Disks keep spinning up...

[jowi]

Ever since i'm running V6 (for about a month now) i've noticed that sometimes, for no apparant reason, disks (on ore more, or all) keeps spinning up. I did not have this with V5. I'm running 4 dockers, dropbox, sabnzbd, sonarr and squeezeserver, even if i disable them, disks get spinned up sometimes. No, i do not have any telnet sessions or smb/nfs browsers/connections open. No activity at all... but still... disks get spinned up once in a while. 

 

I now from V5 that using "inotifywait -mr -e access -e modify /mnt/disk1" (or other disk) can give insight in what is causing it. But for some reason the inotifywait command now always shows "ACCESS, ISDIR" every second on every disk, even when all plugins and dockers etc. are disabled/off.

 

Also, i remember that the inotifywait command also showed some info about who is accesing the disk, but that info also seems to be missing? If there is activity, you can only see WHAT the activity is, not WHO is doing it... e.g. if Apple's TImemachine is accessing a disk, i want to see that.

 

Is there a way to find out WHO is responsible for spinning up a disk? Why is there no such option in V6 built in?

This is one of the most frustrating issues i'm experiencing...

 

*edit* as we speak, all my disks were spinned down, and inotifywait is running, watching disk9. it shows 'ACCESS, iSDIR' constantly on disk9. All of a sudden, disk9 gets spinned up for some reason, but inotifiwait dos NOT show anything else but 'ACCESS, ISDIR' every second... totally useless... in this case, i suspect it is apple's timemachine messing things up, since the backup folder for timemachine is located on disk9. It would be a great help if there was a way to find out if the disk was spinned up from an internal or external process.

Edited June 4, 2018 by jowi

[jowi]

Is there some (python) script that can tell if a disk is idle or spinning? I know this can be seen from the user interface, but i would like to collect data on when and how long they are spinning, to see if there is some logic to all this madness.

[djcslip]

Wonder if you and I are experiencing the same issues? I'm having the same results if i run inotifywait -mr -e access -e modify /mnt/disk1 >> (ACCESS,ISDIR)

 

Edited June 19, 2018 by djcslip

[jowi]

It is hard to say. Inotifywait shows the ISDIR/ACCESS activities on all disks constantly, every 1/2 second, if they are spinned up or down, doesnt matter, and it does NOT give any clue on who is doing it. So i don't think this is what is spinning up the disks. The ISDIR/ACCESS might be some internal unraid proces, don't know...

 

In my case i suspect Sonarr (download agent) for spinning up a lot of disks regularly when it is scanning for episodes etc, and the timemachine backup proces (which in my case runs once every week, NOT once every hour!)... i've been monitoring it closely, at the moment these 2 processes look like to be responsible for my spinup issues, but i'm not sure.

 

The activity plugin seems nothing more then some UI wrapper around inotifywait processes, and while it does show there is access, it does not show who is doing it... you have to deduct from the result you are seeing. Tricky, if not impossible. My experience with the cache_dir thing is even worse, for some reason it actually keeps spinning up all my disks all the time... removed it immediatly... 

[jowi]

No one can point me in the direction of a python script that can tell if a disc has been spinned up?

Like the shell "mdcmd" ? Or does mdcmd also returns some value if a disc is spinning or not? 

[Squid]

hdparm -C /dev/sdX will return whether the drive is spinning or not

[jowi]

root@UNRAID:~# hdparm -C /dev/sdj

 

/dev/sdj:

 drive state is:  active/idle

root@UNRAID:~# hdparm -C /dev/sdj

 

/dev/sdj:

 drive state is:  standby

Great, thanks!