Proposed FreeNAS Build

[danb35]

should find pools of all sorts, healthy, borked, attached, detached.

No, zpool status only reports the status of attached pools.

 

[thedirtmerchant]

Hmmm, I must have been thinking of a different thread where zpool import had returned nothing. OK, so you have a pool, and zpool import is showing it, but when you try to import it, the system crashes? That normally indicates a corrupt pool, which is a Bad Thing (tm). What happens if you try zpool import -F -n Data?

Code:

[root@freenas ~]# zpool import -F -n Data1								   
cannot import 'Data1': pool may be in use from other system, it was last acce
ssed by freenas.local (hostid: 0x3fcbd2c2) on Sun Mar 12 22:06:28 2017		 
use '-f' to import anyway

 

[diedrichg]

Ah! Well there you go! Now at least you have an answer. @danb35 would have to give the blessing but I believe that is what you need to do, -f (force) import it.

 

[thedirtmerchant]

I'm so amused by the zpool import F n Data1 aka "Zpool, Import the Fucking Data!"

 

[danb35]

So far, that's looking OK. zpool import -F -f -n Data.

The -F flag would attempt to force the import, possibly rolling back the pool to a previous consistent state. However, the -n flag says to not actually do that, just report if the system thinks it will work.

 

[thedirtmerchant]

Wow, I put in the command, hit enter, it hung for several seconds, then heard a beep from the tower...seemed to reboot.

 

[thedirtmerchant]

Tried it again from the shell on the server. Basically got the same issue from post88 in this thread.
Scrolling text and/or hex for a bit (much shorter this time, maybe 20 seconds instead of 2 minutes).
Stops for maybe a second.
It says something about
"Tracing command kernel"
"Text dump complete"
and "cpu_reset: Restarting BSP"
and "cpu_reset_proxy: Stopped CPU#
then immediately reboots.

 

[thedirtmerchant]

Just to be sure, I'm supposed to have 3 options here right.
-F
-f
-n

?

 

[DrKK]

That doesn't sound good. I'll go digging in the thread for information on your config, and if I have any comments, I'll post them momentarily.

 

[DrKK]

So I couldn't find it if you posted it, can you give us your exact hardware? Mobo, amount/type of RAM, type of disks, CPU, etc., for reference?

 

[thedirtmerchant]

So I couldn't find it if you posted it, can you give us your exact hardware? Mobo, amount/type of RAM, type of disks, CPU, etc., for reference?

FreeNAS 9.10.2-u2
Supermicro X11SSL-CF mobo
32 GB (2 x 16) Crucial DDR4 PC4 ECC RAM
Intel Xeon E3-1230 v5
SanDisk 120GB SSD (Boot)
2 x 3TB WD Red drives (from previous FreeNAS box, which ran on 9.2.1.3)
Fractal Design Define R5 case

 

[Ericloewe]

If it panics with -n, things are really in poor shape.

 

[diedrichg]

His (2) 3TB disks were in RAIDZ1. It was functioning fine until an issue with the previous hardware, thus the hardware upgrade. He'll have to give the previous hardware specs as well as the issue that arose prompting the upgrade.

 

[danb35]

His (2) 3TB disks were in RAIDZ1.

That's impossible (RAIDZ1 requires at least three disks), and also contradicted by his own earlier post. The output of zpool import indicates that they were mirrored.

 

[danb35]

Just to be sure, I'm supposed to have 3 options here right.

Correct. I'd have to agree with @DrKK and @Ericloewe that this is a very bad sign. I'm really not sure what else to try--as I mentioned earlier, there may be some things that zdb could do, but that's a tool I haven't messed with before.

 

[thedirtmerchant]

That's impossible (RAIDZ1 requires at least three disks), and also contradicted by his own earlier post. The output of zpool import indicates that they were mirrored.

This is my lack of knowledge showing, apologies.

 

[thedirtmerchant]

His (2) 3TB disks were in RAIDZ1. It was functioning fine until an issue with the previous hardware, thus the hardware upgrade. He'll have to give the previous hardware specs as well as the issue that arose prompting the upgrade.

It was actually drive errors that started everything, then the USB key failed, then I decided to do the job correctly with the new hardware.

Basically, I'm thinking the drives are toast...

having said that...would I be able to only connect one of the drives and try the import?
As maybe one of the drives is really bad and the other is only semi-bad?
Or if that was the case, it would import w/o issue?

 

[danb35]

would I be able to only connect one of the drives and try the import?

You certainly could try it. I doubt you have anything to lose, though I'd honestly be surprised if it helped.

 

[thedirtmerchant]

You certainly could try it. I doubt you have anything to lose, though I'd honestly be surprised if it helped.

I'll be trying this later today.

 

[thedirtmerchant]

Next question. I'm starting my burn-in/stress testing.
I'm doing the CPU test using Breakin Version 4.26.1-53.
It's running now and I expect to run for about an hour.
Question though, in the list of tests "ecc" is listed it has a 1 under Pass, 0 under Fail, but the "Last Message" listed by ecc is "No support for ECC checking on this system".
And underneath, I guess listing how long the tests us "22s: Disabling burnin test 'ecc'".

Is that something about Breakin doesn't test ECC? Or (mini-heart attack) I don't have ECC RAM loaded?