Unknown relocation type 8bffffe8 in PLT

[PierceIt]

I recently upgraded to 9.2 - no other changes made since then and the system has been running fine for about a week, but today had a yellow blinking light on the Alert icon, after clicking it it told me that the status of volume1 (RaidZ1) is UNKNOWN.

When I dropped into the Shell and typed "zpool status" or "zpool -v volume1" i get the following response each time:

"Unknown relocation type 8bffffe8 in PLT"

I've googled and searched all over the manual but can't figure out what this means. The server and all it's functions seem to running just fine and the system is responsive.

I have a good backup of the config file and a backup of volume1 onto an external eSATA drive with UFS filesystem and that backup volume is HEALTHY

I re-ran a SCRUB, but of course can see the results of the scrub because i get nothing back when typing "zpool status"

Any ideas ?

thanks in advance,
Ron

 

[PierceIt]

I decided to reboot the server and after the reboot it came up with the Alert icon GREEN and indicating the ZFS volume is now HEALTHY
I was able to run the spool status command too and it said there are no errors but some ZFS features are not in use and I should upgrade ZFS. here is the screen shot of that.

Any ideas/thoughts/comments ?

 

[Dusan]

That isn't a ZFS error. It actually points to a problem with the zpool executable.
How did you run the scrub? By scheduling it in the GUI, or by running zpool scrub in the CLI?
Right now my best guess is that your USB stick is corrupt and you should reinstall it.

 

[PierceIt]

Thanks for the reply Dusan

I scheduled the scrub via the GUI

Just curious what leads you to the USB stick possibility? While it is a new 32GB SanDisk stick, i did run an upgrade on this stick from 9.1.1 to 9.2 via the GUI upgrade option recently.

if memory serves me from what I have read in the manual doing a re-install on the stick would be something like:

• downloading the 9.2 install for the stick (as opposed to the GUI install i have now)
• burn the install to a new stick (bootable)
• boot the machine from new stick
• load the server config from the back-up i made
• auto-import the volumes
• done?

 

[Dusan]

Interesting, so the executable works again after a reboot. I reviewed your past posts and your system seems to be experiencing:

• random time drifts (http://forums.freenas.org/threads/locking-up-periodically-and-gaps-in-report-info.16782/#post-90078)
• processes exiting on signal 10 (http://forums.freenas.org/threads/exited-on-signal-10.11754/#post-91893)
• and now (temporarily) corrupted binaries

Would you care to provide your HW details? I also strongly suggest you run a memory test.

 

[Dusan]

I scheduled the scrub via the GUI

That scrub didn't run at all as the zpool binary was corrupted. You can see that your zpool status output shows no that there were no scrubs performed.

Just curious what leads you to the USB stick possibility?

A corrupted file. However, as it "fixed itself" after a reboot and based on the previous issues you mentioned I now suspect a more serious problem with your HW. Please post your HW details and run a memory test. Are you using ECC memory?

 

[cyberjock]

Oh, Dusan. You stole my thunder.. I'm leaning heavily towards bad RAM. If this is the case, it'll be interesting to see the final result of his pool.

WARNING: DO NOT START A SCRUB AT THIS TIME!!!!!! IF YOUR RAM IS BAD YOU MAY KILL YOUR POOL!

If you want to take offense that I'm yelling at you.. good! A scrub is the last thing you want to do right now.

 

[PierceIt]

Thanks for the help here - much appreciated. Here's the answers to your questions:

• Yah, the time drift issue seemed to have gone away after i updated the BIOS firmware to the latest version. Never saw the issue again after that but found some related time drift issues on other forums about running Linux and other unix variants on the version of firmware I was on previously (and noting that the new firmware does not have the issue) and since I have found this to be true so far too I believe this one is closed out.
• the exiting on signal 10 also went away after a reboot - root cause unknown
• and this recent issue about the UNKOWN status of the ZFS volume and inability to run a spool status check seemed to resolve with a reboot as well - root cause unknown

My hardware is:

• Dell XPS 420 desktop from 2008 (new motherboard, cpu, and ram - not refurbished, but new)
• CPU - 2.4GHz Intel Core 2 Quad Q6600
• RAM - 8 GB RAM (non-ECC however) (qty 4 sticks) 2GB Single DDR2 667MHz (PC2-5300) CL5 Unbuffered UDIMM 240-Pin Desktop Memory Module CT25664AA667
• (qty 5) 1TB SATA drives for total of 5TB storage on volume1 configured in RaidZ1 (1 is a 3 year old seagate, 2 are 2 year old WD blue drives, and 2 are brand new WD green drives)
• (qty 1) 1.5TB external eSATA fusion drive on backup volume configured in UFS (it's new)
• The USB stick is a new 32GB SanDisk stick that I got from Radio Shack on a Black Friday Door Buster special for $4.00 (usually $54.00)

I will download and try the memtest software you provided in the link above and will report back.

 

[PierceIt]

cyberjock - no offense taken - i want it fixed and do not want to lose my pool and thank you for making it clear and catching me before i ran one!

I'll avoid the SCRUB for now

 

[cyberjock]

yeah.. memtest.org is da shiznit. Just boot from the CD/USB and let it run. No settings needed. 3 passes is generally considered "good". If anything fails you'll see nasty errors at the bottom of your screen. If you do, I might want to buy that stick from you as I need a bad RAM stick for testing.

 

[warri]

If you do, I might want to buy that stick from you as I need a bad RAM stick for testing.

What are you up to? ;) Looking forward to your results!
I might have another faulty 2GB DDR2 stick if I didn't throw it away. But if it's still there it is located in Europe and I won't be able to access it until March.

 

[cyberjock]

Yeah, it has to be DDR3.

I don't give away my secrets... ;)

 

[warri]

Alright, I was just suggesting because PierceIt also listed DDR2 - not many people use them nowadays ;)

 

[cyberjock]

Alright, I was just suggesting because PierceIt also listed DDR2 - not many people use them nowadays ;)

Doh. I should have noticed that. LOL

 

[PierceIt]

Ran 5 passes with the memtest and zero errors

I think I'll try and replace the USB stick next

 

[cyberjock]

Ok, so you probably have a bad USB stick. Get a new one, install FreeNAS on it and upload your config file. You should be back in business after that. If the issues continue it's possible your USB controller is broken or you have a motherboard issue.

And it is now safe to scrub your pool, but I'd wait until you think this issue is resolved for your own sanity. ;)

 

[PierceIt]

so I moved the USB stick from a USB port on the USB bus on the back of the desktop to a port on the USB bus on the front of the system and have not had the issue in a couple days since. Still watching.

 

[PierceIt]

also - I ran a Scrub - all seems we'll and fine (so far)

scrub repaired 0 in 1h59m with 0 errors
errors: No known data errors

 

[Yatti420]

How about power supply / capacitors? Did you test the USB aswell? I usually stress test mine for awhile do a few passes.. Theres many windows tools out there for it..