On 08/31/13 05:51, Neha Sharma wrote:
> Hi,
>  
> We are using master and slave postgres and slony for replcating the
> database.
> The master and slave are on different servers in the same LAN. The one
> containg master db was rebooted. When the machine came up, the slony
> started replication but the db state is still unavailable. The postgres
> logs shows the following errors:
> 
> UTCFATAL:  no pg_hba.conf entry for host "172.168.113.4", user
> "localclient", database "msw", SSL off
>  
> where 172.168.113.4 is the IP address of the server which was rebooted.
>  
> Could it be possible that before the database is up, slony starts
> replication. If not then why above error is being shown in postgres logs.
> Please note that slon logs are showing messages corresponding to copying
> of tables at the same time.

That error message only indicates that "some" client, which may or may
not be a slon process, is denied to connect because of the host based
authentication failing. A Slony cluster always consists of multiple slon
processes, each establishing multiple database connections. One of them
copying or replicating data is no guarantee that every other one can
also establish all the connections, they need.

Is Slony actually using the user "localclient" for its connections?


Jan


-- 
Anyone who trades liberty for security deserves neither
liberty nor security. -- Benjamin Franklin