[h2] Unique index or primary key violation in cluster failover test

2013-10-31 Thread Tomáš Remeš
I test HTTP session persistence to db using latest WildFly application server (Beta2 - SNAPSHOT, h2 1.3.173). I use this connection url jdbc:h2:file;AUTO_SERVER=true;DB_CLOSE_ON_EXIT=FALSE, because I want database to survive serve shutdown. Occassionally I am facing following exception:

Re: [h2] Unique index or primary key violation in cluster failover test

2013-10-31 Thread Noel Grandin
On 2013-10-31 13:07, Tomáš Remeš wrote: I test HTTP session persistence to db using latest WildFly application server (Beta2 - SNAPSHOT, h2 1.3.173). I use this connection url jdbc:h2:file;AUTO_SERVER=true;DB_CLOSE_ON_EXIT=FALSE, because I want database to survive serve shutdown.

[h2] Unique index or primary key violation in cluster failover test

2013-10-31 Thread Tomáš Remeš
I test http session persistence to db with latest WildFly application server (Beta2-SNAPSHOT, h2 1.3.173). I have two nodes running in my test, where failover is tested and I use this connection url - jdbc:h2:file;AUTO_SERVER=true. I need to use file storing, because of server shutdown

Re: [h2] Add support for sequence min value, max value and cycle

2013-10-31 Thread Noel Grandin
On 2013-10-30 17:32, Daniel Gredler wrote: I've attached a patch to issue 520 (https://code.google.com/p/h2database/issues/detail?id=520) adding support for some extra sequence attributes (min value, max value and cycle) that are fairly common in other RDBMS. Can a committer have a look

Re: [h2] Unique index or primary key violation in cluster failover test

2013-10-31 Thread Thomas Mueller
Hi, Actually, you are not using H2 version 1.3.173. You are using an old version of H2 (1.3.159 according to the error code, [23505-159]). Could you try with a newer version? Regards, Thomas On Thu, Oct 31, 2013 at 12:07 PM, Tomáš Remeš tomas.re...@gmail.com wrote: I test HTTP session

Re: [h2] Re: IGNORECASE=TRUE and comparing strings

2013-10-31 Thread Thomas Mueller
Hi, The option ignorecase=true will convert all column that are created as varchar to varchar_ignorecase. I think the problem is that text literals are still interpreted as varchar, not varchar_ignorecase, so that if you compare text literals with other text literals, this is still case