Cache Drive does not spin down unless I force it to.

I am running the latest rc 6 test beta, and even prior to this beta my cache drive does not spin down unless I manually spin down the drive. Only have 3 drives at the moment.  1 parity 1 data and cache.  I wouldnt even use a cache but the wife complains copying to server is too slow without it.  Is this normal?  I am running Serviio plugin and unMenu.  I use Serviio for streaming to xbox and ps3. I use unMenu for battery backup safe shutdown and system alert emails.

syslog.txt

What happens if you stop Serviio?  I would suspect it's what's keeping the drive spun up.

I have uninstalled Serviio and JRE since serviio requires it.  Cache drive is still not spinning down.  Only thing non stock is unmenu and I really rather not uninstall that since it gives me the ability to get screen and apc plugin, and safe shutdown etc.

In my sys log I can see where my parity and data drive spin down. But no log command for the cache drive.

 

Jul 17 21:45:44 Tower kernel: mdcmd (33): spindown 0

Jul 17 21:45:45 Tower kernel: mdcmd (34): spindown 1

Enter "lsof /mnt/cache"

Enter "lsof /mnt/cache"

That will only let you see what is using the cache drive at the instant you invoke the command.  It will not let you know what may be accessing the drive periodically.

 

A better command to monitor the drive accesses is described here:

http://lime-technology.com/forum/index.php?topic=21401.msg190179#msg190179

 

Since the cache drive can be accessed through the user-shares, probably need to check accesses via /mnt/user and /mnt/cache

 

Joe L.

Enter "lsof /mnt/cache"

That will only let you see what is using the cache drive at the instant you invoke the command.  It will not let you know what may be accessing the drive periodically.

 

A better command to monitor the drive accesses is described here:

http://lime-technology.com/forum/index.php?topic=21401.msg190179#msg190179

 

Since the cache drive can be accessed through the user-shares, probably need to check accesses via /mnt/user and /mnt/cache

 

Joe L.

Have you set the spin-down time for the cache drive?  (to be sure, set it to an alternate time, then re-set it to the desired time. )  You'll see "hdparm" commands in the syslog setting the disk's internal spin-down timer.

Ok well changing the spindown time on the cache drive from default to 30mins saving then changing to 15 mins instead of default which is also 15 mins worked. Now to test this after a reboot. Will report back!

Ok this seemed to fix this issue.  All drives now spin down automatically.  This is a bug no?  It was set as default just like all my other disks.  All other disks spun down just fine except cache.  I had to change the time to 30mins then back to 15mins in order to get cache drive to spin down.  Just kind of wondering why my thread was moved from the Version 5 RC 6 thread.

This is not isolated to current betas - it has often been necessary to fiddle with spindown timers in order to get the setting to take effect.

 

If, as Joe suggests, the spindown is implemented by hdparm setting on the drive itself then, perhaps, it's explicable.  If spindown timers are set (on the drive), then a drive is changed, or added, the new drive won't have had the hdparm command applied to it?

  Just kind of wondering why my thread was moved from the Version 5 RC 6 thread.

It is not a issue specific to the rc5/6 release.