Pool will not Import

[Mac Thompson]

So my FreeNAS box became unresponsive. No GUI, No Console, No SSH. Hard rebooted. In normal boot mode it gets to importing zpool then hangs. In single user mode we can import the pool in readonly mode and all is good but a it comes up saying it cant mount a particular dataset (which I dont really need anyway). Im stuck. Can't import the pool. Cant delete the potentially problem data in readonly mode. Tried import -R in single user mode and it returned no errors. Any ideas how to get the pool back online.

FreeNAS-9.10.1-U2
HP DL360p Gen 9 Chelsio 10Gbp NIC, 2 x Intel(R) Xeon(R) CPU E5-2630 v3 @ 2.40GHz, 786GB RAM,
MDS600 SAS disk chasis via a HP H240 HBA with 3 x RAIDZ vdevs consisting of 7 x 8TB 7200rpm SAS drives (21 drives in total).
2.4TB stripped L2ARC Intel 3500 NVME SSD
400GB mirrored ZIL Intel 3500 NVME SSD.

 

[SweetAndLow]

output of zpool status when you have the pool imported in read only mode? I suspect that you are going to have to backup everything and rebuild the pool. You should also verify that your disks are not failing.

 

[Mac Thompson]

I think the disk are all good:

 

[SweetAndLow]

I'm unsure if that disk controller actually works correctly with zfs. Can you run smartctl -a /dev/da0(or whatever your disk is called)?

 

[Mac Thompson]

actually it looks like the cache is offline

 

[Mac Thompson]

so im looking for ways to remove the cache when I cant import the pool except in readonly mode...

 

[Mac Thompson]

I'm unsure if that disk controller actually works correctly with zfs. Can you run smartctl -a /dev/da0(or whatever your disk is called)?

yes - the controller is all good

 

[SweetAndLow]

yes - the controller is all good

How do you know? You need to start providing information if you expect help.

Sent from my Nexus 5X using Tapatalk

 

[Mac Thompson]

How do you know? You need to start providing information if you expect help.

Sent from my Nexus 5X using Tapatalk

the controller has been running for a couple of weeks without incident and all disk are reporting online and smartctl reports no errors. It looks like its related to the cache disk

 

[rs225]

Depending on the age of the pool, this could be the L2ARC corruption bug.

Whatever the case, your only viable option is probably to offload your data and create a new pool.

If you think the L2ARC devices are the problem, just remove them when shut down. It should be possible to import the pool without them.