Page 1 of 1

Random freeze on the RevPi core 3.

Posted: 06 Mar 2020, 05:17
by login721
Hello.
We have 5 RevPi core 3, all running the same software(deploy from the same .img file then run the factory-reset script) and connected to a DIO module. 2 of them are in testing environment, 3 are in production. Only one of 3 production RevPi get freeze. We get them running 24/7 from December last year, the Pi get freeze 3 times. No ping response, both RevPi core and DIO power led are solid red.
Because of the software is almost identical, is there any chance that the hardware defect causes the problem?
Thank in advance!


This is detail about out project.
Application

Code: Select all

Monitor inputs with DIO and send HTTP request to REST server. The IO monitor is backed by node-red. Nore-red is also monitoring by pm2 and will be restarted if something bad happened.
<br />

Hardware config

Code: Select all

Each RevPi is powered by 24v 60w PSU. 
The PSU also backed by UPS.
1 USB port is connect to the UPS to collect the power signal report.
<br />

Software:

Code: Select all

Based on the offical revPi image
Linux RevPi18033 4.9.76-rt60-v7+ #1 SMP PREEMPT RT Wed, 04 Sep 2019 16:19:42 +0200 armv7l GNU/Linux
Removed java to get some space
Logs are config to write to tmpfs.
Installed node-red.
Installed NUT to safe shutdown the Pi incase of blackout.
Cron job to backup the log everyday and cleanup the logs when the size became too big.
<br />

Logs

Code: Select all

Unfortunately, there are no logs. The pi get freeze and all the logs in ram get erased when the Pi restarted. The backup logs have no information about the crash.
<br />

Re: Random freeze on the RevPi core 3.

Posted: 06 Mar 2020, 09:43
by pi_admin

Re: Random freeze on the RevPi core 3.

Posted: 06 Mar 2020, 10:01
by login721
Thank for reply.
I read that and installed the update on one of the development Pi. But I wonder why only 1 of the 5 RevPis has the problem while they are running the same software.
The in production RevPis are installed in factory without the internet access(we are not allowed to), so it will take a while until we can update the kernel. Plus, if it's the hardware problem, replace the whole unit and DIO will be a safer choice.