How do i solve this?
It's happening about once a week, when the link between 3 replicating
servers goes down.
The only way i could solve this is backing up and restoring the
database with the limbo transaction, but it takes more than 1 hour,
and there are desperated users needing to use the application.
Is there a way to prevent this from happening?

The hosts names are ok and i can connect to each database alone.

[root@server1 log]# gfix -list kins01
Transaction 65443 is in limbo.
Could not reattach to database for transaction 65443.
Original path: kinsberg.personalsoft.com.br:kins01
Multidatabase transaction:
Host Site: server1.thalys-sp.com.br
Transaction 65443
is not available.
Remote Site: kinsberg.personalsoft.com.br
Database Path: kinsberg.personalsoft.com.br:kins01
[root@server1 log]#


[root@server1 log]# gfix -commit 65443 kins01
Could not reattach to database for transaction 65443.
Original path: kinsberg.personalsoft.com.br:kins01
Insufficient information is available to determine
a proper action for transaction 65443.
Multidatabase transaction:
Host Site: server1.thalys-sp.com.br
Transaction 65443
is not available.
Remote Site: kinsberg.personalsoft.com.br
Database Path: kinsberg.personalsoft.com.br:kins01
unexpected end of input
unknown ISC error 0
[root@server1 log]#


[root@server1 log]# gfix -rollback 65443 kins01
Could not reattach to database for transaction 65443.
Original path: kinsberg.personalsoft.com.br:kins01
Insufficient information is available to determine
a proper action for transaction 65443.
Multidatabase transaction:
Host Site: server1.thalys-sp.com.br
Transaction 65443
is not available.
Remote Site: kinsberg.personalsoft.com.br
Database Path: kinsberg.personalsoft.com.br:kins01
unexpected end of input
unknown ISC error 0
[root@server1 log]#

Regards,

Fabiano.