hi, > we have a similar setup, but we use multiple ipvs + ldirector setups > for this, where the most bandwidth we push through any given pair is > ~500 mbit or so of sustained traffic.
that sounds equal. nice. > the way I see it, if you've got 2200 machines, you can afford to have > multiple gateways into your network... we do have them; it's not 2,200 _physical_ machinesm, btw, it's 2,200 IPs. > that or you've done something completely wrong and have far too much > hardware for far too little traffic / processing not really; we're going to replace existing load balancers, hence the setup. best, timo > Geoff Harrison > Chief Architect > The Hive > (617) 301-6200 > > > On Apr 17, 2009, at 4:41 AM, Timo Schoeler wrote: > >> Hello list, >> >> it seems that the first issue I had with ldirectord [0] is mostly >> solved; that given, I'm about to deploy another LVS box in a different >> environment. I did read the 'what is the max throughput LVS can >> handle, >> and max number of concurrent connection?' thread, but that's LVS, not >> ldirectord... >> >> That raises the question for me if anybody has a similar (or even >> bigger) setup in production. >> >> To give a few numbers: Virtual IPs will be about 600, real servers are >> almost 2200. Services should be tested every 2 or 5 seconds. >> >> Any takers? >> >> Thanks in advance, >> >> Timo >> >> [0] -- http://tinyurl.com/cz76dl _______________________________________________ Please read the documentation before posting - it's available at: http://www.linuxvirtualserver.org/ LinuxVirtualServer.org mailing list - [email protected] Send requests to [email protected] or go to http://lists.graemef.net/mailman/listinfo/lvs-users
