I am experiencing an analogous issue with a RevPi core 3 I received a month ago.
At times it shuts down by itself and all LEDs (revpi and attached modules) are red, so in shutdown/fault mode I guess.
There was no custom software running at fault time the first time the issue appeared and the revpi was freshly unpacked and all updates installed.
Everything had been working with custom software on a RevPi core 1 for months before without a single crash, same power supply and IO modules.
I have noticed that a non-closed remote ssh login seems to cause more frequent crashes.
After the the system was correctly powered through a UPS and earthed the issue appeared again, a normal raspberrypi connected on the same line did not report any shutdown events or faults, so the main line (UPS) is stable.
This time an additional program was running but was constantly profiled and the logs are written to a mounted external USB flash device. The CPU has an average of 2.5% usage per core and RAM is available.
There were no hardware changes, I just left some screen sessions open for testing.
I have analysed the kern.log, auth.log, checked the last logins and nothing strange shows up, I still have them in case.
As said a RevPi core 1 with the same Jessy image and updates had not had any issues with modules attached.
This last crash was on 15th March at about 8:30am guessing by the missing cron events in auth.log. The system has been running since and I am constantly monitoring it.
One suspicious entry was this:
Code: Select all
Mar 11 22:06:29 RevPi4858 kernel: [123702.816477] piControl: too many communication errors -> set inputs to default 0 2 2 3 3 3 0 0
But it happened to appear days before the last crash. The annoying part is that this has happened 4 times now and there is no obvious error source.
Any suggestions as to why the system shuts down under normal conditions? all status LEDs of the RevPi core 3 and of the connected modules are red, as if it had been shut down via software, a full power loss and a reboot would have restarted the system correctly.
Could it be the 24V power source? But there is a UPS and the other pieces of hardware don't go down and it has been running for 6 days now (sign of power stability).
Enrico