Cloudron crashing every ~30 minutes on new install

[ianhyzy]

I'm restoring Cloudron from a DO droplet to a dedicated server. Setup runs mostly as expected, but about 30 minutes after setup, everything stops working, including SSH (it just times out). Rebooting temporarily fixes the issue but it returns again in about the same timeframe. The server host said they didn't see any issues, even after running a stress test for about half an hour.

[girish]

@ianhyzy You can try this one by one:

1. systemctl stop docker . Does this make the server unresponsive after 30 minutes?

2. systemctl stop box docker . Does this also make server unresponsive after 30 minutes?

I can help debug further, one you tell me if 1 or 2 is causing issues.

[absnt_mind]

@ianhyzy Killed my entire dedicated server at OVH about 10 times today. Poor server techs had to go manually start it a bunch of times.

I ended up just cutting a partition and deploying a VPS on 18.04.5 and moved it all to there, which seems fine with the Cloudron updates that I installed last night, which didn't seem so fine on 20.04 x64.

Tech said that the screen was black & wouldn't respond to keyboard input, had to hard re/boot several times. But I haven't looked at any logs yet to see what was happening. So that's all I got for now.

[ianhyzy]

@girish

systemctl stop docker did not stop the problem. I rebooted and tried systemctl stop box docker. Both reproduced the issue. It doesn't happen before cloudron is installed.

[p44]

@ianhyzy Dedicated server is always hosted on DO, or other provider? I'm asking this because I'm having similar problem on a bare metal server. Cannot reproduce exactly the issue but sometimes the server got stuck, cannot reach Ssh, cannot ping, nothing, like frozen... Eg. today I had to hard reset four times.

[p44]

@absnt_mind Me to, I've to kill 4 times today... Really strange that happens to all together, or not?

[ianhyzy]

@p44 I switched from using DO, which did not have this issue, to a dedicated server which does.

[p44]

@ianhyzy Ok. All clear...

[girish]

I think this might be related to the gnome/hibernation issue - https://forum.cloudron.io/topic/4119/gnome-shell-installed-in-cloudron-6 . I have fixed this in 6.1 which is due later today. So, hopefully after an update, we can rule out that this is something sleep/hibernation related.

[p44]

@girish Thank's Girish, do you know when will be released 6.1?

[nebulon]

@p44 we are just finishing up final testing, so later today or tomorrow.

[p44]

@nebulon

[ianhyzy]

@p44 If you're in a hurry, the trick in the linked forum post worked beautifully for me.

[p44]

@ianhyzy said in Cloudron crashing every ~30 minutes on new install:

@p44 If you're in a hurry, the trick in the linked forum post worked beautifully for me.

Thank's @ianhyzy , do you refer to this step?

On existing installs, you can simply run apt remove -y gnome-shell && apt -y autoremove to remove all the xorg and gnome shell stuff. You also have to reboot the server so that services like Xorg are stopped (alternately, you have to stop those manually).

[ianhyzy]

@p44 ah sorry no, this one: apt install libnotify4 --no-install-recommends

[p44]

@ianhyzy Thank's a lot! For now seems to working everything. If problem occurs again, I'll run that command. Otherwise I'll wait for new Cloudron release.

Many thank's!

[p44]

@girish said in Cloudron crashing every ~30 minutes on new install:

I think this might be related to the gnome/hibernation issue - https://forum.cloudron.io/topic/4119/gnome-shell-installed-in-cloudron-6 . I have fixed this in 6.1 which is due later today. So, hopefully after an update, we can rule out that this is something sleep/hibernation related.

This fix will be applied after update? Reboot needed?

[girish]

@p44 said in Cloudron crashing every ~30 minutes on new install:

This fix will be applied after update? Reboot needed?

Yes, reboot the server after applying the update since it will stop all X server etc (if it was running previously).

[p44]

@girish Thank's a lot!