Bug 873192

Created attachment 639059 [details]
cpk-win2k8-32-network-43.cpk

Please refer to the attached for the cpk file.

Best Regards,
Dawn

We cannot find the exact cause of failure from the logs...

Do you see that the drivers are correctly installed on the VM?

Best regards,
Yan.

In any case - please retest with: http://download.devel.redhat.com/brewroot/packages/virtio-win-prewhql/0.1/44/win/

(this is also a fix for https://bugzilla.redhat.com/show_bug.cgi?id=873128 )

(In reply to comment #3)
> We cannot find the exact cause of failure from the logs...
> 
> Do you see that the drivers are correctly installed on the VM?
Yes, the drivers are correctly installed on the VM, I retest again today with keeping eyes on the guests,find when the NDISTest Client is doing the last job of "postconfig", the NDISTest Server side finish "postconfig" job before Client and quickly quit.But the Client is still running "postconfig". 

And another thing is job "InvalidPackets.htm" fail for NDIS6.5, please refer to the attachment "InvalidPacketsFail.png"

Any way, I'll retest it on new driver 44, and update the result.

Thanks
Best Regards,
Dawn
> 
> Best regards,
> Yan.

Created attachment 645562 [details]
InvalidPacketsFail

Is it possible to run same test with REHL 6.3 host?

Thanks,
Yan.

(In reply to comment #7)
> Is it possible to run same test with REHL 6.3 host?
> 
> Thanks,
> Yan.

ongoing now

(In reply to comment #6)
> Created attachment 645562 [details]
> InvalidPacketsFail

Can you double click on the test in the UI list and it should open the test log. Please save it and attach to BZ.

Thanks,
Yan.

(In reply to comment #9)
> (In reply to comment #6)
> > Created attachment 645562 [details]
> > InvalidPacketsFail
> 
> Can you double click on the test in the UI list and it should open the test
> log. Please save it and attach to BZ.
> 
> Thanks,
> Yan.

(In reply to comment #9)
> (In reply to comment #6)
> > Created attachment 645562 [details]
> > InvalidPacketsFail
> 
> Can you double click on the test in the UI list and it should open the test
> log. Please save it and attach to BZ.
> 
> Thanks,
> Yan.

Hi Yan,

Please refer to the attachment of "InvalidPackets.htm" for details.

Best Regards,
Dawn

Created attachment 653272 [details]
InvalidPackets.htm

(In reply to comment #10)
> (In reply to comment #9)
> > (In reply to comment #6)
> > > Created attachment 645562 [details]
> > > InvalidPacketsFail
> > 
> > Can you double click on the test in the UI list and it should open the test
> > log. Please save it and attach to BZ.
> > 
> > Thanks,
> > Yan.
> 
> (In reply to comment #9)
> > (In reply to comment #6)
> > > Created attachment 645562 [details]
> > > InvalidPacketsFail
> > 
> > Can you double click on the test in the UI list and it should open the test
> > log. Please save it and attach to BZ.
> > 
> > Thanks,
> > Yan.
> 
> Hi Yan,
> 
> Please refer to the attachment of "InvalidPackets.htm" for details.
> 
> Best Regards,
> Dawn

Thanks!

Any news on using RHEL6.3 host?

Bridge filtering settings:

Good settings are:
# sysctl -a|grep bridge
net.bridge.bridge-nf-call-arptables = 0 
net.bridge.bridge-nf-call-iptables = 0 
net.bridge.bridge-nf-call-ip6tables = 0

Hi Yan,

Tried to run the same test with rhel6.3 host, still hit the same issue with rhel6.4 host.

Best Regards,
Dawn

(In reply to comment #5)
> (In reply to comment #3)
> > We cannot find the exact cause of failure from the logs...
> > 
> > Do you see that the drivers are correctly installed on the VM?
> Yes, the drivers are correctly installed on the VM, I retest again today
> with keeping eyes on the guests,find when the NDISTest Client is doing the
> last job of "postconfig", the NDISTest Server side finish "postconfig" job
> before Client and quickly quit.But the Client is still running "postconfig". 
> 
> And another thing is job "InvalidPackets.htm" fail for NDIS6.5, please refer
> to the attachment "InvalidPacketsFail.png"
> 

I don't think it is a problem. Postconfig is a local task of removing the protocol driver and reconfiguring the VM if the test did something with it. It is expected for the server to quit. Also the logs are collected from the client side only.


> Any way, I'll retest it on new driver 44, and update the result.
> 
> Thanks
> Best Regards,
> Dawn
> > 
> > Best regards,
> > Yan.

Please check the event log for traces of the system crash.

Thanks,
Yan.

(In reply to comment #18)
> Please check the event log for traces of the system crash.
> 
> Thanks,
> Yan.

Hi yan,

Please refer to the attached "system-event-log.evtx" for logs.

Best Regards,
Dawn

Created attachment 657326 [details]
system-event-log

Hi Dawn,

I see in the event log following errors: 
"The server could not bind to the transport \Device\NetBT_Tcpip_{003E256F-3421-468A-BB7B-AC6F5EFAE7DA} because another computer on the network has the same name.  The server could not start." (computer name "2k8-32-1-46.virtio.com").


Considering that the DTM uses computer names - this might cause some issues. Can you double check that the VMs used for test and possibly others VMs in the network don't have same name.

Thanks,
Yan.

Hi Dawn,

I see in the event log following errors: 
"The server could not bind to the transport \Device\NetBT_Tcpip_{003E256F-3421-468A-BB7B-AC6F5EFAE7DA} because another computer on the network has the same name.  The server could not start." (computer name "2k8-32-1-46.virtio.com").


Considering that the DTM uses computer names - this might cause some issues. Can you double check that the VMs used for test and possibly others VMs in the network don't have same name.

Thanks,
Yan.

(In reply to comment #22)
> Hi Dawn,
> 
> I see in the event log following errors: 
> "The server could not bind to the transport
> \Device\NetBT_Tcpip_{003E256F-3421-468A-BB7B-AC6F5EFAE7DA} because another
> computer on the network has the same name.  The server could not start."
> (computer name "2k8-32-1-46.virtio.com").
> 
> 
> Considering that the DTM uses computer names - this might cause some issues.
> Can you double check that the VMs used for test and possibly others VMs in
> the network don't have same name.
> 
> Thanks,
> Yan.

Hi Yan, 

Yes, actually I saw the logs yesterday, and felt strange since we named the guest as "system name + driver name + build Number + NIC1/2", so shouldn't have the duplicate name,and in the domain if the name duplicate, system will prompt you computer name duplicate. So I still checked two times, didn't find the same name exist for all guest (win2k8-32/64 win2k8-R2).

Best Regards,
Dawn

Since this could be a setup problem (transport binding failure), please try the following:
    1. Close all ndis-test related applications running on HCK clients
    2. Uninstall NetKVM driver for all RedHat adapters on all clients(Device Manager --> Rings Click --> Uninstall --> Delete driver software)
    3. Reboot all clients
    4. Install drivers again for all Redhat adapters on all clients
    5. Reboot all clients
    6. Try to repeat this test

We saw a few similar problems, this sequence usually helps.

(In reply to comment #24)
> Since this could be a setup problem (transport binding failure), please try
> the following:
>     1. Close all ndis-test related applications running on HCK clients

We are asked to Run netkvm jobs on WLK ,Could you confirm whether the following steps works on WLK as well ?

Mike
>     2. Uninstall NetKVM driver for all RedHat adapters on all clients(Device
> Manager --> Rings Click --> Uninstall --> Delete driver software)
>     3. Reboot all clients
>     4. Install drivers again for all Redhat adapters on all clients
>     5. Reboot all clients
>     6. Try to repeat this test
> 
> We saw a few similar problems, this sequence usually helps.

(In reply to comment #25)
> (In reply to comment #24)
> > Since this could be a setup problem (transport binding failure), please try
> > the following:
> >     1. Close all ndis-test related applications running on HCK clients
> 
> We are asked to Run netkvm jobs on WLK ,Could you confirm whether the
> following steps works on WLK as well ?
> 

Yes, these steps should work on WLK as well.

> Mike
> >     2. Uninstall NetKVM driver for all RedHat adapters on all clients(Device
> > Manager --> Rings Click --> Uninstall --> Delete driver software)
> >     3. Reboot all clients
> >     4. Install drivers again for all Redhat adapters on all clients
> >     5. Reboot all clients
> >     6. Try to repeat this test
> > 
> > We saw a few similar problems, this sequence usually helps.

Please try the steps in comment #24

(In reply to comment #27)
> Please try the steps in comment #24

Hi Yan,

Tried with the steps in comment #24 to re-install the netkvm driver, still hit the same issue.

Best Regards,
Dawn

Looks this bug should be filter by https://sysdev.microsoft.com/en-US/Hardware/EC/ECDetails.aspx?id=2966 

I will research why filter does not help.

Retest it on Build49 on win2k8R2 w/ the driver located in win7
I don't think "InvalidPackages" is the key point cause job failed ,it can be filter by https://sysdev.microsoft.com/en-US/Hardware/EC/ECDetails.aspx?id=2966  from my test 


The job failed due to "Make Sure communication helpers stats and adapter stats match" sub job 

Following is the logs :

Failed 
Start Test 12/12/2012 1:06:57.656 PM Make sure communication helper stats and adapter stats match. 
Error 12/12/2012 1:06:57.656 PM Adapter statistics and comm helper statistics do not match! 
File:    Line: 0 
Error Type:   WIN32 
Error Code:   0x50012 
Error Text:   There are no more files. 
End Test 12/12/2012 1:06:57.656 PM Make sure communication helper stats and adapter stats match. 
Result:   Fail 
Repro:   ndistest.exe /logo /auto /client /target:Miniport /tc:{653340A5-E5F0-424E-B98D-C59169DFD541} /TestEnvInfo:TestEnvInfo.xml /support:{F64CE5E3-56C5-4810-BB46-225AC67D6F3E},{653340A5-E5F0-424E-B98D-C59169DFD541} /msg:{ACC8DD1F-8734-4CA9-8352-44218376E597} /jobgroup:WHQL\Lan /ParaVirtualizationDriver:True

(In reply to comment #34)
> Retest it on Build49 on win2k8R2 w/ the driver located in win7
> I don't think "InvalidPackages" is the key point cause job failed ,it can be
> filter by
> https://sysdev.microsoft.com/en-US/Hardware/EC/ECDetails.aspx?id=2966  from
> my test 
> 
> 
> The job failed due to "Make Sure communication helpers stats and adapter
> stats match" sub job 
> 
> Following is the logs :
> 
> Failed 
> Start Test 12/12/2012 1:06:57.656 PM Make sure communication helper stats
> and adapter stats match. 
> Error 12/12/2012 1:06:57.656 PM Adapter statistics and comm helper
> statistics do not match! 
> File:    Line: 0 
> Error Type:   WIN32 
> Error Code:   0x50012 
> Error Text:   There are no more files. 
> End Test 12/12/2012 1:06:57.656 PM Make sure communication helper stats and
> adapter stats match. 
> Result:   Fail 
> Repro:   ndistest.exe /logo /auto /client /target:Miniport
> /tc:{653340A5-E5F0-424E-B98D-C59169DFD541} /TestEnvInfo:TestEnvInfo.xml
> /support:{F64CE5E3-56C5-4810-BB46-225AC67D6F3E},{653340A5-E5F0-424E-B98D-
> C59169DFD541} /msg:{ACC8DD1F-8734-4CA9-8352-44218376E597} /jobgroup:WHQL\Lan
> /ParaVirtualizationDriver:True

Re-run this job on 2k8R2 5 times ,one time passed .
win2k8-32 passed as well ,I will run this job on win2k8-64 Once the win2k8R2 netkvm whql test finished .

Best Regards,
Mike

(In reply to comment #36)

> Re-run this job on 2k8R2 5 times ,one time passed .
> win2k8-32 passed as well ,I will run this job on win2k8-64 Once the win2k8R2
> netkvm whql test finished .
> 
> Best Regards,
> Mike

Did not hit this issue on build49 over win7 32/64 ,win2k8 32/64/R2 bit.
There must be a configuration issue on QE AD testing environment ,I will re-setup it after all block&netkvm job submit to MSFT.

Closing this one

Also this erratum will probably help: https://sysdev.microsoft.com/en-US/Hardware/EC/ECDetails.aspx?id=2966