Setting for delay before "disk missing"

[thepregnantgod]

Drashna, can you explain how I can tweak the setting that Drivepool uses before they mark a disk as "missing?"

 

I have a few Seagate disks that seemingly take forever to spin up and after a very short while, then Drivepool marks them as missing which of course puts the pool in read only mode.

 

The only way I can get them live is to unplug the SAS/SATA cable while they've been powered up for a while, then plug it back in.  Assuming there's activity, I don't have a problem, but I suspect once they spin down and Drivepool tries to write data again and they don't respond fast enough, then missing!

 

I remember a tweak somewhere that can make this much longer before marked as missing.

 

Additionally, do you know of a tweak - perhaps via scanner that doesn't allow these troublesome drives to spin down?

[Christopher (Drashna)]

Well, Enabling the "BitLocker_PoolPartUnlockDetect" setting in the advanced config will do this, but this should be enabled by default in the beta versions. 

 

Also, StableBit Scanner will query the drives every minute or so, waking them up or keeping them awake. Unless you've enabled throttling (in the scanner settings).  So if you have enabled it, disable it. 

 

 

Otherwise, I suspect that it's not just that the disks are idling/sleeping, as that wouldn't cause DrivePool to flag the disks as missing.

 

The volumes would have to be unmounted.  This is generally done by disconnecting the drives.   

 

 

I would recommend checking the Event Viewer, and the "System" section, and see if you're getting any disk, controller or ntfs errors.... especially "surprised removed" errors when this happens.

 

 

 

 

 

that said, we document all of the advanced config stuff here:

http://wiki.covecube.com/StableBit_DrivePool_2.x_Advanced_Settings

 

 

And the only setting that would be close to this is "CoveFs_WaitForVolumesOnMount" and the Ms version this, which control how long the service waits on service start before "timing out" the drives.

 

 

 

And ideally, I would prefer if you grab logs from this happening:

http://wiki.covecube.com/StableBit_DrivePool_2.x_Log_Collection

 

Enable it, let it sit until this happens and then stop/disable the logging once it occurs. 

 

And then run the StableBit Troubleshooter:

http://wiki.covecube.com/StableBit_Troubleshooter

[thepregnantgod]

I am getting a lot of these...

 

Event ID 11 - driver detected an error on \Device\Raidport0.

 

Like about a dozen or so....

[Christopher (Drashna)]

That could be nothing, or it could be an indicator of a problem.   

 

It would depend on the timing and frequency of the occurrences. And if you're seeing any other errors around the same time.

 

If you've upload the files, let me know.