Lazy Exception // Partial Collection Failure // Cisco Prime 3.10.6 VM

[JB Houston 404]

Good morning Cisco Community, 

Can anyone advise on the following?

Issue: Cisco Prime Infrastructure 3.10.6 Virtualised and operating fine with one exception.  Collection Archive doesn't work for our devices and leaves a "Partial Collection" error.  Upon reviewing the Job Dashboard and the error status it says "Lazy Exception...".

What is weird however is, if I delete the device and add it back in again, verify credentials and manually invoke a Scheduled Collection it does complete successfully.  Subsequent Scheduled Archive then fail again with the same error.

Due to a move of 3rd party support provider, Cisco won't accept a TAC case on this at the moment without novation of serial numbers but this is a fully-patched and fully upgraded/supported virtualised installation of 3.10.6 as previously advised to do by Cisco.

Anyone else seen similar issues with Config Archives failing post-upgrade to 3.10?

Thanks in advance folks!

Kindest Regards, 

JB 

[marce1000]

  - @JB Houston 404   Never seen that one before ; if Prime was  upgraded on a VM it  usually takes more resources; I would invest CPU and Memory usage of the Prime VM with tools from the hypervisor and increase those resources if needed ,

  M.

[marce1000]

  - @JB Houston 404   I would also advise to take a look at : https://www.cisco.com/c/en/us/td/docs/net_mgmt/prime/infrastructure/3-8/admin/bk_CiscoPrimeInfrastructure_3_8_AdminGuide/bk_CiscoPrimeInfrastructure_3_8_AdminGuide_chapter_0100.html

                                 Like for instance : https://www.cisco.com/c/en/us/td/docs/net_mgmt/prime/infrastructure/3-8/admin/bk_CiscoPrimeInfrastructure_3_8_AdminGuide/bk_CiscoPrimeInfrastructure_3_8_AdminGuide_chapter_0100.html#task_1074744
                                       Probably useful to try,

  M.

[JB Houston 404]

Thanks buddy, i'll take a peek and see if anything stands out.  The bit that's puzzling me is that if a device is removed/readded, it will do the first Config Archive fine so this tells me the Credential Profiles, SNMP, Telnet/SSH, user account and client configuration is all OK.  It's just a puzzle and i'm not technical when it gets to exception errors in backend DBs.  I really dont know if i'm looking at a cause or a symptom with those errors.  Only thing I can think of is that we completely rebuilt the platform during the upgrade and were (for around 2-3 weeks) running on the Secondary CPI Server.  We had uncovered all sorts of bugs like the /var folder going wild, etc.  Since then our organisation has changed it's support provider and the support agreements haven't yet novated across so we're having difficulty getting in front of TAC for a concise root-cause analysis and solution.  Appreciate all the help and suggestions!

[marce1000]

  - @JB Houston 404    Execute these CLI commands :    1) ncs stop
                                                                                        2) ncs cleanup
                                                                                        3) ncs start

                                    Because the LazyInitializationException errors point to an issue with Prime when accessing it's database
                                    and the 'ncs cleanup' command will cleanup the database ,

  M.

[wajidhassan]

Hi @JB Houston 404 ,

Yes, this behavior with “Lazy Exception” and Partial Collection errors in Cisco Prime 3.10.x isn’t uncommon — especially after upgrades. It’s typically related to stale job data, credential sync issues, or archive process bugs that linger even in patched versions.

Since re-adding the device temporarily resolves the issue, you might try the following workaround:

• Go to Admin > System Settings > Inventory > Collection Settings and lower the collection concurrency.

• Ensure the Config Archive job history is cleared, and restart the Config Archive Daemon (via CLI or GUI).

• Also, verify SNMP/TFTP/SSH credentials are consistent and not timing out.

• If you have custom archive scripts, try disabling them temporarily.

Unfortunately, full resolution often requires TAC or database cleanup, but these steps can help stabilize the behavior in the meantime.

Hope this helps a bit until the support transfer is sorted out.