SYMC:Same IP is being assigned to ports and service instances

Bug #1480517 reported by Rudrajit Tapadar
10
This bug affects 2 people
Affects Status Importance Assigned to Milestone
Juniper Openstack
New
Undecided
Unassigned
OpenContrail
New
Undecided
Unassigned

Bug Description

After the upgrade to 2.02 build 59 we got user complaints that they are being kicked out of ssh sessions frequently. On investigating, I found that LBaaS and VM ports have same IPs assigned and hence flows are only intermittently set up on a hypervisor.

Revision history for this message
Prakash Bailkeri (prakashmb) wrote :

Looks like duplicate ip address allocation issue. Can you check the zookeeper status and validity of data after upgrade? How the zookeeper data backed/restored in the upgrade process?

Also, please confirm which object existed before upgrade and which one is created after upgrade? Whether the VM was created before upgrade or LBaas? I assume by LBaas you are referring to VIP address allocated for LBaas.

Can you please get the api server logs(contrail-api-0-stdout.log* and contrail-api.log.* along with api-0-zk.log* to check the address allocation issue?

Revision history for this message
Rudrajit Tapadar (rtapadar) wrote :

Zookeeper is running on the db nodes. The db nodes were not touched in this upgrade and they are still running 2.01-43, only controllers and vrouters were upgraded to 2.02-59.

Objects created before and after the upgrade both are affected by IP conflicts. I have provided all the logs in the slack channel.

tags: added: config
tags: added: service-chain
summary: - Same IP is being assigned to ports and service instances
+ SYMC:Same IP is being assigned to ports and service instances
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.