Hi Malcolm, On Tue, Mar 08, 2011 at 10:07:05PM +0000, Malcolm Turnbull wrote: > I have a customer with a MS Sharepoint configuration using > > contimeout 4000 > clitimeout 3600000 > srvtimeout 3600000 > > listen MOSS 10.11.12.6:80 > bind 10.11.12.6:443,10.11.12.6:25 > mode tcp > option tcplog > option persist > balance leastconn > stick-table type ip size 10240k expire 30m > stick on src > server ifesdcsps02 10.11.12.62 weight 0 check port 443 inter 2000 rise 2 fall > 3 > server ifesdcsps01 10.11.12.61 weight 1 check port 443 inter 2000 rise 2 fall > 3 > server backup 127.0.0.1:9081 backup > option redispatch > option abortonclose > maxconn 40000 > log global > > > But he keeps on getting some clients randomly disconnect (usually > after about 10-15mins?) > The log just shows several SD (Server Disconnects) at the time of the > probloem: > > Mar 8 13:32:50 lbmaster haproxy[4269]: 10.11.13.28:49669 > [08/Mar/2011:13:28:57.348] MOSS MOSS/ifesdcsps01 0/0/163085 196445 SD > 268/6/6/6/0 0/0 > Mar 8 13:32:50 lbmaster haproxy[4269]: 10.11.13.28:49681 > [08/Mar/2011:13:29:30.769] MOSS MOSS/ifesdcsps01 0/1/129665 2491 SD > 269/7/7/7/0 0/0 > Mar 8 13:32:50 lbmaster haproxy[4269]: 10.11.13.28:49680 > [08/Mar/2011:13:29:30.766] MOSS MOSS/ifesdcsps01 0/1/129668 2491 SD > 270/8/8/8/0 0/0 > Mar 8 13:32:50 lbmaster haproxy[4269]: 10.11.13.28:49679 > [08/Mar/2011:13:29:30.756] MOSS MOSS/ifesdcsps01 0/0/129678 3415 SD > 271/9/9/9/0 0/0 > > It may well be a server config / windows authentication issue but am I > missing something obvious in the log or the configuration? > > Thanks very much in advance, this has puzzled me.
What I suspect is that the client closes normally (eg: FIN) and the server simply responds with an RST, and for this reason is identified as an abort or error. It would be nice to check in a network capture if you see resets emitted from the server which seem to correlate to what you're seeing above. Regards, Willy
