"Daniel Inman" <DanielInman@discussions.microsoft.com> wrote in message
news:914EA9FB-1554-45F7-BC84-47E44D56161C@microsoft.com...
> I am using transactional replication with queued updating. The publisher
is
> located in Norway, with a single subscriber in the US. The Queue Reader
> Agent starts successfully, and transactions originating on either side are
> properly replicated. However, the Queue Reader agent eventually shuts
down
> with the message "Queue Reader Aborting". When I check the session log, I
> find occurrences of the error: "Server does not exist or access denied"
> throughout the entire time the agent runs. Right before the agent
actually
> shuts down, I get the following three errors:
> "Communication link failure"
> "Error in caching messages from SQL Queue"
> "Queue Reader Aborting"
>
> After the queue reader fails, it does not restart automatically - it
> requires manual intervention. We are aware that the connection is
sporatic,
> and can accept the queue building up on both sides. Is the communication
> link failure a result of some threshold being reached? Can that threshold
be
> removed? If not, is there a way to have the queue reader agent
automatically
> restart?