They must be referring to ambient temperature, 40C is too low for normal chip temperature.

Update: My two new (used) Dell H310's arrived from eBay yesterday. I flashed them each with the latest IT firmware from the IR/IT modes thread. Installed the first one, booted up unRAID and all was well.

But about 45mins after I went to bed the card dropped a drive again and now is "non-operational", same as the old card.

I have a Startech slot exhaust fan in the slot below moving hot air away from the Dell card, which I didn't have before.

Thanks for any ideas.

tower-diagnostics-20191219-1056.zip

Did you/can you try the HBA in a different slot? Or a different board?

I can try another slot later today, currently removed the new card and running Memtest to rule out memory. Don't really have another board to test with, I have two other PCs here, but they're in use, and different generation CPUs etc. But I'm not opposed to getting another board if we can determine this one isn't stable anymore. Thanks.

So memtest ran all day, no errors. And now I've tried the HBA in every possible slot, but unRAID shows missing drives each time.

Next thing to try would be a different board.

Yeah makes sense, I will track down another motherboard.

Possibly stupid question I guess, but is it possible the H310 isn't compatible with the 10TB BarraCuda Pro ST10000DM0004 for any reason?

My trouble started when I upgraded one of my 4TB parity drives to this 10TB drive, which I had precleared successfully.

Anyway, I've now moved the 10TB drive from a port on the Dell card, to one of the motherboard ports, and now when I boot unRAID it's showing all drives again, no dropped drives yet. It's been stable for a couple hours, although I didn't start the array yet. I've now shutdown and re-installed my GTX 970, and moved the Dell card to a lower slot. Will see if it remains stable.

System seems stable so far, three hours and no dropped disks yet. But I haven't started the array yet, as Disk 2 has a Red X beside it. Diagnostics attached. Thanks for any replies. I'm lacking in SMART report reading ability.

tower-diagnostics-20191223-0032.zip

11 hours ago, stor44 said:

Possibly stupid question I guess, but is it possible the H310 isn't compatible with the 10TB BarraCuda Pro ST10000DM0004 for any reason?

There's a known issue with those disks and LSI controllers, there's a firmware upgrade available, more info here.

Well isn't that interesting! Unfortunately my drive is a different model. Not Ironwolf, these are Barracuda Pro. Certainly sounds like a similar issue though. Will look into that. Will contact Seagate too, although these are shucked drives, so I doubt they can help.

 

Any thoughts on how to handle the Red X on drive 2? I swapped SATA cables and plugged it into a motherboard port, no change. If the SMART looks ok, should I try this?  https://wiki.unraid.net/Troubleshooting#Re-enable_the_drive

unRAID is up for 14 hours now, no further drive problems, seems stable.

Yep, different model, my mistake, but could be a similar issue, as for the disabled disk if it looks good and emulated disk is mounting correctly you can re-enable. 

Ok, rebuilding the failed disk now, will see in about ten hours.

It got to about 94% complete rebuilding Disk2, but now it's saying Parity2 has errors. Diagnostics attached.

It was previously working at about 60MB/s, now it's slowed to 100kb/s, and says 25 days left.

Thanks.

tower-diagnostics-20191223-2345.zip

Well it finished rebuilding Disk2. It says everything is ok.

tower-diagnostics-20191224-0128.zip

SMART for parity2 is incomplete, reboot and post new diags.

Ok rebooted, new diagnostics here.

tower-diagnostics-20191224-1003.zip

It was a disk problem, though these errors can sometimes be intermittent, but never a good sign, you can run a SMART extended test.

Latest update: Instead of an extended SMART test on Parity2, I removed that drive and replaced it with a new 10TB Seagate, plugged into the motherboard. The parity check completed fine.

I upgraded to unRAID 6.8.0 and updated any plugins and Docker apps. This all ran fine over the weekend.

Today I've replaced the oldest hard drive, Disk6, with a precleared 10TB Seagate, but now several disks have dropped out again during the parity check.

Logs attached.

Fortunately, I have now transitioned my preferred Windows software (Blue Iris, Retrospect) to a Win10 VM, so I can now use my 2nd PC for testing if needed.

Thanks, please advise.

tower-diagnostics-20191230-1203.zip

HBA problem again:

Dec 30 11:31:33 Tower kernel: mpt2sas_cm0: SAS host is non-operational !!!!

 

Thanks. Trying again with the other HBA I got from eBay, in a different slot.

Update: Still had HBA card failing in different slots. So I've given up on my Gigabyte X58A-UD3R, and moved my unRAID setup to my other PC consisting of:

 

Case: Fractal Design R2 XL EATX  //  MB: ASUS Maximus VI Formula, 1603 BIOS  //  Intel Core i7-4790 3.6 GHz

RAM: 24GB HyperX (3x8GB)  //  Dell Perc H310 //  Array: 1 x 10TB Seagate Barracuda Pro // 5 x WD 4TB RED  //  Parity: 2 x 10TB Seagate Barracuda Pro  //  Cache: WD Blue 500GB SSD x 2

Docker: ChannelsDVR  //  Plex  //  Sonarr

unRAID Pro 6.8.0

 

Only problem I'm seeing now is a "softreset" on Sata8 during bootup. Is that a concern other than making it take longer to boot unRAID? Thanks.

 

tower-diagnostics-20200102-0645.zip

Edited January 2, 20206 yr by stor44
Added missing question.

Update: Ok I think things are finally stable. Since moving unRAID and the HBA to my ASUS motherboard, no issues. I'm now running all three 10TB Seagates off the HBA. So yeah, just some random incompatibility with the HBA, old motherboard, and these 10TB drives? I just have my 2 cache drives and optical drive on the motherboard SATA ports.

 

I've also repurposed my old unRAID PC by installing Windows 10. Still using a Dell H310 in there, and DrivePool to pool the drives to backup unRAID. No issues there yet, but it's a mix of old 3TB & 4TB WD Red drives. I'll stay away from adding 10TB drives to that setup.

 

Thanks for all the help, this is an outstanding community.

tower-diagnostics-20200113-0517.zip

24 minutes ago, stor44 said:

So yeah, just some random incompatibility with the HBA,

Apparently yes, though the link I posted above is for different models I've read since on another forum that it also affects that model, and there's no firmware update from Seagate for now, so they shouldn't be used with LSI HBAs.

I’ll keep my eyes open for drive firmware updates. It’s weird that they are working now, connected to the HBA. I think my old motherboard was playing a role.

Anyway, I’ll mark this as solved and move on. Thanks.