[jboss-jira] [JBoss JIRA] (JGRP-1801) DuplicateTest fails when testing OOB multicast to all three senders

Wed Mar 26 10:21:13 EDT 2014

    [ https://issues.jboss.org/browse/JGRP-1801?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12956592#comment-12956592 ] 

Richard Achmatowicz edited comment on JGRP-1801 at 3/26/14 10:20 AM:
---------------------------------------------------------------------

Even if I adjust the timeout to wait for 60 seconds, I still get failures. 
I added in code to print a dot (.) after each second of waiting. In passing tests, I see this:
{noformat}
------------- testOOBMulticastToAll3Senders -----------
Checking for messages from all senders:
.
[C1]: C1: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
[C1]: C2: [9, 2, 1, 10, 3, 5, 6, 7, 4, 8]
[C1]: c3: [1, 10, 2, 3, 5, 4, 8, 9, 7, 6]
Checking for messages from all senders:

[C2]: C1: [1, 2, 10, 3, 4, 6, 8, 9, 5, 7]
[C2]: C2: [3, 1, 2, 4, 5, 6, 7, 8, 9, 10]
[C2]: c3: [2, 5, 6, 1, 10, 3, 7, 8, 9, 4]
Checking for messages from all senders:

[C3]: C1: [1, 2, 10, 3, 4, 5, 6, 7, 8, 9]
[C3]: C2: [1, 10, 2, 3, 4, 5, 6, 7, 8, 9]
[C3]: c3: [1, 8, 2, 3, 4, 5, 9, 10, 7, 6]
{noformat}

but in the faiiing tests, I see this:
{noformat}
------------- testOOBMulticastToAll3Senders -----------
Checking for messages from all senders:

[C1]: C1: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
[C1]: C2: [1, 6, 5, 2, 3, 4, 10, 7, 8, 9]
[C1]: c3: [1, 10, 2, 3, 4, 6, 7, 8, 9, 5]
Checking for messages from all senders:

[C2]: C1: [1, 2, 3, 10, 4, 5, 7, 8, 9, 6]
[C2]: C2: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
[C2]: c3: [1, 5, 2, 3, 4, 10, 6, 7, 8, 9]
Checking for messages from all senders:
............................................................
{noformat}

In other words, even with a timeout of 60 seconds, on channel C3, messages are not being received from one of the three channels. It seems that something (either sending or receiving) is getting "stuck".

Again, disabling the OOB thread pool fixes things.

      was (Author: rachmato):
    Even if I adjust the timeout to wait for 60 seconds, I still get failures. 
I added in code to print a dot (.) after each second of waiting. In passing tests, I see this:
{noformat}
------------- testOOBMulticastToAll3Senders -----------
Checking for messages from all senders:
.
[C1]: C1: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
[C1]: C2: [9, 2, 1, 10, 3, 5, 6, 7, 4, 8]
[C1]: c3: [1, 10, 2, 3, 5, 4, 8, 9, 7, 6]
Checking for messages from all senders:

[C2]: C1: [1, 2, 10, 3, 4, 6, 8, 9, 5, 7]
[C2]: C2: [3, 1, 2, 4, 5, 6, 7, 8, 9, 10]
[C2]: c3: [2, 5, 6, 1, 10, 3, 7, 8, 9, 4]
Checking for messages from all senders:

[C3]: C1: [1, 2, 10, 3, 4, 5, 6, 7, 8, 9]
[C3]: C2: [1, 10, 2, 3, 4, 5, 6, 7, 8, 9]
[C3]: c3: [1, 8, 2, 3, 4, 5, 9, 10, 7, 6]
{noformat}

but in the faiiing tests, I see this:
{noformat}
------------- testOOBMulticastToAll3Senders -----------
Checking for messages from all senders:

[C1]: C1: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
[C1]: C2: [1, 6, 5, 2, 3, 4, 10, 7, 8, 9]
[C1]: c3: [1, 10, 2, 3, 4, 6, 7, 8, 9, 5]
Checking for messages from all senders:

[C2]: C1: [1, 2, 3, 10, 4, 5, 7, 8, 9, 6]
[C2]: C2: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
[C2]: c3: [1, 5, 2, 3, 4, 10, 6, 7, 8, 9]
Checking for messages from all senders:
............................................................
{noformat}

In other words, even with a timeout of 60 seconds, on channel C3, messages are not being received from one of the three channels. It seems that something (either sending or receiving) is getting "stuck".

> DuplicateTest fails when testing OOB multicast to all three senders
> -------------------------------------------------------------------
>
>                 Key: JGRP-1801
>                 URL: https://issues.jboss.org/browse/JGRP-1801
>             Project: JGroups
>          Issue Type: Bug
>    Affects Versions: 3.2.13
>         Environment: Windows (18), Solaris (2), RHEL(1) where (x) is the number of failures seen
>            Reporter: Richard Achmatowicz
>            Assignee: Bela Ban
>             Fix For: 3.2.14
>
>
> DuplicateTest does the following:
> - creates three channels containing the DUPL layer called c1, c2, c3
> - DUPL is used to duplicate messages sent
> - channel receiver for channel each keeps a map of messages received from each sender
> - channels send messages: regular, OOB, or mixed
> - the test checks that the messages received  by each channel are correct in number and in order
> The test testOOBMuloticastToAll3Senders is failing regularly on multple platforms. The test makes each of the channels send OOB multicast messages to the group, but only two of three members ever end up receiving the multicast messages. Al example of the failure:
> {noformat}
> Error Message
> expected size=3, msgs: [C2, C1]
> Stacktrace
> java.lang.AssertionError 
> at org.jgroups.tests.DuplicateTest.check(DuplicateTest.java:217) 
> at org.jgroups.tests.DuplicateTest.testOOBMulticastToAll3Senders(DuplicateTest.java:123) 
> {noformat} 

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira