[QPID-5973] HA cluster state may get stuck in recovering

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.28
• Fix Version/s: 0.31
• Component/s: C++ Clustering
• Assignee: Alan Conway
• Reporter: Alan Conway
• Created: 08/Aug/14 06:43
• Updated: 19/Mar/15 12:23
• Resolved: 08/Aug/14 09:27

Description

HA brokers can become stuck in "recovering" or "catchup" state when running transactional clients with multiple failovers.

To reproduce, in one window run:

while qpid-txtest2 -b 20.0.20.200 --total-messages 1000 --connection-options '

{reconnect:true}

' --tx-count 1000; do true; done

In another window run (ha script attached):

while ha wait -a; do sleep .5; ha kill; done

After some time one or more brokers willl become stuck in catchup or recovering state.