Feature Request: Spin Up

[Brucey7]

Ability to set by user share whether or not to spin up all disks in the share

 

E.g.  13 disks in 1 share, when browsing a media catalogue, there are large delays on each access as individual disks are spun up separately on demand until all 13 are spinning.

[JonathanM]

Ability to set by user share whether or not to spin up all disks in the share

 

E.g.  13 disks in 1 share, when browsing a media catalogue, there are large delays on each access as individual disks are spun up separately on demand until all 13 are spinning.

Can you not accomplish what you want with spinup groups in disk settings? That's been a feature for many years.

[itimpi]

Ability to set by user share whether or not to spin up all disks in the share

 

E.g.  13 disks in 1 share, when browsing a media catalogue, there are large delays on each access as individual disks are spun up separately on demand until all 13 are spinning.

I agree that this will be a useful feature.

 

It is worth pointing out that Joe L. produced a shell script that will spin up all disks for a share as soon as one of them is accessed.  This might act as stopgap and help with the particular case you mention.

[Brucey7]

My Spinup Groups has always been set to Yes.  I am fairly technical and I find all the documentation on this feature as clear as mud.  As a user, I prefer everything in the Web GUI, I like my unRAID to work out-of-the-box and if it's not on a menu, I really don't want to start learning yet another operating system.

 

Spinup Groups has been around a lot longer than unRAID and the purpose of Spinup Groups elsewhere has always been is to minimise the load on the power supply and vibration as many disks are spun up, typically when you first turn on a server, it would spin up disks in groups every few seconds.

 

I can't find any other options on the Web GUI apart from this one switch.

 

It would be more useful to have this option with the user share, with 3 options, No, Yes, Yes but don't keep alive.

 

Accessing a user share could thus spin up all disks in that share, and if (say) you begin to watch a movie it could additionally set with "Yes but don't keep alive" to spin down the unused disks after the default spin down delay.  To implement this, it would have to remember which disk caused the last user share spinup command and any further accesses to that disk wouldn't cause a user share spinup until another disk in that user share was accessed at which point it would then became the master.

 

In my case, I have a movie share spanning 13 disks and the fanart, posters, .nfo files are needed by ember media manager when I browse the catalogue and all disks spinning up would save me from 13 delays.  Once I start watching a movie, the other 12 disks would spin down after the default delay.

 

A lot more elegant solution for me would be to store my fanart/posters/nfo's on my home theatre PC for local access, but alas it's not a feature in Ember Media Manager.  I do also use XBMC but when you have over 15,000 movies, XBMC is just too simplistic.

 

 

[RobJ]

Cache_dirs is an alternative way to handle this, by keeping the directories in memory, and therefore not needing to spin up *any* drives except the one the actual movie being played is on.  You will need to experiment with it, probably try the -u option.  I can't say if this is a total solution for you or not.

[WeeboTech]

It looks like spin up groups exist. I know they do at the controller level.

 

 

Perhaps you could go into the gui and review.

Perhaps even adjust it so the majority of drives you want together are on the same controller and/or adjust drive settings so they spin up on common Spinup group(s) name.

 

 

I don't have a large enough array to experiment with adjusting the Spinup group name from one controller to another.

It's my understanding this is done to force spin ups on a common Spinup group name to avoid delays just as being described.

 

 

The effort on the user share part is identifying all drives in a user share and insuring they all share the same common group name.

 

 

 

[Brucey7]

I do use cachedirs (I wouldn't be without it).  My issue isn't directories being cached, it's 3 smallish files per movie (thumbnails, posters and movie descriptions), my catalog program opens 3 files per movie to display the catalog on the screen.  Going up & down the screen with the arrow keys on my HTPC opens 3 files per movie, the list of movies is stored in the catalog.  So with 13 data drives in a user share, I get 13 delays, thereafter it's fine until they spin down again.

 

Deciding what movie to watch by browsing the catalog invariably results in 13 separate delays.

 

Actually, spinup groups at the controller level is useless, it's like saying spinup groups are linked to the colour of the cable, it's nonsense.  Sooner or later I will have 20 drives in one user share, i.e. on every controller.

 

They key to cracking this nut is linking it to user shares.  First access on a spun down user share could spin up all disks in that share.

[WeeboTech]

I do use cachedirs (I wouldn't be without it).  My issue isn't directories being cached, it's 3 smallish files per movie (thumbnails, posters and movie descriptions), my catalog program opens 3 files per movie to display the catalog on the screen.  Going up & down the screen with the arrow keys on my HTPC opens 3 files per movie, the list of movies is stored in the catalog.  So with 13 data drives in a user share, I get 13 delays, thereafter it's fine until they spin down again.

 

Deciding what movie to watch by browsing the catalog invariably results in 13 separate delays.

 

Actually, spinup groups at the controller level is useless, it's like saying spinup groups are linked to the colour of the cable, it's nonsense.  Sooner or later I will have 20 drives in one user share, i.e. on every controller.

 

They key to cracking this nut is linking it to user shares.  First access on a spun down user share could spin up all disks in that share.

 

 

Any way to use an SSD in the array and migrate the informational files onto that?

[Joe L.]

This script is exactly how I solved the same issue you are describing.

 

This script is a different approach to making media players happy (and wife happy as a result) by immediately spinning up all the disks affiliated with a given "user-share" whenever any file or folder within the user-share is accessed.  This is NOT a green approach as it will result in more spinning disks, so those trying to minimize power consumption can stop reading now.

 

For those with media players that will not wait for disk spinup, and time-out, and make for a less enjoyable movie watching experience, this might work.

 

 

http://lime-technology.com/forum/index.php?topic=4858.0

 

Joe L.

[Brucey7]

@Weebo, I can't use an SSD because of the layout

HD Movies (user share)

    Movie1 (folder)

        Movie1.mkv (only needed if actually play the movie)

        Movie1.nfo (catalog opens to display IMDB scrape)

        Movie1.tbn (catalog open to display thumbnail)

        Movie1-fanart.jpg (catalog opens to display fanart)

  Movie2 (folder)

        Movie2.mkv

        Movie2.nfo

        Movie2.tbn

        Movie2-fanart.jpg

 

all the way down to

    Movie15000 (folder)

      etc etc

 

@Joe L.

I will try your script, thank you very much.

 

 

[Brucey7]

Joe, your script might have a bug, on one of my systems only the disk I access spins up, on the other, all the disks except parity are alway spun up whenever I check with the web gui