Strange update issue on Supermicro C2758 / 22.1.7.x to 22.1.8_amd64

[simu]

Hi

Just to let you/someone know: had the strangest issue in update from 22.1.7(_1?) to 22.1.8_1 (current as of this morning).
Saved config (ful), did the update, updater going into reboot;
In boot process, this showed for an extended period of time on the console (see attachment, don't know how to insert image here, darn complicated - UX... whaaa).

Some pings to FW-LAN interface and WAN (googles DNS) went through, without a change on display of attached screenshot.
Then, without any further ado, no warning, no message or whatsoever, the system turns itself completely off.
Turning on the sytem again manualy, same thing (boot, attached screenshot, some pings, off)
Again, over and over. Did the whole thing maybe 10 times, had the IPMI-Function in suspect for the "poweroff". No, i did not connect IPMI to WAN or any such rubbish...

So, did a download of 22.1.2, install, update, config-restore - and am up and running again.
Not had something like that the past times...

Is this a sign of faulty drive or alike? because, SMART says, that the disks seem fine... (SSD)...?

br. D

[Cerberus]

I dont think it is the drive. On FreeBSD the system normaly goes really verbose (cam errors) on your terminal screen on a drive fault.

There are some reports about issues after people updated to 22.1.8, aliases not working anymore, rules has to be saved again and stuff like that. I updated three KVM to 22.1.8 and had no issues, but these are the more simple setups. Most of my production / heavy duty OPnsense installation run on the business edition and intentionally lags behind the minor releases.

If this happend again, try to logon locally and check if your installation is still alive.

[simu]

Had it again, this morning at around 3.45ish am. And first thing i did was a complete reinstall, did not take that long and was up and running shortly after.

So, this afternoon I had to move that device phisicaly in another rack. Did a proper shutdown, moved it, the usual stuff, power on - same thing as in screenshot.
As I had time, I tried to do a debug, as I did not think, that either the hardware nor the software (OPNs) as such are the trigger.

Checked the CL, there was a Bios-Update somewhat recently. So, checked the Bios. WatchDog was active, so I de-activated. Booted and it ran - for almost 1 and a half minutes, got the console pluged and saw, that it was a seemingly somewhat proper shutdown (like no sudden poweroff).

So, back to bios, checked with a TK-Wiki-Article: PCIe AER Support was enabled - as this might lead to a problem, I disabled it.

Am online now for an hour-ish ...

Keeping an eye on this and will go now for the logs.

[simu]

Think I got it:

Multiple-non-related: Bios stuff from above and then UPS-Power-Issues this afternoon.

So, there is a UPS, that is connected to a Syno-NAS (USB). This NAS provides Net-UPS-Infos to 2 other NAS - and the Opnsense! And it seems, that it was flapping somewhat around that time (effectively, at the time of reboot, a proper reboot). Found some respective log entries:

Code: [Select]

UPS ups@xxx.xxx.xxx.xxx:3493: forced shutdown in progress

Code: [Select]

Executing automatic power-fail shutdown
So, deactivated NUT for instance and will search on, if there is another cause... (UPS has oodles of runtime, and even if, the device is not exactely business critical)...

I'll keep you posted if...