"Permanent errors have been detected" on ZFS pool

[morxy49]

29-32 are Intel ports - the Marvell are 8, 9, 11, 12, 14 & 15 - see pages 20 and 31 of your manual

Thanks! So it's the Intel controller then.

 

[morxy49]

Anyone has any ideas what I can do? :/

 

[Kevin Horton]

Things to try:

1. Reseat data and power supply cables to the drives.
2. Replace data and power supply cables to the drives.
3. Try a known good power supply.

You never did tell us what the power supply is. Which make and model, and how old is it?

 

[morxy49]

Things to try:

1. Reseat data and power supply cables to the drives.
2. Replace data and power supply cables to the drives.
3. Try a known good power supply.

You never did tell us what the power supply is. Which make and model, and how old is it?

Thanks, I will try all that when I get home tonight.
The power supply is a SilverStone SFX Series ST45SF-G 450 Watt which is 3-4 years old.

 

[Kevin Horton]

The power supply is a SilverStone SFX Series ST45SF-G 450 Watt which is 3-4 years old.

Hmm. I wonder if perhaps 450W might be a bit low to handle 12 drives. It may have been enough when the power supply was new, but power supply capacities decrease as the components age.

 

[Stux]

I'd power down and disconnect the good drives and use their power and Sata cabling etc in this pool.

Would certainly tell you if the issue is power etc.

And with an aging 450W second tier PSU it certainly could be.

 

[morxy49]

Sorry for the late response. I've been extremely busy. I've done some troubleshooting the last days though;

• I tried cleaning the server from dust (it was a lot of dust) and disconnecting the 8 disks that were working. I left the 4 non-working Seagate disks exactly as they were, just re-seating the cables. When booting up again everything looked good. I ran a scrub which completed successfully. Everything was up running for two days straight without any errors at all. So that confirms that there's no issue with the disks, cabling, or controller.
• Today I reconnected the 8 working disks as well, so that all 8+4 disks are connected simultaneously. It's been up for 2 hours so far, and there's no issues yet. Let's see if it can keep going this way?
• Right now I also have a power meter plugged in between the NAS and the UPS, measuring the total power draw from the PSU. It's not drawing more than 95-100 Watts, so I have a really hard time believing it is a power issue. The SilverStone SFX ST45SF-G 450W is also supposed to be a high quality PSU with 80+ Gold.

I will update how it goes later. Right now it looks like just a cleaning from all the dust and re-seating the cabling did the trick, but lets see.

 

[danb35]

That's sounding good. Try a scrub on each pool (separately) and make sure each is showing good.

 

[morxy49]

It's still going strong!