[CONTROLLER-1735] Entityownership leastload policy doesn't work normally Created: 03/Jul/17  Updated: 08/Nov/22  Resolved: 08/Nov/22

Status: Resolved
Project: controller
Component/s: eos
Affects Version/s: Carbon
Fix Version/s: None

Type: Bug
Reporter: Xu Renfei Assignee: Ivan Martiniak
Resolution: Won't Do Votes: 0
Labels: pt
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Operating System: All
Platform: All


External issue ID: 8785

 Description   

Sometimes the entityownership leastload policy doesn't work normally.

Steps to reproduce,
1. Bring up a 3 node cluster.
2. Add six entities, now every node owns two entities.
3. Kill the controller process for member-1.(Or isolate the member-1 network). Now the member-2 and member-3 both have three entities.
4. Restart the member-1(Or recovery the member-1 network) and kill member-2.
5. Sometimes the member-1 owns 2 entities and member-3 owns 4 entities. The result we want is member-1 and member-3 both own 3 entities.(If this doesn't appear, try again...)



 Comments   
Comment by Robert Varga [ 24/Aug/17 ]

What version is this? Also can you upload karaf.logs from the instances, please?

Comment by Ivan Martiniak [ 08/Nov/22 ]

This issue relates to older EOS implementation, which included owner-picking strategies. The current version of ODL has a new EOS implementation without these strategies. Hence, a new issue should be created with the further specification of owner-picking procedures.

Generated at Wed Feb 07 19:56:19 UTC 2024 using Jira 8.20.10#820010-sha1:ace47f9899e9ee25d7157d59aa17ab06aee30d3d.