Upgrading Prime Infrastructure

[stephendrkw]

hi all!

I upgraded my Prime Infrastructure VE from 1.4 to 3.0.0.278 (ultimate goal is to go to 3.5.x)

process was build a new VMWare server install Prime 2.2 ova, import backup from 1.4 - this worked successfully!

next step I upgraded from Prime 2.2 to 3.0.0.278 inline, this upgrade worked successfully!

next step upgrade from 3.0.0.278 to 3.1.0.0.132 (only 3.1 download version available on Cisco's download software site)

ncs stopped then following error appears after upgrade.......(MD5 checksum mirrored up)

/admin# application upgrade PI-Upgrade-3.0.X_to_3.1.0.0.132.tar.gz defaultRepo
Save the current ADE-OS running configuration? (yes/no) [yes] ?
Generating configuration...
Saved the ADE-OS running configuration to startup successfully

Please ensure you have a backup of the system before proceeding.
Proceed with the application upgrade ? (yes/no) [yes] ?

DO NOT press ^C while the upgrade is in progress
Aborting upgrade with a ^C may leave the system in a unrecoverable state

Initiating Application Upgrade...
Stage 1 of 7: Transferring file ...
-- complete.
Stage 2 of 7: Unpacking file ...
-- complete.
Stage 3 of 7: Executing pre-install ...

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
System will reboot a few times to complete the upgrade process.
Please monitor the upgrade progress from the system console.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

ERROR : Please run the application upgrade from the system console to
monitor upgrade progress. Use system monitor, serial terminal or a virtual console
to initiate the upgrade.

WARNING : It is strongly recommended to copy the latest backups to an
external repository before proceeding with the upgrade.

% Pre-install step failed. Please check the logs for more details.
admin#

LOGS below

[2019-02-27 11:15:07,431] [http-bio-443-exec-20] [topology] [INFO ] - XMPTSPerspectivesViewService: newProjectSessionInfo() sessionIdBB8EE47BCA2BED4B280081E2EAFBC946 tsSessionID:BB8EE47BCA2BED4B280081E2EAFBC946 projectIds:null
[2019-02-27 11:15:07,432] [http-bio-443-exec-20] [topology] [INFO ] - XMPTopologySessionApp: Adding session object with sessionId BB8EE47BCA2BED4B280081E2EAFBC946
[2019-02-27 11:15:08,696] [http-bio-443-exec-26] [topology] [INFO ] - Retrieving all the devices for group 501,506.
[2019-02-27 11:15:08,700] [http-bio-443-exec-26] [topology] [INFO ] - CommonTopologyModelPopulator: TopologyFilter:TopologyFilter=[EDGE:TopologyFilterInCondition=[type:[MANUALLINK]],]
[2019-02-27 11:15:08,870] [http-bio-443-exec-26] [topology] [INFO ] - CommonTopologyModelPopulator: Time taken to populate all Subgroups and nodes for Group:501506 is:173
[2019-02-27 11:15:08,870] [http-bio-443-exec-26] [topology] [INFO ] - AbstractTSServiceCommand: Data population completed in time:175 milliseconds
[2019-02-27 11:15:08,870] [http-bio-443-exec-26] [topology] [INFO ] - AbstractTSServiceCommand: Number of TSModelElement in XMPTSModel:8
[2019-02-27 11:15:08,870] [http-bio-443-exec-26] [topology] [INFO ] - AbstractTSServiceCommand: Number of Nodes in XMPTSModel:0
[2019-02-27 11:15:08,870] [http-bio-443-exec-26] [topology] [INFO ] - AbstractTSServiceCommand: Number of Group Nodes in XMPTSModel:8
[2019-02-27 11:15:08,870] [http-bio-443-exec-26] [topology] [INFO ] - AbstractTSServiceCommand: Number of Edges in XMPTSModel:0
[2019-02-27 11:15:08,870] [http-bio-443-exec-26] [topology] [INFO ] - AbstractTSServiceCommand: Number of Group Edges in XMPTSModel:0
[2019-02-27 11:15:08,870] [http-bio-443-exec-26] [topology] [INFO ] - AbstractTSServiceCommand: Number of Others in XMPTSModel:0
[2019-02-27 11:15:08,870] [http-bio-443-exec-26] [topology] [INFO ] - AbstractTSServiceCommand: Model Attributes:[XT_UseBendsForMultiEdges,true][XT_PreciseShapeClipping,true][XT_ShowBadging,true][XT_ShowEdgeLabel,true][XT_SelectedGroupId,501506][XT_ShowNodeLabel,true][OverlayFlow,false][XT_ShowGroupBGImage,true][XT_ShowEdge,true][XT_ShowGroupNodeLabel,true][XT_ShowConnector,false][XT_ShowGroupEdge,true][XT_InterestedAlarmList,[MAJOR, CRITICAL]][selectedGroupID,501506]
[2019-02-27 11:15:08,871] [http-bio-443-exec-26] [topology] [INFO ] - AbstractTSServiceCommand: Is Firing Event:true
Couldn't open RGB_DB '/usr/share/X11/rgb'
error opening security policy file /usr/lib64/xserver/SecurityPolicy
Couldn't open RGB_DB '/usr/share/X11/rgb'
error opening security policy file /usr/lib64/xserver/SecurityPolicy
Couldn't open RGB_DB '/usr/share/X11/rgb'
error opening security policy file /usr/lib64/xserver/SecurityPolicy
Couldn't open RGB_DB '/usr/share/X11/rgb'
error opening security policy file /usr/lib64/xserver/SecurityPolicy
Couldn't open RGB_DB '/usr/share/X11/rgb'
error opening security policy file /usr/lib64/xserver/SecurityPolicy

Any ideas here?

[superego]

Try to apply the patch on CPI 3.0.  Run md5 check on the upgrade file, "f348c9d9ca00c5a956b9dda0e5beb3f6" to verify integrity of the file.

Also ensure that /localdisk has 80% free space as well.  Move the back up elsewhere if you have.

[stephendrkw]

I have plenty of disk space left. Downloaded 3.0.3 patch which is the only patch available for 3.0.x, says you can only install on a build from 3.0.3.x I'm running 3.0.0.278 so looks like it's not possible https://www.cisco.com/web/software/284272933/131958/PI_3.0.3_Update_04_Release_Notes.pdf

[patoberli]

That's correct. I fear you have to contact TAC to provide you with the required files.
You could also try to skip 3.1 and go for 3.2. The file is named PI-Upgrade-3.X_to_3.2.0.0.258.tar.gz and is available for download.
Manual: https://www.cisco.com/c/en/us/td/docs/net_mgmt/prime/infrastructure/3-2/quickstart/guide/cpi_qsg.html
Important note from there:
If you are upgrading from a Prime Infrastructure 3.0.x or 3.1.x virtual machine, you should create a new virtual machine with the same deployment configuration option you used previously (for example, Express, Express-Plus, Standard, or Professional), back up your existing virtual machine, and then restore it on the new virtual machine.

[Scott Fella]

Just to add... if your PI is a VM, it’s best to just export your devices and maps and then build a new VM from scratch and import that data. You will have to setup your PI settings again, but that shouldn’t take very long.
Since the early days of WCS, BCS and PI, upgrading to a different train has always caused issues and or corrupted the db. It’s faster to just build a new VM if you have the compute. I’m currently doing that with PI 3.5 since I upgraded our existing from 3.4 to 3.5 and now found issues. It took me about less than 4 hours to get the new one up and functioning.

[stephendrkw]

In the end I upgraded directly from 3.0 to 3.5 and this worked,  I asked our Service Provider to log a TAC rather they suggested to use an upgrade bundle package directly from 3.0.x to 3.5.x - "PI-Upgrade-3.X_to_3.5.0.0.550.tar.gz"

Licenses were transferred correctly, everything looks complete. I'm a little surprised when Cisco say, needs to be 3.1.5.x, my guess this is an newish bundle which is dated 30-Nov-2018

[patoberli]

You were lucky in this case. Officially it's not supported, who knows if you run into problems in the long run. I hope you wont.

[stephendrkw]

Why would the upgrade bundle package be not supported if it's listed on Cisco's download site?...out of interest

[patoberli]

The manual doesn't state your old version as "permitted" upgrade path. Doesn't mean it will not work, but there might be some missing database/file alter scripts that haven't run now.

[stephendrkw]

Thanks, will keep an eye out for any abnormalities.