OSMC freeze after a few hours

Hi all,

I just installed OSMC, last stable version, on an all new Raspberry Pi 2.
The only few things I have done so far are:

  • active the root account

  • modify config files for my IR remote,

  • add my NAS NFS shares in fstab:

    192.168.1.10:/volume1/video /mnt/video nfs noauto,x-systemd.automount

So far everything works very well, but the system hangs after a few hours systematically. It does not respond to keyboard, and I can not connect to the Raspberry via SSH anymore.

The problem is, I don’t know how to find what is responsible for these freezes. Could you please indicate me where I can found any hints on what’s happening? I haven’t see nothing specific in /var/log, but I am probably missing something.

Logs are available here

Thank you very much!

Again, the machine is freezed this morning. I really have no idea on what’s happening…

Are you using any overclock ? If so change it back to normal.

What power adaptor are you using ? Have you tried a different one ?

What USB peripherals are connected ?

A hard freeze including not being able to ping the device is almost always either excessive overclock or a lack of power.

No overclock at all. I’m using a 5V/2A power adapter from Samsung.
A USB dongle for keyboard and an IR receptor are both connected in USB.
And that’s it!

I guess I might have to test another power adapter…

If that’s a phone charger I would recommend a proper adaptor designed for the Pi like the official Pi foundation one. Phone chargers are notoriously bad at powering Raspberry Pi’s reliably.

Yep, it’s a phone charger. I just ordered the “officiel” power adapter to replace it.
In the meantime, I will test if the device still hangs without anything connected to it (no USB device, and no NFS shares).
There is really nothing in the logs…

If there’s nothing at all in the logs and it’s a total freeze (both Kodi gui and network connection dies) then it will be hardware related not software, so hopefully the power adaptor solves it for you.

For what it’s worth I get the exact same symptoms on my Pi 2 if I set the overclock to Turbo (it randomly hangs after anywhere from 30 minutes to a few hours) but it runs perfectly stable at Normal. As with you, nothing gets logged in that case.

The system still freezes when nothing is connected to. I’ve just changed the power adapter with the one coming with my iPad, and the system hangs as soon as I try an ssh!
I guess I will have to wait from my new real power adapter …

iPad chargers are known not to work on Raspberry Pi’s - they only put out about 4.3 volts under load and the Pi must have 4.75 > 5.25 volts at all times not to crash.

My RPi2 does exactly the same with the proper Pi power supply. Mine freezes every other day. I just assumed it was down to my iffy internet connection as we had problems with the net for a fortnight due to a recent storm we had. Now my internet connection is sorted, it’s been fine so far (2 days)

I have nothing connected to USB, use the proper Pi PSU and it’s just used to stream from a NAS box. hopefully it doesn’t freeze again but thought I’d add my experience to your post so you know you’re not the only one :wink:

Well, I just received the new ‘original’ Pi PSU. I have connected it my RPi2, but I don’t know if it solves my problem yet. I’ll come back in a few days to confirm it works.

Well, I was just trying to configure a VPN connection via connman through SSH when the system hanged … again. So this was not PSU related.
Once again, I don’t know exactly how to debug that. /var/log is populated with very few log files.

Systemd systems don’t use conventional plain text log files in /var/log - you use the journalctl command to view system logs.

I’m definitely not aware of all the novelties coming with Debian 8. Connman, init and log systems are very different from what I know for years now.
I will take some time to understand how logs work now and try to track what is exactly happening when the system stops responding. Hope it’s a software issue, and not an hardware one. I’ve already done a memtester check and a full clean reinstall of OSMC after a pass of 0 on my SD card, but nothing new so far.

I have actually a similar problem with crashing. Everything seems fine and works well but after an hour or two of no interaction the system crashes. Sometimes with a obvious crash screen with bugged out colored lines and sometimes with a black screen. And only solutions is to pull the power cord.

Well, again, my RPi2 is freezed this morning. I really don’t see anything in the log, again. Well, something is stranged actually, but this might be something related to the way systemd captures the log:

#  journalctl --list-boots
-2 71c76ad5bcc7457abcf3c8a8d4a7ef71 lun. 2015-07-20 09:17:48 CEST—lun. 2015-07-20 20:03:15 CEST
-1 446d0a2fa58c409e9be1d6d0ba65a000 lun. 2015-07-20 20:03:08 CEST—lun. 2015-07-20 23:35:37 CEST
 0 00a0f2af31a64f1e8fee605d14ef0fe7 lun. 2015-07-20 20:03:08 CEST—mar. 2015-07-21 09:45:23 CEST

It seems that a reboot happened yesterday, 23:35:37 CEST. But I have not rebooted the system … And stranger are the logs themselves:

# journalctl --since yesterday
[...]
juil. 20 23:35:36 osmc sshd[973]: pam_unix(sshd:session): session closed for user osmc
juil. 20 23:35:36 osmc systemd-logind[264]: Removed session c3.
juil. 20 23:35:37 osmc sshd[963]: Received disconnect from 192.168.1.20: 11: disconnected by user
juil. 20 23:35:37 osmc sshd[958]: pam_unix(sshd:session): session closed for user osmc
juil. 20 23:35:37 osmc systemd-logind[264]: Removed session c2.
-- Reboot --
juil. 20 20:03:08 osmc systemd-journal[110]: Runtime journal is using 4.5M (max allowed 36.7M, trying to leave 55.1M free of 362.8M available → current limit 36.7M).
juil. 20 20:03:08 osmc systemd-journal[110]: Runtime journal is using 4.5M (max allowed 36.7M, trying to leave 55.1M free of 362.8M available → current limit 36.7M).
juil. 20 20:03:08 osmc kernel: Booting Linux on physical CPU 0xf00
juil. 20 20:03:08 osmc kernel: Initializing cgroup subsys cpu
juil. 20 20:03:08 osmc kernel: Initializing cgroup subsys cpuacct
[...]
juil. 20 20:03:15 osmc connmand[377]: eth0 {del} route 178.62.84.106 gw 192.168.1.1 scope 0 <UNIVERSE>
juil. 20 20:03:15 osmc kernel: Console: switching to colour dummy device 80x30
-- Reboot --
juil. 20 23:35:37 osmc systemd[960]: Stopping Default.
juil. 20 23:35:37 osmc systemd[960]: Stopped target Default.
juil. 20 23:35:37 osmc systemd[960]: Stopping Basic System.
juil. 20 23:35:37 osmc systemd[960]: Stopped target Basic System.
juil. 20 23:35:37 osmc systemd[960]: Stopping Paths.
juil. 20 23:35:37 osmc systemd[960]: Stopped target Paths.
juil. 20 23:35:37 osmc systemd[960]: Stopping Timers.
juil. 20 23:35:37 osmc systemd[960]: Stopped target Timers.
juil. 20 23:35:37 osmc systemd[960]: Stopping Sockets.
juil. 20 23:35:37 osmc systemd[960]: Stopped target Sockets.
juil. 20 23:35:37 osmc systemd[960]: Starting Shutdown.
juil. 20 23:35:37 osmc systemd[960]: Reached target Shutdown.
juil. 20 23:35:37 osmc systemd[960]: Starting Exit the Session...
juil. 20 23:35:37 osmc systemd[960]: Received SIGRTMIN+24 from PID 6402 (kill).
juil. 20 23:35:37 osmc systemd[961]: pam_unix(systemd-user:session): session closed for user osmc
-- Reboot --
juil. 21 09:45:00 osmc http-time[395]: Updated time from Mon Jul 20 18:03:17 UTC 2015 to Tue Jul 21 07:45:00 UTC 2015 using HTTP query to www.google.com
juil. 21 09:45:00 osmc ntpd[521]: ntpd 4.2.6p5@1.2349-o Fri Apr 10 19:31:04 UTC 2015 (1)
juil. 21 09:45:00 osmc ntpd[532]: proto: precision = 0.833 usec

OSMC got back in time, from 23:35 to 20:03?

Just for information: as long as the system is busy (for me, connected to a VPN server with some file sharing), everything is OK and perfectly stable. But if the RBPi2 does “nothing” special, it freezes.
I suspect some kind of bug in the idle phase of the system, but have no idea on how to track it.

I had to try 3 different chargers, before I got a stable Pi.The original charger is not sufficient enough.
You need minimum a 5V 2.1A charger. The USB in my TV does the job, too.

But more ampere is better, especially if you are running Aeon 6 or similar heavy duty skins or if you are overclocking.

This is rare and never recommended.

I´d say, any TV USB-output will outperform the “original” 1A charger option in the stores.