[Bucardo-general] MultiMaster replication sequence problem

[Vilem Kebrt]

Hello all,
is there any posibility to specify for bucardo that when he is 
recovering from "crash" (primar 1 down), he MUST replicate sequence in 
goat to HIGHER ?
Issue:
     2x db server for radius, both masters...
     1. -> pgsql 8.3, bucardo 4.4.0
     2. -> pgql 8.3
     1. was disabled for about 2h (system was transfered to another HW) 
-> debian lenny, virtual machine.
     2 was working
     when 1.one goes online, he takes data from 2.(correct), but then he 
rewrited sequences added in goat to lower which was on 1. and this issue 
DISABLED both radius servers connected to this primar db's.

This is very serious problem, it disables client's on 3 networks!.
So for my question -> is any posibility to specify for bucardo that 
HIGHER in sequence is better ? replicate resolve_conflict "source" don't 
do this and as i wrote more times ago, THERE IS NO REFERENCE 
DOCUMETATION SOURCE ;(
Please, if anybody can send me some information, i'm thinking about make 
documetation by myself...
with regards Vilem
-- 
S pozdravem
Vilem Kebrt
-----------------------------------
NETAIR, s.r.o.
Vilém Kebrt
512 43 Jablonec nad Jizerou č.p.364
mobil: +420 604 550 132
tel: +420 604 553 444
voip: +420 488 572 050
mail: kebrt at netair.cz
web: www.netair.cz
-----------------------------------