LogSat Software

Message Topic Search Topic Options Post Reply Create New Topic Printable Version Translate Topic

   Anyone running multiple SF's over a slow link such as a WAN?
We're looking at adding another Spam filter for redundancy.  The 2nd unit would be in a different town, and thus a different ISP.  VPN is not really an option.  Anyone doing this while sharing a single database?  Is there anything that I should know?
Question 1b - I assume that a single web interface for multiple spam filters is not a problem, even if the same domains are on both servers?  Am I missing anything here?
Thanks
jerbo128

Author	Message Topic Search Topic Options Post Reply Create New Topic Printable Version Translate Topic
jerbo128 Members Profile Send Private Message Find Members Posts Add to Buddy List Senior Member Joined: 06 March 2006 Status: Offline Points: 178	Post Options Post Reply Quote jerbo128 Report Post Thanks(0) Quote Reply Topic: Multiple SF’s over slow link? Posted: 18 May 2007 at 11:01pm
	Anyone running multiple SF's over a slow link such as a WAN? We're looking at adding another Spam filter for redundancy. The 2nd unit would be in a different town, and thus a different ISP. VPN is not really an option. Anyone doing this while sharing a single database? Is there anything that I should know? Question 1b - I assume that a single web interface for multiple spam filters is not a problem, even if the same domains are on both servers? Am I missing anything here? Thanks jerbo128

Desperado Members Profile Send Private Message Find Members Posts Add to Buddy List Senior Member Joined: 27 January 2005 Location: United States Status: Offline Points: 1143	Post Options Post Reply Quote Desperado Report Post Thanks(0) Quote Reply Posted: 19 May 2007 at 12:36am
	jerbo128, Define a "slow link" We have our backup SF in a datacenter in a different state but only have 1Mb/s allocated to our backup. It writes to our local MSSQL server and sends around 200K quarantined messages per day to the remote (to it) SQL server. The traffic is minimal ... say, an average of less than 250Kb/s. Does this answer your question? Question 1b: Not sure what you are asking. Any quarantine queries are against the single database. We synchronize the black/white lists using stored procedures in SQL though there are several posts in this forum about how others deal with this issue.
	The Desperado Dan Seligmann. Work: http://www.mags.net Personal: http://www.desperado.com

jerbo128 Members Profile Send Private Message Find Members Posts Add to Buddy List Senior Member Joined: 06 March 2006 Status: Offline Points: 178	Post Options Post Reply Quote jerbo128 Report Post Thanks(0) Quote Reply Posted: 19 May 2007 at 10:45am
	Desperado... You answered my question exactly. We are looking at a 1MB connection and wondered about the traffic back to the SQL database. Do you mind if I ask - How exactly do you transport the SQL back to the database? VPN, non standard port, ??? Question 1B - I'll try to explain better. Say a user queries the database. Message "A" came from server 1 and message "B" comes from server 2. If the user flags each message to deliver, I take it that the correct server 1 or 2 will each send it's own message from the database. Server 1 will not send on messages that server 2 quarantined?? Jerbo128

LogSat Members Profile Send Private Message Find Members Posts Add to Buddy List Admin Group Joined: 25 January 2005 Location: United States Status: Offline Points: 4106	Post Options Post Reply Quote LogSat Report Post Thanks(0) Quote Reply Posted: 19 May 2007 at 11:12am
	I'll answer the question 1B. If two SpamFilter are using the same (or replicated) database, each SpamFilter instance will know whichi emails it quarantined. When a user forces the delivery of such emails, the server that quarantined it will be the one that will deliver it. This is done by assigning a different ID to each instance of SpamFilter. If you look at the SpamFilter.ini files on both servers, you should see they both have an entry: tblServersServerID=1 and each one should have an unique ID. This ID should have been automatically generated by SpamFilter on the initial setup. If you then look in the database, there's a table called tblServers. In that table you should see listed your two SpamFilter servers, along with their relative ServerID. Lastly, if you look at the emails in the tblQuarantine, you should see that there is a field called "ServerID" that we use to identify whicih server caught that email. This is how each SpamFilter keeps tracks of the emails it owns.
	Roberto Franceschetti LogSat Software Spam Filter ISP

Desperado Members Profile Send Private Message Find Members Posts Add to Buddy List Senior Member Joined: 27 January 2005 Location: United States Status: Offline Points: 1143	Post Options Post Reply Quote Desperado Report Post Thanks(0) Quote Reply Posted: 19 May 2007 at 1:32pm
	Jerbo128 Actually, I use the "standard" sql port but my firewalls only allow connections between the 2 servers and shifts the ports so that the internet traffic is actually on a non-standard port. We tried VPN but our backup location kept causing the vpn drop and the re-establish time caused timeouts. So ... bottom line, you really can just use the standard SQL port but I advice you lock down the port for only known IP's as MSSQL seems to be a favorite amongst hackers.
	The Desperado Dan Seligmann. Work: http://www.mags.net Personal: http://www.desperado.com

jerbo128 Members Profile Send Private Message Find Members Posts Add to Buddy List Senior Member Joined: 06 March 2006 Status: Offline Points: 178	Post Options Post Reply Quote jerbo128 Report Post Thanks(0) Quote Reply Posted: 19 May 2007 at 3:39pm
	Thanks to both of you. Doesn't sound so hard now. I appreciate the help. jerbo128

LogSat Software

Site Navigation[Skip]

Spam Filter ISP Support Forum

Multiple SF’s over slow link?