Cellar causes Karaf container to freeze if system got network interface changes
between container restarts
----------------------------------------------------------------------------------------------------------
Key: KARAF-1309
URL: https://issues.apache.org/jira/browse/KARAF-1309
Project: Karaf
Issue Type: Bug
Components: cellar-hazelcast
Affects Versions: cellar-2.2.4
Environment: Karaf-2.2.6-SNAPSHOT from 20120403, Cellar-2.2.4-SNAPSHOT
from 20120312
Reporter: Alexey Bespaly
1. Started 4 Karaf instances and installed Cellar on those
cluster:node-list
No. Host Name Port ID
* 1 opti.local 5701 opti.local:5701
2 opti.local 5702 opti.local:5702
3 opti.local 5703 opti.local:5703
4 opti.local 5704 opti.local:5704
opti.local = 192.168.1.86
2. 1,2 - default group; 3,4 - "group1" (not sure if groups are essential here,
but were a part of the test case)
cluster:group-list
Node Group
opti.local:5701 default
opti.local:5702 default
opti.local:5703 group1
* opti.local:5704 group1
3. Stopped Karaf containers
4. Got the VPN client on - tun0 (172.27.210.11) network interface added
5. Restarted Karaf containers
- the 1st and the 2nd ones seemed to be working fine:
cluster:node-list
No. Host Name Port ID
* 1 172.27.210.11 5701 172.27.210.11:5701
2 172.27.210.11 5703 172.27.210.11:5703
3 172.27.210.11 5702 172.27.210.11:5702
4 172.27.210.11 5704 172.27.210.11:5704
- the 3rd container got inresponsive:
karaf@trun> osgi:list
...no response....
log:
17:40:29,821 | INFO | hz.1.InThread | dardLoggerFactory$StandardLogger 51
| - - | /172.27.210.11:5703 [cellar] 5703 is accepting socket connection
from /172.27.210.11:50101
17:40:29,822 | INFO | hz.1.InThread | dardLoggerFactory$StandardLogger 51
| - - | /172.27.210.11:5703 [cellar] 5703 accepted socket connection from
/172.27.210.11:50101
17:40:30,196 | INFO | rint Extender: 3 | dardLoggerFactory$StandardLogger 51
| - - | /172.27.210.11:5703 [cellar] Hazelcast Management Center started at
port 5803.
17:40:30,198 | INFO | rint Extender: 3 | dardLoggerFactory$StandardLogger 51
| - - | /172.27.210.11:5703 [cellar] Address[172.27.210.11:5703] is STARTED
17:40:35,249 | INFO | rint Extender: 3 | .hazelcast.HazelcastGroupManager 243
| 211 - hazelcast - 1.9.4.6 | Registering group default.
- the 4th shows the following static picture:
...skipped...
[ 211] [Active ] [Creating ] [ ] [ 60] hazelcast (1.9.4.6)
Fragments: 213
[ 212] [Active ] [Created ] [ ] [ 60] Apache Karaf :: Cellar ::
Core (2.2.4.SNAPSHOT)
[ 213] [Resolved ] [ ] [ ] [ 60] Apache Karaf :: Cellar ::
Hazelcast (2.2.4.SNAPSHOT)
Hosts: 211
[ 214] [Active ] [GracePeriod ] [ ] [ 60] Apache Karaf :: Cellar ::
Config (2.2.4.SNAPSHOT)
[ 215] [Active ] [GracePeriod ] [ ] [ 60] Apache Karaf :: Cellar ::
Features (2.2.4.SNAPSHOT)
[ 216] [Active ] [GracePeriod ] [ ] [ 60] Apache Karaf :: Cellar ::
Bundle (2.2.4.SNAPSHOT)
[ 217] [Active ] [Created ] [ ] [ 60] Apache Karaf :: Cellar ::
DOSGi (2.2.4.SNAPSHOT)
[ 218] [Active ] [Created ] [ ] [ 60] Apache Karaf :: Cellar ::
Utils (2.2.4.SNAPSHOT)
[ 219] [Active ] [Created ] [ ] [ 60] Apache Karaf :: Cellar ::
Shell (2.2.4.SNAPSHOT)
[ 220] [Active ] [Created ] [ ] [ 60] Apache Karaf :: Cellar ::
Management (2.2.4.SNAPSHOT)
karaf@trun>
...and the status does not get changed over the time
Could it be that the stored config conflicts with the newly detected one and
brings such an instability in?
Logs of 4 Karaf instances are attached.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
