[ 
https://issues.apache.org/jira/browse/ARTEMIS-4821?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17856647#comment-17856647
 ] 

Clebert Suconic commented on ARTEMIS-4821:
------------------------------------------

I did a git bisect and found out your test started to fail at here:

# first bad commit: [e47d8ea7c15e9dbdf450478b79db948e18f04f31] ARTEMIS-4774 
Fixing PageCounters out of sync after AckMnager retries

(I don't know yet.. just putting this now for future reference):

git bisect start
# status: waiting for both good and bad commits
# bad: [a600191dbf948dfe195babb0e7be2186ac572d55] RELEASE update on doc
git bisect bad a600191dbf948dfe195babb0e7be2186ac572d55
# status: waiting for good commit(s), bad commit known
# good: [f05b63b8a19186095858c5c577404f2bbeb4db19] [maven-release-plugin] 
prepare release 2.30.0
git bisect good f05b63b8a19186095858c5c577404f2bbeb4db19
# good: [fbfe81eb4d5f3c3c625a9beaa9647b837ed6058a] ARTEMIS-4581 fix NPE crash 
in PostOfficeImpl
git bisect good fbfe81eb4d5f3c3c625a9beaa9647b837ed6058a
# good: [fc13a39c6fd130656396d433cbacdc4b0994101d] ARTEMIS-4695 
DLS_DEAD_LOCAL_STORE in AMQPLargeMessage.java
git bisect good fc13a39c6fd130656396d433cbacdc4b0994101d
# good: [bf81e5fe313dea2c7936aa98b215636193510084] NO-JIRA Test validating 
multiple mirrors connected to a node
git bisect good bf81e5fe313dea2c7936aa98b215636193510084
# good: [e78136b4294dc61989fe33385f7626535c36c529] ARTEMIS-4743 Renaming 
--loop-sleep as just --sleep
git bisect good e78136b4294dc61989fe33385f7626535c36c529
# bad: [3c84f26413bf140244beda32eb46dc5559cecf08] ARTEMIS-4778: split two 
Nested test classes (one parameterized, other not) into their own class files
git bisect bad 3c84f26413bf140244beda32eb46dc5559cecf08
# bad: [1de617e5036cd1698544ca46c235095d7a1c6fb0] NO-JIRA: fix docs 
locator.createClientSessionFactory should be locator.createSessionFactory
git bisect bad 1de617e5036cd1698544ca46c235095d7a1c6fb0
# bad: [e47d8ea7c15e9dbdf450478b79db948e18f04f31] ARTEMIS-4774 Fixing 
PageCounters out of sync after AckMnager retries
git bisect bad e47d8ea7c15e9dbdf450478b79db948e18f04f31
# good: [ae5d98337f00d5a123d95994b8437cdca1d319c4] NO-JIRA Improve a specific 
QueueControlTest
git bisect good ae5d98337f00d5a123d95994b8437cdca1d319c4
# good: [22540cc3ab5802e1b3019f55791a20e5adca8983] ARTEMIS-4773 Performance 
improvement on page.sync
git bisect good 22540cc3ab5802e1b3019f55791a20e5adca8983
# first bad commit: [e47d8ea7c15e9dbdf450478b79db948e18f04f31] ARTEMIS-4774 
Fixing PageCounters out of sync after AckMnager retries
# bad: [e47d8ea7c15e9dbdf450478b79db948e18f04f31] ARTEMIS-4774 Fixing 
PageCounters out of sync after AckMnager retries
git bisect bad e47d8ea7c15e9dbdf450478b79db948e18f04f31
# first bad commit: [e47d8ea7c15e9dbdf450478b79db948e18f04f31] ARTEMIS-4774 
Fixing PageCounters out of sync after AckMnager retries

> Message count is not correct in case of failover while a bridge is open on a 
> paging address
> -------------------------------------------------------------------------------------------
>
>                 Key: ARTEMIS-4821
>                 URL: https://issues.apache.org/jira/browse/ARTEMIS-4821
>             Project: ActiveMQ Artemis
>          Issue Type: Bug
>    Affects Versions: 2.35.0
>            Reporter: nmeylan
>            Priority: Major
>         Attachments: MessageCount230Test.java, MessageCount235Test.java
>
>
> After upgrading from 2.30 to 2.34 one of my test case is failling 
> {{QueueImpl::getMessageCount()}} does not returns what was expected
> See attached Test [^MessageCount235Test.java]  case 
> *testMessageCount_whenFailoverHappenWhileBridgeIsOpen* which is failing, I 
> attempt other scenario which succeed
> I link also  [^MessageCount230Test.java] for version 2.30 (it is another file 
> because due to renaming of master/slave, SimpleString and migration to 
> junit5, code needs to be adapt), all scenario are succeed
>  
> {code:java}
> Topology: Simulating DC1 (1 primary + backup pair)
> Simulating DC2 1 single primary node
> 1. Produce messages on dc1Primary1
> 2. Start a bridge from dc1Primary1 to dc2Primary1
> 3. Stop dc1Primary1 while transfer has started but not finished
> 4. dc1Backup1 become alive
> 5. Start a bridge from dc1Backup1 to dc2Primary1
> 6. dc2Primary1 has received all messages produced originally
> 7. Expected: messageCount on dc1Backup1 to be equal to 0, but actual is > 0
> Notes: if address is not paging, there is no issue
> {code}
> After investigation it happen only when address is paging.
> Also as I face the issue only when using bridge, issue might be due to async 
> acknowledgement



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
For further information, visit: https://activemq.apache.org/contact


Reply via email to