Shortly after adding a new share last night I discovered that my Unraid server had become completely unreachable on any interface (note, I may have made a mistake trying to connect to the IPMI port so possibly that was still up; it did not show connected in my network overview though. It was late and I was grumpy - gave up). There were a number of unnamed MAC addresses which had evidently been connected over the previous 24hrs including one which was all zeros; very weird.

 

This morning I hauled a monitor and keyboard over and was able to login. Nothing obviously amiss, but upon running ifconfig there was no response and I could not exit either (ctrl-c non-responsive); hard hang. Attempting a controlled shutdown timed out (twice), and the subsequent forced shutdown never happened either. I had to power-cycle the machine.
 

Everything started up normally. I did not have syslog enabled, but I do now. Attached are anonymized diagnostics - I am not yet familiar with troubleshooting issues like this, but hoping something was captured.

 

gondolin-diagnostics-20220429-1253.zip

Edited April 29, 20224 yr by _cjd_
spelling

48 minutes ago, _cjd_ said:

There were a number of unnamd MAC addresses which had evidently been connected over the previous 24hrs

Do you really mean IP addresses? I only noticed private IP addresses in the current syslog. Have you put your server on the internet?

 

Unrelated, your appdata has files on the array.

58 minutes ago, trurl said:

Do you really mean IP addresses? I only noticed private IP addresses in the current syslog. Have you put your server on the internet?

 

Unrelated, your appdata has files on the array.

This is not on the internet (not even any ports opened up at the moment). It is able to reach the internet, of course.

And no, I do not mean IPs; historic devices (e.g. not actively connected) I can only see the associated MAC - however, when I've seen unexpected devices they're mirrored IPs (assumption is docker images using HOST or BRIDGE - I'm not sure why they even show up yet, but that could be related if this was a broadcast storm caused by the way I have Unraid+services configured)

Also, thanks. I thought I fixed use of the array for appdata but evidently not. Where do you see that? Definitely still working to get this optimized.

edit: looks like binhex-krusader, which I haven't even ended up using so I'll just remove that for now - I guess I prefer cmd for stuff where tutorials say to use krusader.

it also looks like mover did not work night before last (nor last night but that's no surprise) - a couple backups still on the cache which should have been moved to the array. That potentially puts this on the rc5 update as far as timing.

Edited April 29, 20224 yr by _cjd_

5 hours ago, _cjd_ said:

Where do you see that?

In diagnostics shares/shareDisks.txt

 

You can see this in webUI by clicking Compute... for the share, or the Compute All button, on the User Shares page.

 

5 hours ago, _cjd_ said:

mover did not work

Mover moves cache-yes shares from cache to array assuming there is space, it ignores cache-no and cache-only. Mover won't move duplicates, and nothing can move open files. Turning on Mover Logging in Scheduler can sometimes help diagnose.