Hi All!
I have a small problem with the deployment of a RAIDB-1 with sequoia,
using two nodes with one backend each one. When I launch the first one,
no problem at all. The problem begins when i try to launch the second
one... I begin to see errors in the console about errors trying to
insert duplicate entries in checkpoint table in both nodes.
It's something similar to the bug:
https://forge.continuent.org/jira/browse/SEQUOIA-628 that seem solved in
branch 2.9.
This is the console of the first node when i launch the second
controller in node2.
17:49:52,022 INFO controller.core.Controller Sequoia controller
(2.10.3)
17:49:52,097 INFO controller.core.Controller Loading configuration
file: /sequoia/sequoia-2.10.3-bin/config/controller/controller.xml
17:49:52,163 INFO controller.core.Controller JMX is enabled
17:49:52,186 INFO controller.core.Controller Starting JMX server on
host: 192.168.0.6
17:49:52,448 INFO controller.backup.BackupManager Registering backuper
mysql to handle format MySQL raw dump
17:49:52,469 INFO backend.DatabaseBackend.DB1 Adding connection manager
for virtual user "openser"
17:49:52,469 INFO backend.DatabaseBackend.DB1 Adding connection manager
for virtual user "openserro"
17:49:52,807 INFO controller.RequestManager.opensequoia Request manager
will parse requests with the following granularity: TABLE
17:49:52,817 INFO controller.virtualdatabase.opensequoia Using Hedera
properties file: /hedera_jgroups.properties
-------------------------------------------------------
GMS: address is 192.168.0.6:32857
-------------------------------------------------------
17:49:57,359 INFO controller.virtualdatabase.opensequoia Group
opensequoia connected to Member(address=/192.168.0.6:32857,
uid=opensequoia)
17:49:57,360 INFO controller.virtualdatabase.opensequoia First
controller in group opensequoia
17:49:57,363 WARN controller.RequestManager.opensequoia Backend DB1 was
not properly stopped, manual recovery will be needed (Backend:DB1,
VirtualDatabase:opensequoia, State:enabled, Checkpoint:)
17:49:57,365 INFO sequoia.controller.recoverylog Checking recovery log
consistency
17:49:57,369 INFO controller.virtualdatabase.opensequoia Unknown last
state for backend DB1. Leaving node in disabled state.
17:49:57,369 INFO controller.core.Controller Adding VirtualDatabase
opensequoia
17:49:57,374 INFO controller.core.Controller Waiting for connections on
0.0.0.0:25322
17:49:57,380 INFO controller.core.Controller Controller started on
2007.01.05 01 at 05:49:57 PM CET
17:49:57,381 INFO controller.core.Controller Controller
192.168.0.6:25322 ready, listening to requests ...
++++++++++++++++++++++++++++++++++++++++ ERROR +++++++++++++++++++++++++
17:50:04,111 ERROR sequoia.controller.recoverylog Logger thread was
unable to log Storing checkpoint Member(address=/192.168.0.7:32771,
uid=opensequoia) joined group
opensequoia-192.168.0.6:25322-20070105175003698+0100 with id 0 because
of java.sql.SQLException: Unable to store checkpoint
Member(address=/192.168.0.7:32771, uid=opensequoia) joined group
opensequoia-192.168.0.6:25322-20070105175003698+0100 in recovery log
database (Duplicate entry 'Member(address=/192.168.0.7:3277' for key 1)
java.sql.SQLException: Unable to store checkpoint
Member(address=/192.168.0.7:32771, uid=opensequoia) joined group
opensequoia-192.168.0.6:25322-20070105175003698+0100 in recovery log
database (Duplicate entry 'Member(address=/192.168.0.7:3277' for key 1)
at
org.continuent.sequoia.controller.recoverylog.LoggerThread.storeCheckpointWithLogId(LoggerThread.java:398)
at
org.continuent.sequoia.controller.recoverylog.events.StoreCheckpointWithLogIdEvent.execute(StoreCheckpointWithLogIdEvent.java:69)
at
org.continuent.sequoia.controller.recoverylog.LoggerThread.run(LoggerThread.java:732)
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
17:50:06,005 INFO controller.virtualdatabase.opensequoia Checking
virtual database configuration from 192.168.0.7:25323
17:50:06,248 INFO controller.virtualdatabase.opensequoia Controller
192.168.0.7:25323 is compatible with the local configuration
17:50:06,368 INFO controller.virtualdatabase.opensequoia Updating
backend list for controller Member(address=/192.168.0.7:32771,
uid=opensequoia)
The problem is that this error is continuous. Is like every checkpoint
is inserted always two times. I tried to create checkpoint table with
primary key name, but another problem appear and is that sequoia try to
insert commits which string length is greater than 32, which sequoia
uses in its scripts to create the column name from checkpoint
(varchar(32)).
I'm using mysql drivers 3.1.14 and not hsql... I don't know if this can
affect to this problem.
Any suggestion would be appreciated... Thank you in advanced.
_______________________________________________
Sequoia mailing list
[email protected]
https://forge.continuent.org/mailman/listinfo/sequoia
