WebUI becomes more and more unresponsive on two servers

[bobobeastie]

I have two servers, both 1st gen Zen based, one threadripper and one Ryzen.  The first sign seems to be a loading animation on unassigned devices in the Main page that never finishes loading.  At this point I am unable to get diagnostic files either from the gui, or from putty.  With putty, it just hangs at "starting diagnostics collection...".  After that, eventually, none of the pages load, and I am forced to tell the server to "reboot" twice in putty, the first one doesn't do anything.  After the second reboot, which works, and the systems boot, a parity check starts, but is not able to finish before all the issues start, and they seem to stop updating their status at some point near their end.  Sometimes I see after a reboot that the previous parity check had finished successfully.  Also, mover seems to run slowly and sometimes the transfer speeds on the array stop showing non zero values.

 

Originally I was updating to the most current 6.7 RC's, though I think I skipped .2.  But yesterday I downgraded to 6.6.7 by replacing the bz files on the flash drives with the bz files downloaded from the unraid web page, and I'm still having these issues.  I left the other bz files from the RC's that were not replaced with 6.6.7 bz's.

[bobobeastie]

I was able to grab a log file with "cat /var/log/syslog > /boot/log.txt" in putty, which then I found in the flash share over my network, it is attached, I notice connection timeouts at the end. log.txt

[bobobeastie]

So I think it was due to having shares from the opposite server mounted in unassigned devices, and then when that remote server restarted and became unavailable, unassigned devices became unresponsive, which I think must have messed up the whole web interface?  Which caused a very frustrating game alternating server issues.  There's a remote share error in the log file.  I removed the shares and knock on wood, so far so good.  I have the shares mounted so I can run a dupeguru container from unraid to find duplicate files, so if I understand the cause correctly, I would like to find a solution other than just not mounting the shares.