Panic on FreeNAS Boot

[rcs123]

Hi

I'm running FreeNAS on a HP Micro Server Gen8 with 12GB of memory.

I use a Santa Cruz USB stick on the motherboard USB port internal to the case, to boot from. In the past I had an older USB stick and that failed and started reporting errors so I cloned it to a new stick and it was fine for a while.

I've also got hold of an SSD and was going to use that as the boot disk replacing the USB until I started having some problems described below.

Its been running fine for a couple of years, but recently the shares exported on the disk pool would stop working after a few minutes. I ran disk health checks and even the whole HP healthcheck on memory disks etc. and nothing was reported as an error.

Even with all these problems the system would reboot fine, and come up ok and then after a few minutes the shares would just disappear. I couldn't see any error logs or anything. The Web Interface would also freeze and the only way to recover was to reboot.

However recently it has developed an error when it panics and drops into the debugger just as the data pool is being mounted on boot.
I get exactly the same message each time.

I have tried removing one disk at a time and the error continues to occur on reboot until there is either no disks or just one disk left.

It may not be the wisest move, but in frustration I booted without the disks and added them in one at a time. The disks were recognised and appeared in the list of disks in the GUI, but the data pool didn't come up (not a surprise) and any attempt to import the data pool fails to provide any disks to select from.

I've reseated memory and cleared any dust etc.

Is there anything I can do ?

Thanks
Richard

 

[Meyers]

I'd suspect a bad boot device and would reinstall to that SSD. If you don't have a config backup you'll have to set everything back up again, but the pool should import.

 

[Meyers]

Oh I see, this is the data pool causing this. I'd probably Google some of those error messages but I'm wondering if your pool is corrupted. I hope you have backups.

 

[rcs123]

I've installed and booted from the SSD. Had some hassle with the HP Micro Server Gen 8 to get it to boot from the SSD with the other disks in. In the end I had to change the controller to legacy in the BIOS and then I had two SATA controllers to select the boot order.
Previously I had set it to use AHCI (but not the built in RAID controller).

After booting and getting the configuration restored (I had a backup), exactly the same error occurs when booting from the SSD.

I didn't have a backup, just didn't get round to it...

Is it worth trying anything else before I trash the disks and rebuild the ZFS data pool?

 

[Meyers]

Previously I had set it to use AHCI

Did this start happening after changing to legacy? I've had odd things happen when I've done this.

 

[rcs123]

It was happening with AHCI too. I had to switch to legacy to get the SSD to be the boot disk.

I have given up and have rebuilt the data volume now with the SSD as my boot device.

I realised tvat having all four disks in the pool as two vDevs of two disks protected me from a couple of disk failures but not from a corruption of the pool or just one of the vDevs.

I am now experimenting with two pools each with a single vDev of two disks as a mirror pair. I am going to set up a replication task from the FreeNas server to itself with the second pool as the target.

I really need a off system backup too and will look at options for this.

Luckily most of my data is on the original devices as I was using the FreeNas mainly as a backup server.

Thanks for your help.

 

[Meyers]

Ahh well, I'd probably Google that panic message a bit but I suspect you're probably going to have to wipe it and start over.

What kind of disks are these? Did you do any sort of burn-in before using them? If not, there's a sticked post on the forums detailing what to do. I do that on every new system I build. I'm not saying that would have prevented this from happening, but it would be a good idea to really stress these disks before relying on them again (and to do this on any new disk).

 

[rcs123]

They are Western Digital blue NAS disks.

I will tske a look at how to burn in test these now. Thanks fir your help and suggestions.

 

[Meyers]

You might want to read through this thread as well. Make sure you have SMART tests and scrubs running regularly.