Nexus Log-Message: "unkown enum 4 - xmlsa"

[mullzkBern_2]

Hi everybody

Our Nexus 5000 produce a lot of Log-Messages which we do not understand. Almost hourly (but not on a exact schedule), we get an entry:

"%USER-2-SYSTEM_MSG: unknown enum 4 - xmlsa"

System Reference, Google and a search on Support Community do not bring up anything. Has anybody an idea what this Message means and what we could do against it?

Issue occurs on all our Nexus5548 and Nexus5596, running NX-OS 5.1(3)N1(1a).

Using DCNM 5.2.2c and LMS 4.0.1 (although the N5K are not managed in LMS)

Thanks in advance and greetings from Berne, Switzerland
Stefan Mueller

[Ivan Shirshin]

Hi Stefan,

Do you know if there are any show command executed at that time? There are several bugs related to show command or DCNM device discovery that were fixed in 5.2(1)N1:

CSCtz46192    Nexus 5k - show interface status returns unknown enum value for fex int

CSCtk10449    XML is not implemented properly for STP commands.

That errors do not have service impact and only mean that the XML agent is not able to properly get data for XML responses.

Kind Regards,
Ivan

**Please grade this post if you find it useful.

[mullzkBern_2]

Hi Ivan.

There are no Users executing show commands at that time - but its possible that DCNM does so.

I'll think about upgrading to 5.2(1)N1.

Thanks for the quick Reply, Kind Regards

Stefan

[Ivan Shirshin]

welcome, Stefan.

[peterjhollis · Accepted solution]

FYI:

I have been experiencing this error since installing our new environment. A side effect of these errors on the 5k's was the logging levels resetting to their default values.

I pinpointed the cause of the error with DCNM and user accounts. If I ran the DCNM client on more than one workstation using the same login credentials, within 15 minutes, all the 5k's would produce this enum error. Creating individual users accounts for each monitoring workstation and different accounts for each user the enum error has not appeared.

Tested by logging the same account on two seperate workstations and the enum error appeared on the 5k's as expected.

Hope this helps..

Regards

Peter

DCNM Ver 5.2(2e),S7

[mullzkBern_2]

Hi Peter

Thx for this one, most probably you are right as the problem occured when we tested DCNM and did not occur since. The Test-Installation of DCNM used the same user account as the LMS.

Greetings from Berne, Stefan

[peterjhollis]

Right, think I have got to the bottom of this!

DCNM requires a minimum level for logging features set on all managed devices. If the logging levels are set to low, DCNM will attempt to reset the logging levels and reports this error. DCNM sucessfully changes the logging levels on the devices within the running config but the startup config remains unchanged.

What I found is that if the logging levels are changed during device discovery and if you get onto the physical device with DCNM running and reset the logging levels to what you want, it causes DCNM to falter in several places including setting devices as unmanaged and.or reporting all kinds of server errors.

I have enabled all features listed in the default logging levels and set to the required level and have not had a problem for 7 days now.

The only problem I have now is setting the DCNM Server Administrator/DCNM Server Log Settings to default 'Warning' level but the sys.pipe log file still records everything.

Hope the above makes sense...

Peter