I have a bunch of CentOS servers with proprietary applications running on them which are part of a third-party security camera system supplied by an external vendor (i.e. I don't have access to the source for the applications). This includes a number of system services, most of which are managed by
systemd, that all must run as
root, along with a couple "binary blob" proprietary kernel drivers.
On two of these systems, we have encountered an issue where the system will, on boot, restart between 2 and 10 times before it finally boots and stays booted. I've gone through
journalctl, etc. Nothing sticks out (i.e. I don't see any mention of kernel panics, driver failures, etc.), and the only logging I see is a mention that the system went down. I also inspected via
last reboot and the only info I can glean from it is that
root rebooted the system, nothing more. I'm certain at this point, that the reboots are being caused by some process requesting a restart, or maybe a buggy
Is there a way I can modify/configure these systems so that I can track down which process is causing the reboot (i.e. is it a kernel driver that requested the reboot, or was it one of these system services, or was it some ACPI-related issue with power or thermal management, etc.)? For all intents and purposes, it appears that something (i.e. a process/service, not an interactive user) is causing these shutdowns, and it's very frustrating having to wait 10-20 minutes for the system to auto-reboot 10 times before finally allowing me to log in to it.