Mox reboots randomly after TOS 6.0 upgrade

I have the Same Problem is there any solution for it. After I go back to a pre 6.0 snapshot I’ve to flash my MOX. where can I download the 5.4.5 Version of Turris OS?

After flashing with 6.0.1 I didn’t install any aditional package and the Issue still exists.

I uploaded my log-file to GDrive: srv/log/messages.

2 Likes

Same here. Reboots randomly. :frowning: On the other hand soft reboot finally works. :slight_smile:

1 Like

Ok, I can’t take it anymore… Will rollback the upgrade and disable automatic updates for the time being.

2 Likes

Thanks, I might try that. It turns out that it’s rebooting roughly every 2 to 3 hours (last time just 15 minutes ago), which gets to be a pain if I need the office VPN.

Any sign of rescue by The Management?

Not that I am aware of, unfortunately. Except the gitlab issue mentioned at the very beginning of this thread.

I’ve tried to ask in another thread (Turris Mox USB (from module A) not working any more (for mass storage) after TOS 6.0 upgrade - #12 by Mossroy), but got reprimanded about cross-reference.

I understand the situation is probably complicated for Turris staff, I’m not trying to blame them.

However, I’m surprised that these Mox issues have not been mentioned in the “aftermath” thread: Turris OS 6.0 aftermath

Like some other ones, I did a rollback on this upgrade, and disabled automatic upgrades until there is a more sustainable solution. See Turris OS 6.0 is released - #143 by Mossroy

2 Likes

Thanks for the roll-back instructions. How long did you wait before rebooting?

I am 1 of the early backers of the indiegogo campaign and ever since I got the mox, every time it does a soft reboot after an update it hangs. I normally worked around this by doing a power cycle.

However, since the latest update my router gets rebooted randomly and because of the hanging problem on reboot, it can’t boot up properly, forcing me to power cycle the thing constantly. I have since rolled back to TurrisOS 5.4.4 or 5.4.5, not sure if it would solve the problem.

If I get a brand new module A, will I still be experiencing the hang on soft reboot problem? Does this problem still persist?

My configuration is AGBEE with the PoE and 1GB RAM upgrade together with the miniPCIe wifi card

Maybe one minute.

It should be very fast, as I suppose it’s mostly switching to a btrfs snapshot

Thanks that all worked.

At least rebooting now without unplugging, it is progress.

1 Like

Has anyone, who had the random reboot isue, tried out 6.0.2?
Is there still the same problem?

I followed the various links on the bug reports and it seems to be a kernel bug/regression, with the chip manufacturer very slow to respond. So my understanding is that it is a bit out of Turris team’s control and it might take time.
I am also a MOX owner and luckily (for me) I am stuck on a development version of TOS 6 that does not suffer from this.

2 Likes

Still rebooting…​​​​

1 Like

Hello,

we can acknowledge that the random reboots do happen on some MOX devices. We have a fix for that, which was not applied in the Turris OS 6.0.2 version, though. It should be released together with the Turris OS 6.0.3 version.

The fix is easy: you can run the following command:

`echo "performance" > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor`

Unfortunately, it remains only until the next reboot. After the next start of the router, you will need to apply it again until the fix is released in Turris OS.

Please, accept our apology for the inconvenience.

9 Likes

Thank you very much for this reply

What is the effect on the power consumption? Did it go up when compared to old 4.14 kernel on TOS 5.4?

Is that a fix or rather a workaround? Does it limit some functionality?

Looks like you can find full the description of the issue here

If I’m not misstaken it is a fix as the new kernel changed the default governor

EDIT: humm, lookslike my MOX with 5.4.3 have ondemand governer. So as for the delfault change Idon’t know.
Anyway, moxes CPU don’t consume a lot of power so it is probably negligible as mentioned in the gitlab issue.

1 Like

why not add this to /etc/rc.local to be executed on each boot ?