[Sequoia] Sequoia 2.10.10

[Julio Leyva]

Hi all

We are running sequoia-2.10.10 against 2 postgresql 8.3 backends
Our front end uses Java,Tomcat , apache
In one of the tomcat logs we see these error messages that are coming out from Sequoia

16s time shift detected, from Wed May 19 11:41:07 AST 2010 to Wed May
Brutal changes of date/time can lead to erroneous controller failure detections!



In the sequoia logs we see these messages

2010-05-19 11:41:52,174 WARN virtualdatabase.VirtualDatabaseWorkerThread.test Client (login:postgres,host:localhost closed connection with server)
2010-05-19 11:41:52,180 INFO controller.virtualdatabase.test Transparent client failover operated for transaction 237058
2010-05-19 11:41:52,181 WARN virtualdatabase.VirtualDatabaseWorkerThread.test Client (login:postgres,host:localhost closed connection with server)
2010-05-19 11:41:52,183 INFO controller.virtualdatabase.test Transparent client failover operated for transaction 237055
2010-05-19 11:41:52,173 WARN virtualdatabase.VirtualDatabaseWorkerThread.test Client (login:postgres,host:localhost closed connection with server)
2010-05-19 11:41:52,173 WARN virtualdatabase.VirtualDatabaseWorkerThread.test Client (login:postgres,host:localhost closed connection with server)
2010-05-19 11:41:52,184 INFO virtualdatabase.VirtualDatabaseWorkerThread.test Error during command execution (Transparent failover for statements that potentially return multiple results is only supported in distributed configurations.)


Any ideas why that should happened?

Thanks

[Emmanuel Cecchet]

Hi Julio,

It looks like your clock is drifting fast and ntp is probably doing big
adjustments. Running ntp more frequently would prevent such clock shifts.

For the second part, it looks like you are using a single controller
configuration (not a distributed virtual database with group
communication). I this configuration you cannot have transparent
failover to another controller (which is basically what the message says).

Hope this helps
Emmanuel

Post by Julio Leyva
We are running sequoia-2.10.10 against 2 postgresql 8.3 backends
Our front end uses Java,Tomcat , apache
In one of the tomcat logs we see these error messages that are coming out from Sequoia
16s time shift detected, from Wed May 19 11:41:07 AST 2010 to Wed May
Brutal changes of date/time can lead to erroneous controller failure detections!
In the sequoia logs we see these messages
2010-05-19 11:41:52,174 WARN virtualdatabase.VirtualDatabaseWorkerThread.test Client (login:postgres,host:localhost closed connection with server)
2010-05-19 11:41:52,180 INFO controller.virtualdatabase.test Transparent client failover operated for transaction 237058
2010-05-19 11:41:52,181 WARN virtualdatabase.VirtualDatabaseWorkerThread.test Client (login:postgres,host:localhost closed connection with server)
2010-05-19 11:41:52,183 INFO controller.virtualdatabase.test Transparent client failover operated for transaction 237055
2010-05-19 11:41:52,173 WARN virtualdatabase.VirtualDatabaseWorkerThread.test Client (login:postgres,host:localhost closed connection with server)
2010-05-19 11:41:52,173 WARN virtualdatabase.VirtualDatabaseWorkerThread.test Client (login:postgres,host:localhost closed connection with server)
2010-05-19 11:41:52,184 INFO virtualdatabase.VirtualDatabaseWorkerThread.test Error during command execution (Transparent failover for statements that potentially return multiple results is only supported in distributed configurations.)
Any ideas why that should happened?
Thanks
------------------------------------------------------------------------
_______________________________________________
Sequoia mailing list
http://forge.continuent.org/mailman/listinfo/sequoia