No connection from Management Pack for storage devices to Cisco DCNM

[Jetix]

Hi there,

Is there anybody outthere who has successfully managed to create a connection from vROps 7 with Management Pack for Storage Devices Version 7 to Cisco DCNM 11.x? Any help would be appreciated.

Regards,

Marc

[GayathriS]

Just a question out of curiosity , did you check if vrops 7.0 supports this management pack of Cisco ?

Regards

Gayathri

[GayathriS]

I did check this from vrops compatibility guide and I dont see storage management pack for CISCO and vrops 7.0 :

Please consider marking this answer as correct and helpful if this helps you.

regards

GS

[Jetix]

Here is the link to the product documentation:

VMware Solution Exchange

And a printscreen, which says that you can discover storage switches from Cisco DCNM:

On the other hand: the connection from the Management Pack for Storage Devices to Brocade Network Advisor works just perfect.

[GayathriS]

Cool thanks for the details.

Could you let me know where you are stuck and what errors are you getting ?

regards
Gayathri

[Jetix]

When I try to create a new "Fabric Server Adapter" with the address "https://fqdn_of_dcnm:5989/cimv2", as described in the user guide "vRealize Operations Management Pack for Storage Devices 7.0 Guide", I get the following error message:

Inside the logfile I get the following (and belive me, it's not a certificate issue):

FabricServer.Trust.has.not.been.established..Connection.will.not.take.place.=Trust has not been established. Connection will not take place.
FabricServer.No.trusted.certificates.available.=No trusted certificates available.
FabricServer.Use.correct.username.and.password.and.provide.a.trusted.certificate.=Use correct username and password and provide a trusted certificate.
FabricServer.Adapter.instance.has.been.configured.to.trust.multiple.certificates,.when.only.one.is.allowed..Please.remove.any.old,.unneeded.certificates.and.try.again.=Adapter instance has been configured to trust multiple certificates, when only one is allowed. Please remove any old, unneeded certificates and try again.
FabricServer.Configure.Failed..No.trusted.certificates.available.=Configure Failed. No trusted certificates available.
FabricServer.Failed.to.invoke.WBEM.operations.against.a.WBEM.Server=Failed to invoke WBEM operations against a WBEM Server

I already opened a SR at VMware, unfortunately they are not even familiar with their own management packs.

[GayathriS]

Error message is talking about certificate not trusted.

What certs are you using on vrops ?

are they not expired , they are valid.

regards

GS

[Jetix]

We use public trusted certificates by SwissSign. We also created a test environment and left all certificates unchanged, without success. VMware support confirmed, that there are no certificate issues. It has something to do with WBEM on the target site. But since VMware probably has also tested the whole thing with DCNM, they should just tell me how to get it to work. Are you using the "Management Pack for Storage Devices" to communicate with Cisco DCNM?

[tristankelkerma]

Hello,

Have you found a solution ? We have the same issue here.

[Jetix · Accepted solution]

Hi folks...

During the last weeks/months I did a lot of troubleshooting regarding MPSD (Management Pack for Storage Devices). So the main points were:

• Namespace detection
  First, I enabled debug-logging for the MPSD, which showed that the MPSD always wanted to connect to a namespace called /CIMv2. The problem is, that Cisco DCNMs default namespace is /interop. Once I modified the properties.file of the MPSD to /interop, I was able to establish a connection from the MPSD to Cisco DCNM. Unfortunately, the error message regarding WBEM didn't disappear.
• WBEM
  Usually, WBEM (Web-based enterprise management) runs per RFC-Definition on UDP-427. Cisco DCNM has no processes listening on UDP-427. After checking with Cisco they confirmed that SMI-S and WBEM are tied together and run on port TCP-5589, which is in fact SMI-S. But MPSD requires WBEM, as it is hard coded.
• Performance
  The time need for a full collection cycle was around one hour with only a couple of SAN switches.

To cut a long story shot: I went back to VMware and was connected to the main-developer of the MPSD. He's a really nice guy and was very helpful. From his side, he improved the MPSD with code modifications (correct namespace detection, disable WBEM advertising for Cisco DCNM, performance optimizations, improving API-calls, etc.). From my side, I did a lot of testing with their newer builds, troubleshooting and log-studies:smileylaugh:... However, I'm a bit proud to announce that VMware has finally fixed it. This is the way, how customers and suppliers should work together. To whom it may concern: thank you very much, great job, well done, keep it up!!!

Regards,

Jetix

Update: VMware has released Management Pack for Storage Devices 7.5.0, Build 13462779, which should provide compatibility to collect data from Cisco DCNM. Unfortunately, there was an issue with "Switch Port Statistics" collection, so that only one switch out of N provided these metrics. However, the issue is fixed and if you are affected you should immediatly request the hotfix from VMware. The hotfix has the following build number: 13672434