[SOLVED] Firewall hanging every day -- turnstile_broadcast() crash

Started by Taomyn, February 02, 2017, 07:26:16 PM

Previous topic - Next topic
The reporter doesn't report any errors after the reboot.
I haven't had the possibility to check the console, but for the last three days it was stable.
Updated to 17.1.1 yesterday too.

I also switched the console to 'vt', might this be a workaround/fix for the hang?

Quote from: abraxxa on February 10, 2017, 08:44:59 PM
I also switched the console to 'vt', might this be a workaround/fix for the hang?


No, at least it wasn't for me - since updating to 17.1.1 I'm testing mine without a scheduled reboot to see if it's fixed. 1day 4hrs so far....

Well so much for that test - it just hung again so will re-enable the daily reboots.


Report submitted as always.

Same issue on my box. lexcom 3i270d
Installed fresh one week ago.

Died always after one day.

Intel(R) Atom(TM) CPU N270 @ 1.60GHz
2 CPUs: 1 package(s) x 1 core(s) x 2 HTT thread. 1GBRam
NanoBSD form 4GB-CF


Quote from: mistergefahrensucher on February 13, 2017, 07:35:21 PM
Died always after one day.


Did you submit a report through the firewall? The "Report" page is a little mis-leading as it will look like there is nothing to report, but hit the button anyway as it will say there is and fill in the details.


The more info we send in, hopefully the sooner we can get this fixed.

Sorry can't provide you with chrashlogs. because the firewall will not boot anymore.
So i switched back to my old CF-Card with a booting pfsense 2.3.1
At the moment i have not the time for investigating.
My intension was to go to opnsense because pfsense will not support i386 and NANOBSD in the comming release.
Have seen some other posts where also users have problems with BSD 11.0 on Atom Hardware.
At the moment i will stay on pfsense 2.3.1 with BSD 10.3.



So far not much luck here. It's a corruption of the VM lists somehow, a locked object returns without a lock. The code does not expect this to happen at all and on 10.3 apparently it did not.

It's somewhere in the amd64 code, so maybe i386 has better luck running smoothly. Can somebody confirm/deny?

Regarding the crashes in turnstile_broadcast(), a FreeBSD developer provided a test kernel:

# opnsense-update -kr 17.1.1-rwdebug
# /usr/local/etc/rc.reboot

It still panics, but will print vital debug information when doing so.

Running it and providing the output will hopefully help resolve this.


Thank you all,
Franco

I missed something.   How is this solved?  I'm hitting this exact issue.



The issue is fixed in 17.1.2 and FreeBSD is still asking for users to reproduce this with the debug kernel posted above in order to get it fixed there... I would just like to pass this along, but I can understand our users' stance as well.


Cheers,
Franco