In message <20111221083337.gb5...@fantomas.sk>, Matus UHLAR - fantomas writes: > >In message <20111220182222.ga3...@fantomas.sk>, Matus UHLAR - fantomas write > s: > >> Long time ago when we were trying to have multiple web servers for > >> redundancy and balancing, we have found that multiple IP's is not a > >> good solution (parts of web pages didn't load). We selected L3 > >> switches then... > > On 21.12.11 09:26, Mark Andrews wrote: > >Which is really the result of badly designed clients. Clients are getting > >better with address affinity and fast failover on unreachable servers. > > It's been long time ago (~10 years). And even if they did failover, 30s > (tcp connection timeout) delays are very ugly when loading a web page.
Indeed. 150-250ms [1] is a more realistic timeout for starting a second connection attempt. You use the connection which completes first and close the others if they complete. Mark [1] http://tools.ietf.org/html/draft-ietf-v6ops-happy-eyeballs-07 -- Mark Andrews, ISC 1 Seymour St., Dundas Valley, NSW 2117, Australia PHONE: +61 2 9871 4742 INTERNET: ma...@isc.org _______________________________________________ Please visit https://lists.isc.org/mailman/listinfo/bind-users to unsubscribe from this list bind-users mailing list bind-users@lists.isc.org https://lists.isc.org/mailman/listinfo/bind-users
