Dropping a drive

[ytulpan]

I have a drive that keeps dropping out of the pool every few days. When I see the alert, I look into Event Manager and there is no event whatsoever, HD sentinel shows nothing, the disk shows OK in the device tree and in disk manager. I can access the drive via the OS (Win 2008R2) with no issues.

 

Thing is, the only way that works for me to reset the situation is a reboot. Not very convenient. The drive is a green 2TB Seagate ST2000DL003. It is connected via a port multiplier to an JMB SATA port on the MB. I have 14 other identical drives and 5 WD 4TB red drives, similarly connected with no issues.

 

Help !

[Christopher (Drashna)]

First thing I would recommend is to check the SATA connections, and maybe swap out the cable, if you have the option.

 

If that doesn't help, right click on the disk in question, select "Disk Settings", and check the "Do not query SMART data" option. 

 

If that helps, the enabling the "Smart_NoWMI" option may help:

http://wiki.covecube.com/StableBit_Scanner_Advanced_Settings

 

Also, There is the "burst test". This is good for troubleshooting problematic controllers. It is possible that the controller is failing/crashing/reseting under heavy load.

 

[ytulpan]

Thanks,

 

I don't have the scanner installed. I had some trouble with installing it and I left it as such...

 

I think your advice refers to it.

 

Anyway I have given the VIP treatment to my disks, set all of them to max performance and updated the controller driver, let's hope the voodoo  will help.

 

The strange thing is that I do have a competent disk monitor suite and it didn't complain, and also the event log doesn't show anything.   

[Christopher (Drashna)]

Sorry for misunderstanding.

 

If the drives are actually dropping out of the system....

 

Well, check the Event Viewer on the system. The "System" section should have "disk" or "controller" related events listed. That should shed some light on what is going on.

 

Also, what controller(s) are you using, specifically? Or motherboard?

And have you tried resetting or swapping out cables?

 

 

Also, is this happening under heavy load? Or a light load? Or...?

[ytulpan]

Thanks Drashna,

 

I have no new dropouts since I massaged the drives and replaced the controller driver.

 

The controller is a JMB362 on a Gigabyte GA-P55-UD6 MB. It's connected to a SiI3726 port multiplier.

 

The 2 issues concerning me are :

 

1. There was no indication anywhere of a problem (eventlog, HD sentinel). Also the drive was accessible with no problems by the file system.

2. I could not make DP to see the drive again without rebooting (tried ofline/online, enable/disable in device manager)

 

 

[Christopher (Drashna)]

If you've had no issues with the drives since changing the controller's drivers, then that's great. And it sounds like the drivers are buggy.

 

 

As for why this happened at all.... if there is an issue with the driver, it would affect the Virtual Disk service. And since we rely on that a lot, it would affect DrivePool as well.

 

Could you post the Error Reports from the system? 

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

Well, could you upload the entire folder, and not just the error reports folder?

[ytulpan]

I am uploading everything under the "Service" folder.

 

But still, if there was a problem with the VDS, shouldn't I see something in the event log ? 

 

And also, how do I force DP to see the drive short of rebooting ? Can I reset the driver ? I should remind you that I cannot restart the service - it hangs on stopping.

[Christopher (Drashna)]

Restart the service (services.msc, and look for "StableBit DrivePool Service".

Since that controls the UI...

 

If it's hanging still, make sure you're on the newest version (2.1.0.561 is the newest "Windows" version, 1.3.6.7582 is newest WHS version).

 

If it's still hanging on that version, then reset the settings (basically, delete everything you've just uploaded.

http://wiki.covecube.com/StableBit_DrivePool_Q2299585B

 

If it still persists after this, then get a memory dump of the service:

http://wiki.covecube.com/StableBit_DrivePool_Service_Memory_Dump

Issue reported here:

https://stablebit.com/Admin/IssueAnalysis/5254

[ytulpan]

Will do. Thanks.

[Christopher (Drashna)]

Sorry for the delay!

 

Looks like it was definitely an IO error. And it's definitely hardware.

 

Specifically, it looks like it's this disk: "\\?\Volume{6f89df43-e9b4-11e3-8d17-00241d7f56d6}\"

You can identify which disk that is by running "cmd /k mountvol". That should list all the disks and the "volumes" like this.

 

If you're still experiencing the issue after replacing the controller, then it may be a bad cable or the disk itself.

[ytulpan]

Thank you Christopher,

 

It could be that the rack power supply is failing. Anyway, thanks for the diagnostic.

 

Yossi

[Christopher (Drashna)]

I'm very sorry to hear that! It's never fun when a power supply fails.

 

Hopefully, you can get it replaced quickly.

 

And if there is anything else that we can do or if you have any other questions, don't hesitate to ask.