[CONTROLLER-1785] Transactions fail on start of shutdown nodes Created: 04/Nov/17  Updated: 22/May/18  Resolved: 22/May/18

Status: Resolved
Project: controller
Component/s: clustering
Affects Version/s: None
Fix Version/s: Carbon, Nitrogen, Oxygen SR1, Fluorine

Type: Bug Priority: Highest
Reporter: Sai MarapaReddy Assignee: Unassigned
Resolution: Won't Do Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

regular cluster setup



 Description   

It happens in all versions (Boron/Carbon/Nitrogen/Master)
Steps to reproduce
1. Start a 3 node cluster
2. Transactions (car tests) succeeds.
3. Stop all 3 nodes
4. Start 2 nodes
5. Try car tests (It fails) (karaf logs show errors that car tests failed)

Note:- We checked step 5 after considerable amount of time after step 4.



 Comments   
Comment by Tom Pantelis [ 05/Nov/17 ]

Which 2 nodes did you start? One of them has to be the first seed node listed. It's a little weird - see akka's clustering docs for details.

Comment by Ajay Lele [ 06/Nov/17 ]

The down node was indeed the first seed node on all nodes. Is it valid to give own address as first seed node on each node? It seems to work and that way there is no dependency on any particular other node to be up

Comment by Ajay Lele [ 07/Nov/17 ]

Looks like using own address as 1st seed-node will not work and it will cause separate clusters to get formed. So seems like the node listed as the 1st seed-node MUST be running when the other nodes (atleast 1 other seed-node?) are started
ref. https://doc.akka.io/docs/akka/2.4.7/java/cluster-usage.html#Joining_to_Seed_Nodes

Generated at Wed Feb 07 19:56:27 UTC 2024 using Jira 8.20.10#820010-sha1:ace47f9899e9ee25d7157d59aa17ab06aee30d3d.