Cannot backup Virtual Machines in a CSV with DPM 2010

  • Follow


I have a 4 Node Server 2008 R2 Cluster with 26 HA Virtual Machines in 1 CSV 
volume.  When I create the protection group and start the initial replica, 
DPM 2010 Beta only grabs a couple of megabytes (anywhere from .19 to 15.95) 
then just stays there until I cancel the job.

I am using software based VSS (built in Microsoft version), and have enabled 
CSV serialization on the cluster.  My storage for the CSV is a Dell 
Equallogic iSCSI SAN.

When i look at the Event/Application logs all i see is this:

Revert operation for volume 
\\?\Volume{f7f8c65d-c32a-11de-8e8b-001d09645911}\ has begun.  Volume is being 
reverted to the shadow copy with id {55ed6ccf-72f1-4180-8869-a5bec0753ef0}. 

Any ideas?

Thanks!


0
Reply Utf 11/20/2009 7:50:01 PM

Did you implement the node and CSV serialization? When you are using the 
built in the provider you need to implement CSV and node serialization. This 
process is described in DPM2010BetaProtectingHyperV.docx (part of the 
documentation). The reasons for doing this is mentioned in that document.

-- 
Thanks
Shyama Hembram
This posting is provided "AS IS" with no warranties, and confers no rights.


"Sal Icban" <Sal Icban@discussions.microsoft.com> wrote in message 
news:23B0440C-7477-4230-86CD-57FCE2340C3E@microsoft.com...
> I have a 4 Node Server 2008 R2 Cluster with 26 HA Virtual Machines in 1 
> CSV
> volume.  When I create the protection group and start the initial replica,
> DPM 2010 Beta only grabs a couple of megabytes (anywhere from .19 to 
> 15.95)
> then just stays there until I cancel the job.
>
> I am using software based VSS (built in Microsoft version), and have 
> enabled
> CSV serialization on the cluster.  My storage for the CSV is a Dell
> Equallogic iSCSI SAN.
>
> When i look at the Event/Application logs all i see is this:
>
> Revert operation for volume
> \\?\Volume{f7f8c65d-c32a-11de-8e8b-001d09645911}\ has begun.  Volume is 
> being
> reverted to the shadow copy with id 
> {55ed6ccf-72f1-4180-8869-a5bec0753ef0}.
>
> Any ideas?
>
> Thanks!
>
> 
0
Reply Shyama 11/23/2009 6:00:46 AM


Hi,

Yes I implemented both Node and CSV Serialization and it still gets stuck 
after a couple of megabytes.  After enabling, i recreated the protection 
group and added 2 VM's for testing.  The initial replica is getting stuck.

Thanks

"Shyama Hembram[MSFT]" wrote:

> Did you implement the node and CSV serialization? When you are using the 
> built in the provider you need to implement CSV and node serialization. This 
> process is described in DPM2010BetaProtectingHyperV.docx (part of the 
> documentation). The reasons for doing this is mentioned in that document.
> 
> -- 
> Thanks
> Shyama Hembram
> This posting is provided "AS IS" with no warranties, and confers no rights.
> 
> 
> "Sal Icban" <Sal Icban@discussions.microsoft.com> wrote in message 
> news:23B0440C-7477-4230-86CD-57FCE2340C3E@microsoft.com...
> > I have a 4 Node Server 2008 R2 Cluster with 26 HA Virtual Machines in 1 
> > CSV
> > volume.  When I create the protection group and start the initial replica,
> > DPM 2010 Beta only grabs a couple of megabytes (anywhere from .19 to 
> > 15.95)
> > then just stays there until I cancel the job.
> >
> > I am using software based VSS (built in Microsoft version), and have 
> > enabled
> > CSV serialization on the cluster.  My storage for the CSV is a Dell
> > Equallogic iSCSI SAN.
> >
> > When i look at the Event/Application logs all i see is this:
> >
> > Revert operation for volume
> > \\?\Volume{f7f8c65d-c32a-11de-8e8b-001d09645911}\ has begun.  Volume is 
> > being
> > reverted to the shadow copy with id 
> > {55ed6ccf-72f1-4180-8869-a5bec0753ef0}.
> >
> > Any ideas?
> >
> > Thanks!
> >
> > 
> .
> 
0
Reply Utf 11/23/2009 7:11:02 PM

Sal,

Can you try running backup for any other data source and see if the job 
hangs in that case as well? You could try protecting even a few files on a 
non system drive.

Also, can you create and share the following dumps at the time when your 
respective job is hung. Hyper-V Server here is the one that hosts the VM at 
the time of backup.
1.	dpmra crash dump on Hyper-V side
2.	dpmra crash dump on DPM side
3.	msdpm crash dump on Hyper-V side.

You can create these by locating the respective processes in Task Manager.

Also the respective error logs:

dpmracurr.errlog on the hyper-V server under C:\Program Files\Microsoft Data 
Protection Manager\DPM\Temp
dpmracurr.errlog and msdpmcurr.errlog on the DPM server under C:\Program 
Files\Microsoft DPM\DPM\Temp

In case you do not have a shared location to place these, send me an email 
at ansingh at microsoft dot com.

Thanks,
--Angad.


-- 
This posting is provided “AS IS” with no warranties, and confers no rights

"Sal Icban" <SalIcban@discussions.microsoft.com> wrote in message 
news:A88A64EC-BC23-4A3A-9631-61175C4BF014@microsoft.com...
> Hi,
>
> Yes I implemented both Node and CSV Serialization and it still gets stuck
> after a couple of megabytes.  After enabling, i recreated the protection
> group and added 2 VM's for testing.  The initial replica is getting stuck.
>
> Thanks
>
> "Shyama Hembram[MSFT]" wrote:
>
>> Did you implement the node and CSV serialization? When you are using the
>> built in the provider you need to implement CSV and node serialization. 
>> This
>> process is described in DPM2010BetaProtectingHyperV.docx (part of the
>> documentation). The reasons for doing this is mentioned in that document.
>>
>> -- 
>> Thanks
>> Shyama Hembram
>> This posting is provided "AS IS" with no warranties, and confers no 
>> rights.
>>
>>
>> "Sal Icban" <Sal Icban@discussions.microsoft.com> wrote in message
>> news:23B0440C-7477-4230-86CD-57FCE2340C3E@microsoft.com...
>> > I have a 4 Node Server 2008 R2 Cluster with 26 HA Virtual Machines in 1
>> > CSV
>> > volume.  When I create the protection group and start the initial 
>> > replica,
>> > DPM 2010 Beta only grabs a couple of megabytes (anywhere from .19 to
>> > 15.95)
>> > then just stays there until I cancel the job.
>> >
>> > I am using software based VSS (built in Microsoft version), and have
>> > enabled
>> > CSV serialization on the cluster.  My storage for the CSV is a Dell
>> > Equallogic iSCSI SAN.
>> >
>> > When i look at the Event/Application logs all i see is this:
>> >
>> > Revert operation for volume
>> > \\?\Volume{f7f8c65d-c32a-11de-8e8b-001d09645911}\ has begun.  Volume is
>> > being
>> > reverted to the shadow copy with id
>> > {55ed6ccf-72f1-4180-8869-a5bec0753ef0}.
>> >
>> > Any ideas?
>> >
>> > Thanks!
>> >
>> >
>> .
>> 
0
Reply Angad 11/24/2009 7:01:21 AM

Angad,

I just sent you an email.  I am not able to send you the logs via email (too 
big).  I tried running a backup of something else (random files) on a non 
system drive and it is also getting stuck.

Thanks

"Angad Pal Singh [MSFT]" wrote:

> Sal,
> 
> Can you try running backup for any other data source and see if the job 
> hangs in that case as well? You could try protecting even a few files on a 
> non system drive.
> 
> Also, can you create and share the following dumps at the time when your 
> respective job is hung. Hyper-V Server here is the one that hosts the VM at 
> the time of backup.
> 1.	dpmra crash dump on Hyper-V side
> 2.	dpmra crash dump on DPM side
> 3.	msdpm crash dump on Hyper-V side.
> 
> You can create these by locating the respective processes in Task Manager.
> 
> Also the respective error logs:
> 
> dpmracurr.errlog on the hyper-V server under C:\Program Files\Microsoft Data 
> Protection Manager\DPM\Temp
> dpmracurr.errlog and msdpmcurr.errlog on the DPM server under C:\Program 
> Files\Microsoft DPM\DPM\Temp
> 
> In case you do not have a shared location to place these, send me an email 
> at ansingh at microsoft dot com.
> 
> Thanks,
> --Angad.
> 
> 
> -- 
> This posting is provided “AS IS” with no warranties, and confers no rights
> 
> "Sal Icban" <SalIcban@discussions.microsoft.com> wrote in message 
> news:A88A64EC-BC23-4A3A-9631-61175C4BF014@microsoft.com...
> > Hi,
> >
> > Yes I implemented both Node and CSV Serialization and it still gets stuck
> > after a couple of megabytes.  After enabling, i recreated the protection
> > group and added 2 VM's for testing.  The initial replica is getting stuck.
> >
> > Thanks
> >
> > "Shyama Hembram[MSFT]" wrote:
> >
> >> Did you implement the node and CSV serialization? When you are using the
> >> built in the provider you need to implement CSV and node serialization. 
> >> This
> >> process is described in DPM2010BetaProtectingHyperV.docx (part of the
> >> documentation). The reasons for doing this is mentioned in that document.
> >>
> >> -- 
> >> Thanks
> >> Shyama Hembram
> >> This posting is provided "AS IS" with no warranties, and confers no 
> >> rights.
> >>
> >>
> >> "Sal Icban" <Sal Icban@discussions.microsoft.com> wrote in message
> >> news:23B0440C-7477-4230-86CD-57FCE2340C3E@microsoft.com...
> >> > I have a 4 Node Server 2008 R2 Cluster with 26 HA Virtual Machines in 1
> >> > CSV
> >> > volume.  When I create the protection group and start the initial 
> >> > replica,
> >> > DPM 2010 Beta only grabs a couple of megabytes (anywhere from .19 to
> >> > 15.95)
> >> > then just stays there until I cancel the job.
> >> >
> >> > I am using software based VSS (built in Microsoft version), and have
> >> > enabled
> >> > CSV serialization on the cluster.  My storage for the CSV is a Dell
> >> > Equallogic iSCSI SAN.
> >> >
> >> > When i look at the Event/Application logs all i see is this:
> >> >
> >> > Revert operation for volume
> >> > \\?\Volume{f7f8c65d-c32a-11de-8e8b-001d09645911}\ has begun.  Volume is
> >> > being
> >> > reverted to the shadow copy with id
> >> > {55ed6ccf-72f1-4180-8869-a5bec0753ef0}.
> >> >
> >> > Any ideas?
> >> >
> >> > Thanks!
> >> >
> >> >
> >> .
> >> 
0
Reply Utf 11/24/2009 6:29:01 PM

4 Replies
1029 Views

(page loaded in 0.107 seconds)


Reply: