Viewing main causing drives to spin up

[unevent]

Is there a way to prevent drives from being spun up when viewing tower/main?  All drives can be spun down and loading tower from a browser causes several drives to spin up.  Sometimes a couple, sometimes all.  6.1.6.

 

[Squid]

Is there a way to prevent drives from being spun up when viewing tower/main?  All drives can be spun down and loading tower from a browser causes several drives to spin up.  Sometimes a couple, sometimes all.  6.1.6.

Never seen this behaviour.  How do you know that they are being spun up?

[unevent]

Is there a way to prevent drives from being spun up when viewing tower/main?  All drives can be spun down and loading tower from a browser causes several drives to spin up.  Sometimes a couple, sometimes all.  6.1.6.

Never seen this behaviour.  How do you know that they are being spun up?

I can hear them spin up and there is a delay to page load.  If drives are already spining it is fine, page load is quick.  I dont want drives to spin up just because i view the GUI.

[bonienl]

Is there a way to prevent drives from being spun up when viewing tower/main?  All drives can be spun down and loading tower from a browser causes several drives to spin up.  Sometimes a couple, sometimes all.  6.1.6.

 

This is not normal behavior. GUI shouldn't start anything which could trigger a disk spin up.

 

Does this happen to all your disks, if not what is the difference? Other type of connection SATA, IDE? Specific controller or directly connected to motherboard?

 

[tucansam]

This happens on both my 6.x servers, and has happened since 5.x.  The web GUI is slow to load, because it seems to wait until some (not all) disks spin up, and then loads the web GUI.  Has been this way for years across every install I've ever done, even plain vanilla (which is what I am running now, no dockers, no virtualization, no extra programs running, just unraid)

 

[bonienl]

This happens on both my 6.x servers, and has happened since 5.x.  The web GUI is slow to load, because it seems to wait until some (not all) disks spin up, and then loads the web GUI.  Has been this way for years across every install I've ever done, even plain vanilla (which is what I am running now, no dockers, no virtualization, no extra programs running, just unraid)

 

Can you post a diagnostics file. Your case is definitely not the norm.

[unevent]

 

This is not normal behavior. GUI shouldn't start anything which could trigger a disk spin up.

 

Does this happen to all your disks, if not what is the difference? Other type of connection SATA, IDE? Specific controller or directly connected to motherboard?

 

Number of disks appears random.  Sometimes two, sometimes three or four.  At least once it was all, in staggered spinup.  I tried disabling page refresh, but made no difference.

 

As far as drives on motherboard vs addon, will have to get back on that.  They are all SATA and are all Toshiba 3TB.

 

[unevent]

This happens on both my 6.x servers, and has happened since 5.x.  The web GUI is slow to load, because it seems to wait until some (not all) disks spin up, and then loads the web GUI.  Has been this way for years across every install I've ever done, even plain vanilla (which is what I am running now, no dockers, no virtualization, no extra programs running, just unraid)

Don't recall this behavior on stock 5.x or with Dynamix on 5.  Only seen this in 6 (6.1.3 and whatever the latest is, 6.1.6?).

[unevent]

Update: Doesn't matter if on motherboard or the addon cards.  Also, its not spinning up to get temps as the drives are active but no temps reported for some of them as seen in the screen shot.  If I do a manual spin up or use the drive, the temps are reported as normal.  Running 'smartctl -i -n standby /dev/sdX1' reports ACTIVE OR IDLE for those spun up and STANDBY for the drives spun down.

[unevent]

Still doing it.  Spun up 3 & 4 this morning when loading tower/main.  It never showed drive temps for the drives it spun up.  Is it querying drive smart data on load?  Is there source code available for main?

[Squid]

Temperatures for drives are checked by default every 30 minutes (so as to not introduce performance issues by bombing the drive with smart requests), unless you manually spin up a drive(s).  The setting for that is Settings / Disk Settings

[unevent]

Temperatures for drives are checked by default every 30 minutes (so as to not introduce performance issues by bombing the drive with smart requests), unless you manually spin up a drive(s).  The setting for that is Settings / Disk Settings

Not interested in the disk temps, they report normal when drive is spun up manually or for normal file access.  When main loads it will sometimes spin one or more drives up, but even though it spun a drive up for no reason it will not report the temp for it no matter how many times page is refreshed.  It will report when the drive is accessed for a file or manually spun up. Temps are not my concern, the lack of reason for spinning up drives just to view the GUI a couple times a day is.  It appears to be some sort of smart query, but there are passive ways to check if the drive is spun up and act accordingly.

[Squid]

Temperatures for drives are checked by default every 30 minutes (so as to not introduce performance issues by bombing the drive with smart requests), unless you manually spin up a drive(s).  The setting for that is Settings / Disk Settings

Not interested in the disk temps, they report normal when drive is spun up manually or for normal file access.  When main loads it will sometimes spin one or more drives up, but even though it spun a drive up for no reason it will not report the temp for it no matter how many times page is refreshed.  It will report when the drive is accessed for a file or manually spun up. Temps are not my concern, the lack of reason for spinning up drives just to view the GUI a couple times a day is.  It appears to be some sort of smart query, but there are passive ways to check if the drive is spun up and act accordingly.

I understand that the temperature reporting for the drives is not your major concern.  I answered your question about it.  Also, I believe that bonienl had asked for you to post your diagnostics for the system so that we can try and see if there are any clues within it, but I haven't seen it yet.

[unevent]

Temperatures for drives are checked by default every 30 minutes (so as to not introduce performance issues by bombing the drive with smart requests), unless you manually spin up a drive(s).  The setting for that is Settings / Disk Settings

Not interested in the disk temps, they report normal when drive is spun up manually or for normal file access.  When main loads it will sometimes spin one or more drives up, but even though it spun a drive up for no reason it will not report the temp for it no matter how many times page is refreshed.  It will report when the drive is accessed for a file or manually spun up. Temps are not my concern, the lack of reason for spinning up drives just to view the GUI a couple times a day is.  It appears to be some sort of smart query, but there are passive ways to check if the drive is spun up and act accordingly.

I understand that the temperature reporting for the drives is not your major concern.  I answered your question about it.  Also, I believe that bonienl had asked for you to post your diagnostics for the system so that we can try and see if there are any clues within it, but I haven't seen it yet.

He asked tucan for diag when he quoted his post but i can post as well.  Away from server, will be later this evening.

[kizer]

Have you installed any plugins or I should say "Unverified Plugins" or are you running a out of the box 6x Build?

[unevent]

Have you installed any plugins or I should say "Unverified Plugins" or are you running a out of the box 6x Build?

 

Diag attached.  Only plugin is Virtualbox running two VMs and running seven Dockers, all from cache drive.  Drives spin up when access the web GUI (I am standing right next to the server when I view the GUI and hear thew drives start up and there is usually a delay in the GUI main loading.  I only check GUI about twice a day, once in morning and once in the evening for drive temps and logs.

tower-diagnostics-20151217-1728.zip

[Squid]

What are these two lines (and do they coincide with the drives spinning up)?

 

Dec 17 17:26:58 Tower rpc.mountd[3024]: authenticated mount request from 192.168.150.36:913 for /mnt/cache (/mnt/cache)
Dec 17 17:26:58 Tower rpc.mountd[3024]: authenticated mount request from 192.168.150.36:1011 for /mnt/user/Movies (/mnt/user/Movies)

 

Or, let me rephrase... I know its NFS related, but what I'm wondering if your using a laptop or something to check the GUI, and when you wake it its actually doing something to cause the drives to spin up... 

[unevent]

What are these two lines (and do they coincide with the drives spinning up)?

 

Dec 17 17:26:58 Tower rpc.mountd[3024]: authenticated mount request from 192.168.150.36:913 for /mnt/cache (/mnt/cache)
Dec 17 17:26:58 Tower rpc.mountd[3024]: authenticated mount request from 192.168.150.36:1011 for /mnt/user/Movies (/mnt/user/Movies)

 

Or, let me rephrase... I know its NFS related, but what I'm wondering if your using a laptop or something to check the GUI, and when you wake it its actually doing something to cause the drives to spin up...

 

I run autofs on ubuntu which automounts those NFS shares when I run a script that does an 'ls' on those shares.  After a period of time they auto unmount.  Since I do not run cache_dirs at the moment, running that script will cause drives to spin up.  IOW, independent/unrelated to loading the GUI/main.  Strange no one else is having this issue as there is nothing special about my build or what I am running.  Firefox on Linux or Chrome on Android produces same result on main.  If the drive spins down after timeout vs. being spun down manually seems to make the difference on when or if they spin on main load after an extended period of time.

 

[Squid]

What happens if you hit the webUI from your Phone?

[unevent]

What happens if you hit the webUI from your Phone?

 

Same as doing it from PC as mentioned.  There is a delay as one or more drives spin up then main is shown.  What drives that were spun up do not show temp info regardless of number of times I refresh.  If I spin drives down manually nothing spins up.  If I spin drives up temp shows.  If they timeout and spin down and has been couple hours, almost guaranteed to spin one or more up when loading GUI from PC or phone.

[bonienl]

Have you tried to eliminate the problem? Start your server in safe-mode without plugins and check what happens. Another check is to disable temporary dockers.

 

[unevent]

Have you tried to eliminate the problem? Start your server in safe-mode without plugins and check what happens. Another check is to disable temporary dockers.

 

Will give it a shot over the weekend.

[unevent]

Have you tried to eliminate the problem? Start your server in safe-mode without plugins and check what happens. Another check is to disable temporary dockers.

 

Will give it a shot over the weekend.

Update: reboot with no changes solved it temporarily.  Plugin dependencies were giving me issues so i switched to Dockers  shortly before this problem started occuring.  Dockers have been hit and miss for reliability.  Had plex Docker lock everything up because it would not start and when it did it always wanted to push an update even though it already updated.  Out of the box it ran great.  Stopped it to add a new folder map and never wanted to start again even after reboot.  Hell with that, back to plugin version for that app.