Prime infra 3.5.1 HA stuck in secondary syncing and loose Gui access

[Augustgood]

Hi all,

Our new deployment of prime infrastructure with HA don't work. The secondary node stuck in secondary syncing and the vip ip GUI is not accessible.

Someone can help me ?

Thx all

[balaji.bandi]

is the primary working as expected?  how is your network between Primary and Secondary? any other FW  involved or just Layer 2?

Can you explain is this new setup, or after HA build having issue ?

some FAQ for reference :

https://community.cisco.com/t5/networking-documents/cisco-prime-infrastructure-faq/ta-p/3149669

[Augustgood]

The primary, secondary node and vip are on the same subnet. No firewall, prime version 3.5.1 (standard ) under vmware esxi 5.5. This is  a new deployment (fresh install), all the service run from the primary and secondary... but the secondary node show syncing from more than 8 hours from when we configured HA.

Thx for reply.

[renjithg]

Hi Nicoles,

This is HA design.

For HA healthy state, primary  role would be Primary Active and Secondary server role would be Secondary sync. This means database is in sync with primary.

Also, When secondary server is in secondary standby role, only health monitor page is accessible due to fact that database is only available to sync data.

Please refer HA document for understanding different states of server in HA.

https://www.cisco.com/c/en/us/td/docs/net_mgmt/prime/infrastructure/3-5/admin/guide/bk_CiscoPrimeInfastructure_3_5_AdminGuide/bk_CiscoPrimeInfastructure_3_5_AdminGuide_chapter_01010.html

Hope this clarifies the question.

[Augustgood]

Hi,

us i told you in the object we loose gui access from the primary node (and VIP). This isn't a normal behaviour...?

[renjithg]

provide ncs ha status from cli session of both primary and secondary server.

Ensure all services are running in primary server. ncs status.

Also, do not keep HA failover setting as automatic since it can lead to split brain scenario where network hiccup between the pairs can make both server take primary role if no response received within specified time between each other.

Also, it is very important to note network throughput between HA pairs must be 100Mb/sec and consistant since oracle db data is shared and missing of certain files can cause oracle to be be move to offline which cause service to halt.

[SeanYoung02972]

I know this is an old ticket, but I had the exact same issue so this might help someone else.

The Routing Table on my Primary Node had updated with a static route entry to point 10.0.0.0/255.0.0.0 to gateway 0.0.0.0. I removed this entry and everything started functioning properly.