace disk9, you can also sync parity at the same time.

[tential]

Where can I look to get an idea of where disk 9 (and the old disk 10 which is now labeled sdq as I guess I'll have to replace that next) is located in my system as in where it's connected?  I'm not sure how to get that info I've been poking around my diagnostics looking for it.

[JorgeB]

Both

 are connected to the 4 port Marvell controller, check the serial numbers for which is which.

[tential]

3 hours ago, johnnie.black said:

Both

 are connected to the 4 port Marvell controller, check the serial numbers for which is which.

 

Thanks, they have to be the bottom most HDDs then.  That's quite easy.  

Ok another issue just happened to me though.

I did my sync of Drive 10 to an 8TB drive (All I have).

I just took the array offline to do what you said.  It's now telling me that it's expecting the drive to be 5TB in size, because that was the original drive size. However, I used an 8TB  drive to restore that drive.  What do I do now?

[JorgeB]

 

You just assign the 8TB drive and start the array, it will expand the file system.

[tential]

2 minutes ago, johnnie.black said:

 

You just assign the 8TB drive and start the array, it will expand the file system.

It's telling me it will delete all the data?

That's ok? 

[JorgeB]

4 minutes ago, tential said:

That's ok? 

No, I though you were talking about disk9, post current diags.

[tential]

Here is my current diagstower-diagnostics-20180301-0702.zip

[JorgeB]

OK, I think I know what happened, you replaced the old 5TB disk10 with the new 8TB while its status was invalid, so unRAID didn't consider a new disk and kept the original filesystem size, this can be considered a bug though it's an unusual situation.

 

Unfortunately this means you'll need to rebuild disk10 before replacing disk9, to be safer I would use a new disk (e.g. the one you were going to use to replace disk9) and keep the previously rebuilt one untouched until the rebuild is done, after it's done you use it to replace disk9.

[tential]

Ok, well this time it's giving me the correct 8 TB Total Size, too bad I'll have to wait til tomorrow though.  Oh well!  No biggie!  Thanks.

[tential]

This time there were no errors on Disk 9.  Is that expected?

I didn't stop the array yet this time around, not sure if that matters.  If I'm all good, I'm ready to start rebuilding Disk 10/Parity. 

tower-diagnostics-20180302-0535.zip

[JorgeB]

17 minutes ago, tential said:

This time there were no errors on Disk 9.  Is that expected?

No, the same or more errors were expected, disk has no pending sectors now, which shouldn't happened just from reads.

 

18 minutes ago, tential said:

If I'm all good, I'm ready to start rebuilding Disk 10/Parity. 

You mean disk9/parity, right?

[tential]

11 minutes ago, johnnie.black said:

No, the same or more errors were expected, disk has no pending sectors now, which shouldn't happened just from reads.

 

You mean disk9/parity, right?

Err Yes Disk 9 / Parity.

 

Ya, that's odd Disk 9 is fine I'm not sure how that happened!  Definitely has to be related to my old power supply.  

 

Can I start rebuilding Disk 9 to a new drive/parity or is there something else I should do first? 

Do I still need to do  anything to Disk 9?  Is the disk worth keeping once I take it out or should toss it?

[JorgeB]

It's probably still a good idea to replace it, but you can also just resync parity for now and if there are no errors again it's good for now.

[tential]

10 minutes ago, johnnie.black said:

It's probably still a good idea to replace it, but you can also just resync parity for now and if there are no errors again it's good for now.

 

Ok, I just did Parity for now.  I'll check back in tomorrow let ya know my error situation.  Thanks!

[JorgeB]

I forgot to mention before and almost forgot again, the cache pool is showing some issues, it also completely full, when possible move everything you can to the array, backup any remaining important data and run a scrub, check that there are no uncorrectable errors.

[tential]

20 hours ago, johnnie.black said:

I forgot to mention before and almost forgot again, the cache pool is showing some issues, it also completely full, when possible move everything you can to the array, backup any remaining important data and run a scrub, check that there are no uncorrectable errors.

 

Ya, Transmission turns on when I start the server and if I don't turn it off it can fill my cache drive up really quick.  I'm going to move everything second I get everything fixed.  

I attached my latest diagnostic now that it's finished.

Again no errors!

Should I take care of the cache drive now or is there another step to take first?

tower-diagnostics-20180303-0543.zip

[JorgeB]

10 minutes ago, tential said:

Should I take care of the cache drive now

Yes, don't forget to empty it and backup first before running the scrub.