[
https://issues.apache.org/jira/browse/QPID-5975?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Alan Conway updated QPID-5975:
------------------------------
Attachment: ha.tgz
Attached ha.tgz
Updated with second reproducer:
ha-tx-txtest2.sh: Runs qpid-txtest2 with failover only during the transfer
phase. Ignores TransactionUnknown errors from the transfer phase but runs
the
verify phase to ensure no extra/missing messages.
ha-tx-atomic.sh: Runs transactional qpid-send/qpid-receive with fail-over
and
verifies that the number of enqueues/dequeues and message depth are a
multiple
of the transaction size.
> HA extra/missing messages when running qpid-txtest2 in a loop with failover.
> -----------------------------------------------------------------------------
>
> Key: QPID-5975
> URL: https://issues.apache.org/jira/browse/QPID-5975
> Project: Qpid
> Issue Type: Bug
> Components: C++ Clustering
> Affects Versions: 0.28
> Reporter: Alan Conway
> Assignee: Alan Conway
> Attachments: ha.tgz
>
>
> Running qpid-txtest2 in a loop against a 3 node cluster while killing primary
> brokers to force failover eventually results in the test failing due to extra
> or missing messages.
> To reproduce:
> In separate windows:
> $ while ha stat; qpid-txtest2 -b 20.0.20.200 --total-messages 100
> --connection-options '{reconnect:true}' --messages-per-tx 10 --tx-count 10;
> do true; done
> $ while ha wait -a 2; do sleep .5; ha kill; done # ha script attached.
> Results:
> Eventually (up to 10 minutes) the test fails with a message like:
> The following ids were missing:
> 'msg-100'
> 'msg-51'
> 'msg-52'
> 'msg-53'
> OR
> The following extra ids were encountered:
> 'msg-1'
> 'msg-1'
> 'msg-10'
> Expected result:
> Test runs forever, no failures.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
