Transactional Replication Breaks with General Network Error

I am running sql 7.0 on a active/passive cluster. I am
using transactional replication to push from this server
to another sql 7.0 server. I keep getting intermittent
errors on the distributor agent with error message
"General Network Error. Check your network Documentation"
When I click on details. I get category "data source" and
source "servername" and number "11", and a second entry
states category "odbc" and source "odbc sql server driver"
and number "08S01". I have tried osql and odbc connection
to the server being replicated to and they work fine. I
have reinstalled MDAC and done a rebind. Nothing seems to
fix it. After the agent starts, i get a message
stating "delivering replicated transactions" and shortly
thereafter the agent dies with the error stated above. I
have stopped each agent and cleaned up the replication
history and started the log reader agent and then the
distributor agent. Can't figure this one out.

 

Normally when you get the general network error it means that there is some
error on your network which causes replication to hiccup and fail.

You should change your querytimeout value to a higher value to help to
prevent this.

One of the things that has been observed to help with this problem is to
open an isql connection to the subscriber on your publisher and do something
like this

while (1<>2)
begin
select * from sysobjects
select getdate()
waitfor DELAY 00:05:00
end

This keep alive script will fail when the connection goes down and give you
a time of failure which can help with the troubleshooting effort, but it has
been known to keep alive lossy connections - and replication jobs.