Hi, I''ve followed the instructions in the nano document, and apart from my own ISP trouble, it seems to work fine. I have a question about the dgd however. The dgd works on the actual gateway, but in my experience it seems that most problems can be expected a few hops down the route, not on the gateway. One of those two gateways is located here anyway, so if that goes down then chances are other stuff is wrong as well ;) So my question is: is there some way that I can detect that the actual route to the net is unusable, even though the gateway is up ? If not, when either one of the lines is down, users will notice they''re not getting a connection randomly, depending on what gateway is used for that connection. Thomas -- The Dave/Dina Project : future TV today ! - http://davedina.apestaart.org/ <-*- -*-> Lover fair We''ll be looking sharp I swear I want them all to stop and stare When we take''em down <-*- thomas@apestaart.org -*-> URGent, the best radio on the Internet - 24/7 ! - http://urgent.rug.ac.be/
Thomas Vander Stichele(thomas@urgent.rug.ac.be)@Wed, Mar 27, 2002 at 10:45:10AM +0100:> Hi, > > I''ve followed the instructions in the nano document, and apart from my own > ISP trouble, it seems to work fine. I have a question about the dgd > however. > > The dgd works on the actual gateway, but in my experience it seems that > most problems can be expected a few hops down the route, not on the > gateway. One of those two gateways is located here anyway, so if that > goes down then chances are other stuff is wrong as well ;) > > So my question is: is there some way that I can detect that the actual > route to the net is unusable, even though the gateway is up ? If not, when > either one of the lines is down, users will notice they''re not getting a > connection randomly, depending on what gateway is used for that > connection. > > ThomasIt sounds to me like you need something which is proactively monitoring your network health/status. The points in the network your referring to, are not under your sirect control, so there''s not a lot tc/ip can directly do to solve your problem. but with some clever scripting, you could attempt to detect a fault in the network, and take appropriate action. such as delete the faulty route for all the clients that were using it, and forec them onto the other route. you''ll still need to be able to test the dead route from the box, such that you can activate again, once its working. i hope i understood the problem correctly, and havent gone off on one. -vince -- PGP key: http://codex.net/pgp/pgp.asc Computers make very fast, very accurate mistakes.
> > So my question is: is there some way that I can detect that the actual > > route to the net is unusable, even though the gateway is up ? If not, when > > either one of the lines is down, users will notice they''re not getting a > > connection randomly, depending on what gateway is used for that > > connection. > > > > Thomas > > It sounds to me like you need something which is proactively monitoring > your network health/status. The points in the network your referring > to, are not under your sirect control, so there''s not a lot tc/ip can > directly do to solve your problem. > > but with some clever scripting, you could attempt to detect a fault in > the network, and take appropriate action. such as delete the faulty > route for all the clients that were using it, and forec them onto the > other route. you''ll still need to be able to test the dead route from > the box, such that you can activate again, once its working. > > i hope i understood the problem correctly, and havent gone off on one.No, you understood it correctly, and this is what I used to do. However, i was looking for a better solution and the load balancing is, in a normal situation, better. What I wanted to know was, how can I combine this scripting that checks for usable routes to the net with the current nano setup I''m using. I would like for it to integrate cleanly, ie. not having to change routes on the fly each time, but instead maybe mark the gateway as dead by hand, for example. If it''s not possible, I might as well just go back to the older setup without the load balancing, because in that case we just routed some of the traffic over one gateway and some of the other traffic over another, folding stuff back over the working line if one of the went down. I just really like the cleanliness of the nano approach with julian''s patches, and would like to add on that instead ;) Thomas -- The Dave/Dina Project : future TV today ! - http://davedina.apestaart.org/ <-*- -*-> I have these hands teeming with love for you But you''re not here to touch You said you''d wait but it''s killing me When I need something that much <-*- thomas@apestaart.org -*-> URGent, the best radio on the Internet - 24/7 ! - http://urgent.rug.ac.be/