This discussion has been locked. The information referenced herein may be inaccurate due to age, software updates, or external references.
You can no longer post new replies to this discussion. If you have a similar question you can start a new discussion in this forum.

Network Device serial not updated in NCM

Hi All,

   

we have on network switch replaced by network team. changed the community strings from SNMPV2C to V3 after that run the list resources for this device in console. 

added the NCM and test is successfully. machine type updated but serial number not updated in NCM tab. can you please advise how to overcome the situation. 

please provide the solutions 

Version is orion platfrom 2020.2.HF2 

  

  • I've had similar issues in the past.  They were caused by using the same machine name and management IP address.  I don't know why, but my NPM solution never did a proper full and complete rediscovery if I kept the same IP address & device name on a replacement. 

    Two work-arounds for you:

    1.  Copy the machine name, management IP address, and management credentials (including snmp string, etc.).  Then delete the node from NPM.  Finally, re-add the node as if it were totally new to your network, entering in the copied management IP address & credentials.  You should be able to correctly discover the serial number then, as well as other things that don't properly get rediscovered when hardware is replaced while keeping the same management IP address & credentials.

    or 2.  Give the replacement switch a new name & new IP address.  Add it to NPM as a new node.  Delete the previous node.

    This isn't how I I expect NPM to perform.  I've done many forklift upgrades, going from Cisco earlier switches to more current switch models, and they all end up leaving multiple things improperly discovered by NPM when the original node name & management IP addresses are taken over by the replacement hardware.

    Swift packets!

    Rick Schroeder

  • I concur.   Only I usually use option 1.   

  • Thank you for quick Response. 

    But we will re add the device in cosnole. we will loose the historical data. so that is the reason we didn't re added the device. 

    we observed machine type is updated successfully but Chassis ID not updated. i don't understand. 


    Regards,

    Kishore. 

  • Agreed, the solution isn't satisfactory because it causes you to lose historical data for that switch.  You may be able to find someone in Thwack, or via Solarwinds Technical Support, who can offer you a solution that uses SWQL to dump only some of the data associated with the old switch, thus allowing NPM to poll and learn the new switch completely.

    I never found that process.  Since current data is more important to me than an uninterrupted and linear set of historical data, I pulled the historical data I wanted, made screen shots of it for future reference, and then deleted the old node and re-added it to properly discover the new node's hardware properties.

    I'm sorry I don't have a different solution for you.  It seems, for now, that you can have inaccurate current data and a stream of historical data, or accurate current hardware and lose your easy access to historical data.  Given that my organization did not have an interest in retaining historical port statistics for more than a set number of weeks, it wasn't a very big sacrifice to delete the node and start rebuilding that data again.

    An alternative I considered:

    1.  Give the replacement switch a new name & IP address and start monitoring & discovering it.

    2.  Stop polling the old switch's name & IP address.  You'll keep the historical data for as long as the period you've set for retaining history (up to two years).  And you'll be building a new set of historical data using the new switch's name & IP address.

    3.  Once the time that you retain historical data has passed for the old switch, delete the old switch.

    If you DO find a solution that properly retains the data AND that correctly dumps the old hardware info and re-learns it, that would be good to share here.

  • Interesting, I'd not noticed this but I've just replaced a bunch of switches with 9300's and installed some new 9300's and none are showing the serial number. So even brand new ones added to NPM don't show the serial number. I assume something has changed with the MIB variable. I'll have to have a play around to see if I can get it to show.

    And for the record, the 9500 also doesn't show the serial number either as a replacement or a newly added node into NPM.

  • I fixed my issue. I realised my Nightly Hardware Inventory had expired. So I just manually ran a Hardware Inventory job over 2 x 9300's and 2 x 9500's (both changed and newly installed devices). Now all switches show the serial number in NCM.

    But as you can see in the second image, this 9300 is still pulling old EOL info that was for the 2960 it replaced. So I'll have to see if that disappears or not.

  • So overnight I ran the NCM Nightly Hardware Inventory and now all my 9300/9500's are showing their serial number in NCM. For the ones that I replaced, they were still showing old EOL data. So I manually deleted that information for those nodes.

    All good now Slight smile