Tray Monitor hangs/crashes out of sleep mode

[Bugman]

Dean: This is a very old issue that occurs from time to time on a Win 7 machine (CQC 4.8.0) that I use as my movie viewing PC with the XBMC driver (now on Kodi build). Itunes was not installed in the tray application install screen, as I'm just using XBMC at this time. I have this machine go to sleep when not in use as it supports a FlexRaid server with around 12 TB of storage, so don't want the drives running when not in use.

Probably around 20% of the time it comes out of sleep mode (WOL command) with the symptom that the Tray Mon is an active process but is hung up or non functional with regard to the XBMC driver. When I go to the Admin Interface and try reconfiguring the XBMC driver, it shows the Tray Monitor on the machine, but hangs at waiting for comm resource. This occurs with and without XBMC running before sleep mode kicks in.

The most recent instance (log entry below), is that the Tray Monitor goes offline and disappears from the task bar. Typically I need to end the process and start the Tray Monitor up manually from the program menu. Restarting the process does nothing. I had the Tray App in verbose mode, so if there are logs elsewhere to reference other than the main logs let me know.

I would love to track this down as it's annoying as hell...I'll keep the machine in this state if that might help track down the issue.

Next steps?

Here's the most recent log entry:

01/17 08:48:34-zotac, CQCServer, CQCDrv_XBMCThread26
{
CQCKit, MEng.System.CQC.Runtime.CQCLogger.128, Status/App Status
Connect: Exception - An unknown exception recieved from server object
}
01/17 08:48:34-CQC-HTPC, CQCTrayMon, CIDOrbSrvWorkThread_3
{
CIDOrb, CIDOrb_ThisFacility.cpp.552, Failed/Unknown, Error: 6502/0/0
An unknown exception recieved from server object
RemoveRecord
<CQCServer> CQCKit_AppCtrlClientProxy.cpp - 585

[Dean Roddey]

So you are using an app control driver, for XMBC? It looks like that thread is getting some sort of system exception. I'll take a peak at it. These things are stupidly hard to debug because of the sleep thing being involved, but I'll try to think my way through it.

In the meantime, have you considered letting Win7 shut down the drives without putting the machine to sleep? The power management stuff allows you to stop the drives spinning after a period of inactivity.

[Bugman]

So you are using an app control driver, for XMBC?

In the meantime, have you considered letting Win7 shut down the drives without putting the machine to sleep? The power management stuff allows you to stop the drives spinning after a period of inactivity.

Yes, for app control of XBMC.

Let me check on the drive shutdown option, but there may be a complication as I'm running the drives through a SAS controller, and expander.

[Dean Roddey]

Based on the time stamp, is that error from the time it went to sleep, or the time it woke up? I.e. was the damage done on shutdown but didn't really show up until it tried to start up, or did it really happen at startup?

[Dean Roddey]

I'm beginning to suspect that the system isn't giving the program enough time to get itself cleaned up, and it's leaving it sort of in limbo land. When it wakes up, it tries to pick back up again in this half-baked state and croaks. Unfortunately I can't tell from the available bits if the error is actually happening on shutdown (and then it gets frozen in that state until later woken up), or if it's just freezing it halfway through cleaning up, and it happens later when it wakes it back up halfway through a cleanup. I may have to add some logging to help with that.

The whole sleep thing is sort of silly in that it doesn't allow, AFAIR, for programs to ask to be given more time. I understand that it's a user convenience thing, to allow for quick shutdown. But programs that have to deal with the outside world just can't guarantee that they can clean up that fast.

[Bugman]

Based on the time stamp, is that error from the time it went to sleep, or the time it woke up? I.e. was the damage done on shutdown but didn't really show up until it tried to start up, or did it really happen at startup?

Dean: Must have been at startup given the 8AM+ timestamp...

Willing to test anything thrown way