Hello,

I am relatively new to UnRaid (2 weeks tomorrow), but not to Linux or computers in general.  I work in IT, so I am usually fairly self sufficient, with excellent google-fu.

 

I have got all my media and other data moved, and the apps which were running on my old server migrated to dockers (and one VM) on UnRaid.

 

Other than some weirdness moving/copying files from qBittorrent when I enabled cached writes, it's been pretty smooth sailing, with any issues I have had yielding quickly to the aforementioned google-fu. 

 

BUT,  I was thinking about going from 'disks always spinning' to 'spin those disks down as much as you can.'  Every time I look at the Main tab, disk1 has write activity.  In the last 7 minutes, there have been 5500 writes to disk1, and zero to any other array disk (other than parity of course).  I thought I had found the culprit earlier this morning.  When I set up the system, I did not start with a cache disk, so my docker.img and some other libvirt stuff that should have been in cache were on disk1.  I stopped the VM and Dockers, then manually moved the files to cache using the file manager.  I have restarted everything now, and still I get writes to disk1.  I can't seem to nail down what process is writing, nor what file is is writing to.  I have used the open files plugin, lsof and iotop, and none of those has nailed down any process.  I ran a find -mmin -60 (which finds files with an mtime in the previous 60 minutes) on all the folders in disk1, and though I found a few files dated in the far future (2049, 2097, etc) and used touch to reset their mtimes, nothing else popped up. 

 

So I have a mystery process writing to a mystery file.  The disk1 writes seem more frequent when everything is running (dockers and VM), but it still happens while they are all stopped, just much less frequently.  As I am posting this here, I hear my old Senior Engineer from many years ago in my head, extollign the virtues of the WIDR rule (when in doubt, reboot).  I am about to do that, and if there is an improvement in the situation afterwards, I will post back here.

 

And here, I stop to raise a glass to Mr. M. Munford and his WIDR rule.  I guess something was lingering in the background, or maybe because the server was in GUI mode, but after a normal reboot, I have been watching the writes, and after some initial noise, all seems clear.  One of the dockers (Sonarr i think) seems to be scanning folders, because I see a bunch of reads right now.  I'm sure that will calm down once whichever docker is feeling insecure after the reboot decides everythign is really OK after all.

Sorry for the false alarm, but maybe someone else will find this sometime and it will help them.