[6.6.2] Some situation disks show standby but actually spindup

diagnostic.zip please

This is an old issue, I reported this over two years ago during the 6.2 beta cycle, though it doesn't happen with all disk brands/models, e.g., on my main server currently I have one of each WD 6TB green, blue and red and they don't have this issue, but most Toshibas do, looking at your screenshot looks like some WD do too, easy giveaway is the disk icon stays grey but temperature starts showing, and it won't  spin down again on its own.

 

P.S. diags from this server are useless for this, but if the OP doesn't post his I will post later from one of my other servers.

Edited October 16, 2018 by johnnie.black

47 minutes ago, johnnie.black said:

This is an old issue, I reported this over two years ago during the 6.2 beta cycle

Ok I see what's happening.  Thanks for the report, I guess it got lost among all the other posts in that topic at that time.

 

There's an easy way to fix, and a hard way.  The easy way is to prevent any of the SMART buttons from being enabled if the device is spun down, thus requiring user to spin up device first.

 

The hard way is to determine if the device is currently spinning or not.  Believe it or not, this is not that easy to work out an algorithm that works across all device types and interface types (eg, SATA vs SAS).  At present 'emhttp' keeps track of spinning status based on "last I/O time" along with knowing when a device is explicitly spun down via an emhttp api call.  The bug you're seeing is that webGUI issues a SMART command out-of-band from emhttp, meaning emhttp doesn't know this command has been issued, and the command is not counted as an I/O command.

 

So we'll have to think about what's the best way to solve this.

 

I also was curious about the GUI drive status a while back. 

 

So I wrote a basic script to check on the spin states ever hour using both smartctrl and hdparm to a file.

Eg   smartctl -n standby /dev/sdg      and     hdparm -C /dev/sdg

 

Summary Results

smartctl                                                        hdparm

                                                                             

Device is in ACTIVE or IDLE mode                   drive state is:  active/idle

Device is in IDLE_A mode                               drive state is:  unknown

Device is in IDLE_B mode                               drive state is:  unknown

Device is in STANDBY mode, exit(2)                drive state is:  standby

 

I use 8TB (Archive) Seagate drives and discovered that Seagate’s are not either Active or Standby but have (Active,  IDLE_A, IDLE_B, IDLE_C and  Standby)  states  [reference  “tp608-powerchoice-tech-provides-gb.pdf” ]      The first three are full spin states with various electronic power savings.  IDLE_C (SAS only) is a slow spin state.

 

NB: Despite what the Seagate doc said, the drives seem to spend ‘hours’ in both IDLE_A and IDLE_B mode before entering “Standby”  

 

So the drives can be  “not active” but also “full spin”, and if the GUI uses hdparm then also “unkown”.  

 

Note and thanks all.

 

3 hours ago, limetech said:

At present 'emhttp' keeps track of spinning status based on "last I/O time"

That are different as I think Unraid will keep track the active/idle status of disk and determine need command out  spindown or not again.

 

I understand and agree if keep track disk status may cause unnecessary spinup and create another problem. Different disk type/model/firmware also have different behavior.

Edited October 17, 2018 by Benson