On Sat, 1 Apr 2006 03:50:29 +0100
Gavin Hamill <gdh at laterooms.com> wrote:

Just a little addendum..

> 2006-04-01 03:13:58 BST DEBUG1 remoteWorkerThread_1: helper thread for provider 1 created
> 2006-04-01 03:13:58 BST DEBUG1 remoteWorkerThread_1: connected to data provider 1 on 'host=194.24.250.137 dbname=laterooms user=XXX port=5432 password=XXXXXXXXX'

Doing 'select * from pg_stat_activity' shows the time at which the 'idle in transaction' process began...

 <IDLE> in transaction | 2006-04-01 03:14:02.538096+01

i.e. just moments after the copy_set is done and the first real sync will have been brought down from Node1, no doubt..

The plot thickens, but I'm still at a loss as to why this is happening :(

Cheers,
Gavin.