my Vero2 is rebooting every few minutes (so interrupting anything else I’m watching because of the CEC).
This is happening since yesterday morning there was a power outage (I was at my office) and have found the battery of the UPS flat and had to reset the whole shebang.
So I’m kind of suspecting a physical power problem in either the UPS or the Vero2 power supply.
Is there anything I can explore on the Vero2 logs to try and pinpoint the problem ?
Right now only solution was to physically unpower the unit (because the sudo shutdown -h 0 or doing a power off from the menu leads to a reboot and not to the desired shutdown).
Thank a lot.
How often does the device reboot?
Does it get to Kodi before it reboots?
That was a bug in the July version of OSMC, it should be fixed in the August update.
Are you able to log in via SSH for long enough to get some logs?
The easiest step might be to power it directly for a while and see if it’s okay.
You bright kid Yes correct good idea. First try will be powering with current power unit bypassing the UPS.
Will also try uploading the logs.
Will update the thread accordingly.
Would/does OSMC support monitoring an APS UPC to trigger a programmatic clean shutdown in case of an extended power outage ?
Thank you for now.
It would be up to APS to provide this solution.
Powering directly from outlet.
Installed 28 August update.
Here are the logs.
BTW this is not new I must search the site but the screen size is not reported correctly and need to find a way to force it to 1080p
Hi @ActionA thanks for chiming in. On other Linux systems I had, there was a little daemon monitoring the APC UPS data port and you could use its signals to act upon them for example triggering a shutdown when the battery level or time limits you had set expired.
Hope you can spot something.
Thanks in advance.
That is something you would have to seek support from APC for. Otherwise, any debian solution should work.
16:48:05 T:3031232512 NOTICE: application stopped...
16:48:05 T:3031232512 DEBUG: ActiveAE DSP - destroyed
16:48:05 T:3031232512 DEBUG: PVRManager - destroyed
16:48:05 T:3031232512 DEBUG: SECTION:UnloadDll(libcurl.so.4)
16:48:05 T:3031232512 DEBUG: LogindUPowerSyscall - delay lock released
LoginDUPowerSyscall means someone one the system has requested a reboot. That’s why the device has rebooted there. It’s not watchdog (hardware related) or Kodi wouldn’t be aware of the reboot request.
See Do not use "idle mode" in my Vero2 - #72 by sam_nazarko
sorry but I’m not sure I understand.
Vero2 does not run nor has anything installed other than OSMC.
Of course the logs have been generated after myself rebooting the system via KODI’s menu and the timestamp is roughly when I did initiate the shutdown manually.
Could that LoginDUPowerSyscall be a product of my shutdown or of some sort of reboot request after an update ?
Thanks for clarifying if you can.
The shutdown issue which resulted in a reboot has been fixed in this month’s update. A shutdown request will also invoke ‘LogindUPowerSyscall’, as Kodi makes known the user request to power down or reboot the system via systemd’s logind interface.
All your log shows me is that a power cycle / or shutdown was requested, and the system turned off.
You can try a clean install to rule out a problem with the installation.
my bad maybe have misled you with a wrong description.
The logs have been produced only with my manual reboot, NOT after a spontaneous one, which at the moment has not happened.
So my plan is to keep Vero2 on the power outlet for a few hours and catch a spontaneous reboot if it occurs.
If it does not, I’ll shutdown manually, move the power source to the UPS outlet and test it again.
If a spontaneous reboot does occur will upload the logs again in this thread.
Ok here we are
This is the log uploaded just a few seconds after a SPONTANOUS reboot which happened roughly at my 17:45 (local time 15:45 GMT) with Kodi’s homepage appearing at roughly 17:46
Vero2 was powered by the original PSU from the outlet wall so I’d surmise the APC UPS is NOT the culprit.
Thanks for taking a little more time investigating.
Very quick glance shows:
Sep 02 17:34:22 osmc kernel: [<c008d8bc>] (oom_kill_process+0x398/0x3ec) from [<c008ddc0>] (out_of_memory+0x2ac/0x2dc)
Sep 02 17:34:22 osmc kernel: [<c008ddc0>] (out_of_memory+0x2ac/0x2dc) from [<c0091cb0>] (__alloc_pages_nodemask+0x888/0x8a0)
Are you running a lot of services on this device? You could try disabling them to see if this helps
Not that I know
Maybe just reflashing the card and reconfiguring my Samba source, sound settings and stuff would be the best way ?
The installer point to a July installation and to an Android one … which one is recommended ?
What is leaving me puzzled is that until this morning’s power outage poor Vero2 was a faithful dependable friend
Android isn’t designed to replace OSMC, so you probably don’t want that.
Try disabling TVHeadend.
The real source of system’s complexity is USERS
I would not have remembered I installed and enabled TVHeadEnd a few days back, so when I stated “the system behaved well until the power fault” I was not complete nor totally correct !
I have uninstalled TvHeadEnd from OSMC’s app store and will see if that fixes the glitch.
Thanks a bunch.
PS BLUSH … where is the SD card in Vero2 ??? I can’t seem to see a slot from which I could have inserted a new SD card. Is it inside ? In case I’d need it in the future a Sandisk 8GB SD (not microSD) would be a wise choice ?
The internal storage in the Vero 2 is a NAND chip. You can’t remove it, it’s soldered to the board.
It has an SD card slot on the side, which you can use to reflash the internal storage, and boot Android off an SD. You can put a full sized SD card in the side.
So if I ever need to “reflash it” I’ll just prepare an SD with the card prepared by the installer, stick it in there and let it do its magic ?
Thank you so much Sam. Have a good day for now.
Did you in the meantime updated to august version to ensure you have the fix for shutdown bug installed?