I recently had a power outage and even though the OS received the shutdown signal from NUT and upon login it warned me that the system will soon be shutdown, it wasn’t. Unfortunately, the system went out when the power went out.
I want to investigate the issue after the fact, but it seems journal/dmesg logs aren’t persisted by default.
Our journalctl, a newer log manager prevalent in modern linux, is as per our upstream of openSUSE. And we inherit their default on this front.
But you may find something in our own logs in:
/opt/rockstor/var/log/
With the main file there being rockstor.log.
I suspect that the battery in the UPS did not last long enough for the graceful shutdown orchestrated by NUT. You could set NUT to have a much smaller grace period so that it lies within the capacity of your UPS batteries. See:
How we use NUT, behind the scenes, is explained here:
Let us know what your investigations show, and the second technical guide references should give you enough pointers to do such things as test shutdowns etc. As you can then see if NUT is doing it’s thing and the time that it allows with your currrent setup: without actually cutting the power to the system. There is a NUT client command line app that we speak to that you can similarly ask to send a ‘fake’ power out signal to the NUT server so that it acts as if the power has been cut to the UPS. This is a good test as you can then make sure a shutdown actually takes place. Likely as you saw a message indicating this, it does then actually take place - as by then it is out of our hand and the system is about to be shut-down but at a pre-determined delay that likely exceeded your battery capacity. Either that or your batteries died unexpectedly.