Problem finding disks and/or HBA after power outage

[yarcod]

Hi!

I've had my server with a virtualised instance of FreeNAS running for more than two years now without any issues. Until the other day, when we had a power outage at home and bad shit started to happen.

First of all, I have FreeNAS running as a VM under ESXi, with a 1015M in passthrough. Hardware otherwise is a SuperMicro X10SDV-6C-TLN4F with 64GB RAM, 3 WD Red 4TB in RAIDZ1 (I know, I know, I just cannot afford getting more disks :/ ). When power returned I realised that ESXi somehow had been corrupted and could not boot. Hell, I'll need to reinstall it. Upgraded from 6.0 to 6.7U2 at the same time. I also upgraded the VMs to version 6.7 -- I actually don't really know what that means, though. Then I started FreeNAS and it looked fine except I could not see my disks. My main pool was empty and reported status "unknown".

I would guess that I simply need to "re-attach" my HBA card or disks to FreeNAS, but I cannot find any information on how to do that. Searching lead to nothing, which makes me believe it's too simple to have been mentioned by anyone!

Thanks in advance!

 

[co0kiemon5t3r]

Upgrading may have detached the card, I think there were some changes in 6.7 May be worth running through the steps to detach and re-attach and passthrough the card to the VM?

If you have another system capable of taking that many disks, you could try to import the volume to confirm data and disks are good, if you suspect the HBA card died in the outage (unlikely)

 

[HoneyBadger]

May be worth running through the steps to detach and re-attach and passthrough the card to the VM?

This is my suggestion as well. Likely it's as simple as the new install of ESXi not having the passthrough feature enabled, so your FreeNAS VM can't access the HBA.

 

[F!$hp0nd]

Two things, make sure you have the correct driver installed on ESX 6.7 for your HBA card. I know 6.7 dropped preinstalled drivers for a bunch of cards that it use to natively support. Make sure to use the esxcli update command not install (as the install command overwrites essiential files, while update just places installs critical packages but does not overwrite dependencies). Once this is done, reboot the box and then go to the ESX web ui, go to Devices under Manage on the web and then select the HBA card and click the Passthrough button. Hope this helps.

 

[yarcod]

Two things, make sure you have the correct driver installed on ESX 6.7 for your HBA card. I know 6.7 dropped preinstalled drivers for a bunch of cards that it use to natively support. Make sure to use the esxcli update command not install (as the install command overwrites essiential files, while update just places installs critical packages but does not overwrite dependencies). Once this is done, reboot the box and then go to the ESX web ui, go to Devices under Manage on the web and then select the HBA card and click the Passthrough button. Hope this helps.

I ran an update and reattached the HBA to the FreeNAS VM again and it seems to work. However, it is doing something very CPU intensive right now but I can't really seem to find what. I can access my data and everything (as far as I can see) but I'm curious what it is doing. Any ideas?

Anyway, many thanks for the help @co0kiemon5t3r, @HoneyBadger and @F!$hp0nd! :)

 

[yarcod]

While the disks and pool is back, I cannot really use FreeNAS as it is. CPU usage is really high (which makes my passively cooled server unusable) and it seems to be due to the pool itself. I suspected that I might have done some weird things previously, and that that had been the reason for the slowdown. So I decided to create a new VM and import the pool there instead. The issue with high CPU usage remains! I don't really know where to start looking - what in the pool could do this? I should mention that the pool is almost filled up, if that makes a difference

 

[co0kiemon5t3r]

While the disks and pool is back, I cannot really use FreeNAS as it is. CPU usage is really high (which makes my passively cooled server unusable) and it seems to be due to the pool itself. I suspected that I might have done some weird things previously, and that that had been the reason for the slowdown. So I decided to create a new VM and import the pool there instead. The issue with high CPU usage remains! I don't really know where to start looking - what in the pool could do this? I should mention that the pool is almost filled up, if that makes a difference

You will see degraded performance above 80% util n the pool as a rule of thumb, but it shouldn't hit CPU too much.
I remember a bunch of options for power saving and allocating CPU cores, have you exhausted settings on ESXi?