Prime Infrastructure 2.0: Running an inventory sync removes "snmp-server location" in running-config

[bastianbrachner]

Hello,

we're running Prime Infrastructure 2.0 (version VA-2.0.0.0.294).

I was scared as I figured out one hour ago that running an inventory sync (Device Work Center) removes the command snmp-server location and adds the command snmp-server host "PI-ip address" "snmp rw community" in the running-config.

Has someone the same behaviour or can reenact this scenario?!

Can someone imagine that this behaviour has something to do with a configuration setting in PI?

I'm perplexed and got gray hair now.

Thank you for your answers or tips.

Bastian

[AFROJ AHMAD · Accepted solution]

Hi,

check my below post:

https://supportforums.cisco.com/message/4057630#4057630

Thanks-

Afroz

[Do rate the useful post]

[Marvin Rhoads]

Interesting, I am using PI 2.0 and have not seen this bug mainfest, either with SNMP v3 or v2c.

[Michel Hegeraat]

This is because most of us don't use a RW community Marvin.

It is fixed in the update for prime 2.0 that has now becoma available.

Prime can download the update itself, but you can also download a zip the contain the update.

It is available on the CCO download patches.

Cheers,

Michel

[bastianbrachner]

Hello Afroz,

thank you for your answer. Your/Barts post describes my incident.

Sorry but my search in the discussions did not figure out Barts discussion post.

I'll contact TAC, hopfully they have a bugfix which can be installed.

I don't want to edit each device credential regarding the snmp community of nearly 2k devices.

Is it possible to edit the snmp community for all devices within one step?

I'll rate your answer when I've the confirmation of the TAC.

Best Regards

Bastian

[Rob Johnson]

Bastian,

As Michael mentioned, the fix for the issue where the SNMP location gets set back to a blank is fixed in the update available under Prime 2.0 patches and is named pi_update_2.0-3.zip.

This patch is installed via the PI GUI under Administration> Software Updates

I had the same problem on a customer's server as well as my lab 2.0 server as well.  Installing the patch fixes it right up.

[bastianbrachner]

Hello Rob,

thanks for your answer. Should this fix be displayed in PI under Administration->Software Update like Michel posted because I get the message "Login to cisco.com was successful." but the fix is not displayed (No data available).

I've downloaded the fix manually and will install it today.

Bastian

[Rob Johnson]

I haven't tried the method to pull it down from Cisco.com.

I too just downloaded the .zip file; unzipped it to get the .ubf file; selected upload update file; selected the update then clicked install; stopped and restarted PI

[bastianbrachner]

Hello Rob,

I just successfully installed the patch. The snmp-server location is not touched anymore while running a inventory job.

But PI still enters this string in the running-configuration:

snmp-server host "PI-IP address" " our snmp RW community- string"

although there already exists the entry:

snmp-server host "PI-IP address" version 2c "community-string (not the snmp rw-community-string)"

Do you or anyone else has the same behaviour that PI adds automaically himself as snmp trap receiver with the SNMP-RW community?

Regards

Bastian

[Rob Johnson]

Bastian,

I'll check this out from my lab box in the morning then will let you know.

.

[J. Aske · Accepted solution]

In PI2.0 this gets set under the Inventory Collection setting.

Goto Administration > System Settings

Choose Inventory on left side panel

Uncheck the option "Enable Syslog and Traps on a device"