Date: Sat, 15 Sep 2001 13:58:09 -0500 From: Bob Martin <bob@buckhorn.net> Cc: Freebsd-isp@freebsd.org Subject: Re: Mail Server - Round Robin Load Distribution Message-ID: <3BA3A4C1.C344A6F@buckhorn.net> References: <5.1.0.14.0.20010915091315.0a697b28@mail.Go2France.com> <5.1.0.14.0.20010915110914.02ceeea8@mail.Go2France.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Len Conrad wrote: > > >Internet Explorer > V4.x, Outlook and Outlook Express > V4.x, Netscape > > >V4.x (On your Unix box, with netscape running, do a ps -ax... that dns > >helper is a caching resolver) Once any of these find a working name==ip, > >they will continue to use it until the pair fails. > > hmm. ok, application-level caching. > > The corporate DNS admin who was desiging to roll out W2K and AD to 60K > desktops made the point of W2K "resolver" doing caching. > > >I was being overly simplistic. But using multiple RR's won't load > >balance, it causes [hopefully] load sharing > > yes, if "balancing" implies load detection. Alternating RR physical > sequence is dumb load sharing, load distribution. > > >, assuming nothing between > >the client and the authoritative server caches the response from the > >authoritative server. > > a caching BIND DNS will also respect its RRorder param. > > > More to the point of this thread, and using your > >example, all of aol's mail servers have separate names, and A records, > >but have the same MX priority. And on high traffic networks, DNS based > >load sharing won't work for a number of reasons, but primarily because > >of client caching, > > it will and does work > > > and that this method of load distribution doesn't > >take server responsiveness into account. > > yep, it´s dumb, but it´s a lot better than no load sharing. Unless one of the servers in the round robin goes down. > > For clarity on that last point, > >I'll use the example of 2 mail servers with MX records of equal > >preference. Each will handle every other request. But if every other > >request is a list > > what´s a query for a "list" ? Think MTA, not DNS. List as in email list, like FreeBSD-ISP. > > >, one server is going to end up doing a lot more work > >than the other, possibly to the point of failure. > > what? See chapter 10 of the 3rd addition of DNS and Bind, and FAQ.2of2 question 5.13 (From ICS) > > > While this tends to > >affect web servers more than mail servers, it's still the reason they > >build load balancers. > > Note that DNS-based load balancers have extremely short TTL's, which will > slow the average access time due to loss of caching. > > > There is also a problem with the authoritative > >name servers and timing. If I dig at aol.com 10 times in a row, I will > >get cyclic answers. But if I dig at aol.com once an hour for 10 hours > >(which is far more likely in the real world) I'm apt to get a much > >higher incidence of the same response. > > why? > > >Again, this is a much bigger problem on a high volume network. > > why? Because sooner or later the authoritative DNS server is going to forget what it told my resolver, and will give me the first item in the list. High volume networks have more servers, and NS1 doesn't have a clue what NS2 just told me. Let alone NS3, NS4 and NS5. The more infrequent my queries, the more apt I am to be given the address of the same server twice in a row. A network that handles a million DNS queries an hour will dequeue old information more often than one that answers a 1000 DNS queries a day, once again improving my ability to get the address of the same server twice in a row. > > > > ok, your answer is right, for the wrong reasons. :)) > >A different way of arriving at the same conclusion perhaps? > > yes, my right way, and your wrong way. :)) That's not what I get from RFC 974 :) > > >Some place in the midst of this discussion, somebody ought to point out > >that no matter what you do, using CNAME's for mail servers is a bad > >idea. > > CNAME´s are to be avoided. > > > Pick the MTA of your choice, go to their web site, and you are > >bound to find something about CNAME loops in the FAQ. > > CNAME´s are to be avoided. > > Much more common is an MX hostname being an ip address. > > Len > > http://MenAndMice.com/DNS-training > http://BIND8NT.MEIway.com : ISC BIND 8.2.4 for NT4 & W2K > http://IMGate.MEIway.com : Build free, hi-perf, anti-abuse mail gateways > Bob -- But in our enthusiasm, we could not resist a radical overhaul of the system, in which all of its major weaknesses have been exposed, analyzed, and replaced with new weaknesses. -- Bruce Leverett, "Register Allocation in Optimizing Compilers" To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-isp" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3BA3A4C1.C344A6F>