[ENTMQBR-1018] When live-slave fails-back to master, it turns off everything down, even its console - JBoss Issue Tracker

XML

Word

Printable

Details

Type: Bug
Status: Done
Priority: Major
Resolution: Done
Affects Version/s: AMQ 7.1.0.CR2
Fix Version/s: AMQ 7.2.0.GA
Component/s: clustering, console, high-availability
Labels:
- polarion-tc-missing
Environment:

1 HA pair Master-Slave

Target Release:

AMQ 7.2.0.GA
Steps to Reproduce:

Hide

1) Set up HA pair with slave <allow-failback>true</allow-failback>
2) Kill master
3) Make sure slave becomes live
4) Start master again, slave gives control to master and becomes a slave
5) Console of slave (backup now) becomes inaccessible

Show
1) Set up HA pair with slave <allow-failback>true</allow-failback> 2) Kill master 3) Make sure slave becomes live 4) Start master again, slave gives control to master and becomes a slave 5) Console of slave (backup now) becomes inaccessible
Affects:

Release Notes
Workaround Description:

Hide

restart backup

Show
restart backup
Release Notes Text:
In high-availability configurations, AMQ Console is now accessible when a slave broker returns control to the master broker. Previously, AMQ Console would become unavailable for the slave broker when it gave control back to the master broker.
Release Notes Docs Status:
Documented as Resolved Issue

Description

When slave goes from live back to backup, it shuts down everything. Even its own hawtio console. That should not happen. By default console is always available.

Active backup, after master when down:

[root@mtoth-rhel74-5 ~]# netstat -nlp |grep java | sort

tcp6       0      0 :::1099                 :::*                    LISTEN      2449/java

tcp6       0      0 172.0.0.14:1883         :::*                    LISTEN      2449/java

tcp6       0      0 172.0.0.14:5445         :::*                    LISTEN      2449/java

tcp6       0      0 172.0.0.14:5672         :::*                    LISTEN      2449/java

tcp6       0      0 172.0.0.14:61613        :::*                    LISTEN      2449/java

tcp6       0      0 172.0.0.14:61616        :::*                    LISTEN      2449/java

tcp6       0      0 :::34259                :::*                    LISTEN      2449/java

>> tcp6       0      0 :::8161                 :::*                    LISTEN      2449/java

udp6       0      0 231.7.7.7:9876          :::*                                2449/java

udp6       0      0 231.7.7.7:9876          :::*                                2449/java

udp6       0      0 :::56983                :::*                                2449/java

Giving control back to resurrected live broker

[root@mtoth-rhel74-5 ~]# netstat -nlp |grep java | sort

tcp6       0      0 :::1099                 :::*                    LISTEN      2449/java

tcp6       0      0 :::34259                :::*                    LISTEN      2449/java

udp6       0      0 231.7.7.7:9876          :::*                                2449/java

Observe console port 8161 being inaccessible (console goes down)

Gliffy Diagrams

Attachments

Issue Links

follows up on

ENTMQBR-932 Failover to Slave Does Not Occur is Killing master with Kill -9 in a more than 2 broker cluster

Done

is duplicated by

ENTMQBR-1200 When live-slave fails-back to master, it turns off everything down, even its console

Done

is related to

ENTMQBR-1189 [HA-replication 1 pair] - no failback allowed, when master is killed and started again master does not connect to slave HA pair, but creates one himself

Done

ARTEMIS-1737 Loading...

Activity

People

Assignee:

Stanislav Knot

Reporter:

Michal Toth

Tester:

Michal Toth

Votes:

0 Vote for this issue

Watchers:

9 Start watching this issue

Dates

Created:

01/Feb/18 8:43 AM

Updated:

17/Aug/18 6:09 AM

Resolved:

29/Jun/18 3:26 AM