Backup Exec Device & Media Service won't start

[leonstr]

Hi,

I've got a cluster node where the Backup Exec services are configured as resources but the Backup Exec Device & Media Service won't come online, it fails immediately.

I've traced it with Process Monitor and it's exiting with return code 128.

If I try to run pvlsvr.exe with the -console switch the process terminates after a few seconds. Because the -console switch opens a new console window, when the process terminates the window closes and I can't see the output.

Is there a way to get debug output from pvlsvr.exe? Other services appear to have a -debug switch but this doesn't seem to work for this service and I can't find any documentation on configuring this.

Backup Exec 2010, pvlsvr.exe v13.0.2896.118, Windows Server 2003 SP2.

Thanks in advance,

Leon

[CraigV]

Hi Leon,

I'm going to suggest you upgrade to BE 2010 R3 and patch your cluster fully with the available updates. Run the Cluster configuration wizard again if need be...

But BE 2010 R3 seems to have sorted out a couple of funny issues that other OPs have been experiencing. You can use your current license keys, so there is no need to change them. Before doing the upgrade, grab a copy of your Catalogs and Data folders in case something goes wrong with the upgrade.

You can download the latest version here:

https://www4.symantec.com/Vrt/offer?a_id=93355

Thanks!

[Colin_Weaver]

if you run SGMON.exe with Devcie and Media and Capture to file checked and then try to start Devcie and media it should capture something

[leonstr]

CraigV: Thanks but I've got a bunch of largely identical clusters so I'm reluctant to upgrade - all the others are working (including the partner node of this one) so this should work too.

Colin Weaver: Thanks, I'll fail over to the problematic node and see what I get.

[sbora]

Can you also check if a tape device is connected to the server? If yes, can you disconnect the tape device from BE and check if device and media service starts fine? 

[leonstr]

This is what SGMon logs when I run pvlsvr -console (as the same user as the service):

 :
PVLSVR:   [09/07/11 17:50:59] [6388]     PvlDevice::DriveUtilThreadProc()
            Drive = {76E85AF7-E382-4735-813C-23687D541638}, "IBM 3"
            DeviceState = 0003
PVLSVR:   [09/07/11 17:50:59] [4156]     PvlDevice::ChangerUtilThreadProc()
            Library = {E5D8CDCA-0DBF-45F4-8D55-D68FE4F865AE}, "IBM 4"
PVLSVR:   [09/07/11 17:51:00] [1748]     PvlDevice::MonitorThreadProc()
            DateTimeSync: Local = 2011/09/07 17:51:00, System = 2011/09/07 16:51:00 GMT, Database = 2011/09/07 16:51:00 GMT
PVLSVR:   [09/07/11 17:51:00] [1748]     [StorageManager::NagLicenseChecking()] Storage Manager not initialized
PVLSVR:   [09/07/11 17:51:03] [4648]     PvlSvrServiceStart() Stop Service Request - start
PVLSVR:   [09/07/11 17:51:03] [4648]     PhysVolLibServer::ReleasePhysVolLibServer()
PVLSVR:   [09/07/11 17:51:04] [4648]     PvlSvrServiceStart() Service stopped
SESCLU:   [09/07/11 17:51:31] [0000]     SescLu - CContentUpdateManager::Initialize multiple tries to Initialize
:

For comparison this is the log on the other (supposedly identical) cluster node where the services are working:

 :
PVLSVR:   [07/09/11 17:47:41] [7236]     PvlDevice::DriveUtilThreadProc()
            Drive = {872600D7-3DBD-4920-B7C7-25D89769E64A}, "IBM 1"
            DeviceState = 0003
PVLSVR:   [07/09/11 17:47:41] [6756]     PvlDevice::ChangerUtilThreadProc()
            Library = {68DAE149-C1BA-448F-B4C1-6233BC80BFBA}, "IBM 2"
PVLSVR:   [07/09/11 17:47:46] [4184]     PvlSvrServiceStart() Stop Service Request - start
PVLSVR:   [07/09/11 17:47:46] [4184]     PhysVolLibServer::ReleasePhysVolLibServer()
PVLSVR:   [07/09/11 17:47:47] [4184]     PvlSvrServiceStart() Service stopped
SESCLU:   [09/07/11 17:47:58] [0000]     SescLu - QueryContentSeqData() Moniker: {C60DC234-65F9-4674-94AE-62158EFCA433} Requested Sequence: 110906024
:

Obviously I've included what I'm guessing are the salient entries as there's a lot of output, I can post more if required. I'm hoping that running pvlsvr -console is equivalent to what's happening when it's started as a service and that this isn't a red herring.

Any suggestions as to what to try next gratefully received.

sbora: The logged output suggests to me that it is communicating with the device correctly. (The server is remote so I'll have to plan a visit if I need to test this).

[CraigV]

...mmm...you might want to either look at repairing the BEDB using BEutility.exe, or running the cluster configuration wizard on that particular cluster again.