FreeNAS 8.0.2 panic vm_page_insert... Need help!

[HotRod]

[SOLVED] FreeNAS 8.0.2 panic vm_page_insert... Need help!

Hi. I have just recently reinstalled my server with Freenas v. 8.0.2 release and find that the server "crash" every day.

This is the message I get in the console (not a headless system):
panic: vm_page_insert: page already inserted
cpuid = 5
Uptime: 23h51m47s

The server does not respond to querys (webgui/ssh) or console input..
Have to use the powerbutton... Not good...

The system boots up normally afterwards and runs another day, and then crashes/freezes...


Any suggestions on how to fix this? Or at least what it means?

 

[William Grzybowski]

Are you using i386 or amd64? how much of ram, what cpu?

 

[HotRod]

I'm running amd64, 6GB RAM and Intel i7 cpu

 

[William Grzybowski]

Do you happen to have a USB3 port? If so, try to disable it in BIOS, and/or remove xhci_load="YES" from /boot/loader.conf

 

[HotRod]

I will have a look tomorrow when I have hands on the server, but I think that there is only USB3 ports on the mainboard...
I will try out your suggestions, thanks. I'll get back to you as soon as I have tested it :)

 

[HotRod]

How can I edit the loader.conf? It seems to be read-only..
Tried to edit with VI, but failed to do so.

 

[William Grzybowski]

mount -uw /

 

[HotRod]

Ok, thanks.
I switched off USB3 support in bios and commented out the line in loader.conf like this #xhci_load="YES"
Just have to wait and see if it stops freezing up...

 

[HotRod]

This worked!
I'm not sure if it was the disabling of USB3 or the edit of loader.conf that did the trick, but it has now been running stable for several days.

Thanks William

 

[foursky]

Works like a champ

I had this problem come up after running fine for a week on a new config. Disabling USB3 and commenting out xhci_load="YES" from /boot/loader.conf worked like a champ for fixing my vm_page_insert panic problem also.

 

[gcooper]

I had this problem come up after running fine for a week on a new config. Disabling USB3 and commenting out xhci_load="YES" from /boot/loader.conf worked like a champ for fixing my vm_page_insert panic problem also.

USB 3.x support is disabled by default in 8.0.3 based on user feedback.

 

[foursky]

Actually, the problem persisted after this. I replaced the RAM, and suspect that I might have had a memory timing issue. Uptime prior to the RAM swap out was <5 minutes, and now is measured at 3 days, 5 hours. After I disabled USB3, I enjoyed a a few hours worth of stability, and was too quick to post. I quickly went back to 3-5 minutes of uptime. I was waiting for a few days of stability before redacting my comment...

I would comment out the xhci_load line, and remark about the "possible" issue in the release notes. I want to enable USB 3.x after I have a week of stability, as this is important to me in the long run...

Thanks,
Robert

 

[gcooper]

Actually, the problem persisted after this. I replaced the RAM, and suspect that I might have had a memory timing issue. Uptime prior to the RAM swap out was <5 minutes, and now is measured at 3 days, 5 hours. After I disabled USB3, I enjoyed a a few hours worth of stability, and was too quick to post. I quickly went back to 3-5 minutes of uptime. I was waiting for a few days of stability before redacting my comment...

I would comment out the xhci_load line, and remark about the "possible" issue in the release notes. I want to enable USB 3.x after I have a week of stability, as this is important to me in the long run...

All the issues we've seen before with xhci have punted back Trap 12 errors and it's very clear that xhci is the culprit from the kernel panic message.