[Pgcluster-general] 4 packages (1.3, 1.5, 1.7 and 1.9) were updated

[At.Mitani]

Hi all,

I updated 4 packages as this subject.

Main changes are as follows,
(1)Bug fix
Filip Rembialkowski pointed out that the regeneracy mode of pglb might not work in two set of cluster DBs environment.
On a related this issue, I found several problems on 64bit processor.

(2)Add new parameters in cluster.conf
Additional parameters of rsync were suggested by Rick Vernam as follows,
-- Rsync_Timeout
  this is I/O timeout of rsync
-- Rsync_Bwlimit
  this is I/O bandwidth limitation os rsync

(3)Auto retry of rsync
If rsync failed, it is going to retry and adjusting bandwidth limitation automatically.
In order to measure bandwidth between master DB and target DB,
one parameter was added in cluster.conf.
-- Ping_Path
  this is a full path of ping command.
  yes, the measurement tool is using the ping command.

(4)Timeout detection of lifecheck was changed
Anatoly Savchenkov pointed out that PQsetdbLogin() may hang up sometimes (depends on TCP stack configuration of host OS).
Therefore I rewrited PGRcreateConn() to use connect_timeout as his suggestion.

It is using the same configuration files, replication server and loadbalance server among these four packages.
I don't recomend use as mixed version, however it is possible.

Now, the corresponding versions of PostgreSQL are as follows,
  pgcluster-1.3.1rc8  => PostgreSQL 8.0.15
  pgcluster-1.5.0rc18 => PostgreSQL 8.1.11
  pgcluster-1.7.0rc9  => PostgreSQL 8.2.6
  pgcluster-1.9.0rc2  => PostgreSQL 8.3.0

Best Regards,
-- 
At.Mitani <mitani at sraw.co.jp>