cybercluster (1.0.0) stable; urgency=low

  * Finished testing with large loads, rc10 becomes stable 1.0.
  * Enabled debug build in case we are given a core dump. 

 -- zozo <zb@cybertec.at>  Tue, 07 Aug 2007 08:50:00 +0200

cybercluster (1.0.0rc10) unstable; urgency=low

  * Finalize per-thread stack size at 1MB as it's stable
    for both 32 and 64-bit.
  * Use infinite timeout when a cluster node waits for answer
    from the replication server. This eliminates timeout problems
    and cut transactions when nodes are too busy serving other
    clients.
  * Re-enable lock conflist detection in the backend node.
  * Modified a global variable to be per-session variable.
  * Don't call pthread_detach() for threads that were already exited,
    i.e. pthread_join() returned successfully.

 -- zozo <zb@cybertec.at>  Mon, 06 Aug 2007 14:45:00 +0200 

cybercluster (1.0.0rc9) unstable; urgency=low

  * Experiment with per-thread stack size in pgreplicate
    to reduce memory requirements.
  * Delay pthread_attr_destroy() after the child threads exited 


 -- zozo <zb@cybertec.at>  Thu, 02 Aug 2007 17:00:00 +0200

cybercluster (1.0.0rc8) unstable; urgency=low

  * Fix pglb to activate recovered hosts

 -- zozo <zb@cybertec.at>  Sun, 29 Jul 2007 11:30:00 +0200

cybercluster (1.0.0rc7) unstable; urgency=low

  * Fix rsync-based recovery without backing up first (option -u)

 -- zozo <zb@cybertec.at>  Tue, 10 Jul 2007 12:25:00 +0200

cybercluster (1.0.0rc6) unstable; urgency=low

  * Fix a locking bug in the backend

 -- zozo <zb@cybertec.at>  Fri, 06 Jul 2007 13:30:00 +0200

cybercluster (1.0.0rc2) unstable; urgency=low

  * Fix rsync-based recovery

 -- zozo <zb@cybertec.at>  Fri, 29 Jun 2007 00:31:00 +0200

cybercluster (1.0.0rc1) unstable; urgency=low

  * Initial Release.

 -- zozo <zb@cybertec.at>  Fri, 18 May 2007 15:02:01 +0200

Local variables:
mode: debian-changelog
End:
