http://bugzilla.spamassassin.org/show_bug.cgi?id=3983
------- Additional Comments From [EMAIL PROTECTED] 2005-01-07 21:54 -------
i have this running on many systems and its working great. however, i have one
that is a problem, and it really the only oddball out of the bunch. its a dual
hyperthreaded zeon box as opposed to all the others being uniprocessor P4's.
here are the details...
Spamd stops accepting connections after processing several thousand messages.
- new apache forking model
- running smp kernel, dual ht zeons
using the following settings...
max_children 15
max_per_child 25
max_spare 5
min_children 5
min_spare 2
when spamd stops accepting connections, an strace shows a hang on connect()..
# echo | strace spamc -t 5
..
..
socket(PF_INET, SOCK_STREAM, IPPROTO_TCP) = 3
connect(3, {sin_family=AF_INET, sin_port=htons(783),
sin_addr=inet_addr("127.0.0.1")}}, 16 <unfinished ...>
tcpdump shows the syn going to spamd, but spamd never acks back.
as soon as i svc -d and svc -u on /service/spamd it is available and working
again. if i try an hup it with a svc -t, that doesnt help.
last few debug lines related to preforking..
@4000000041df61e201060fe4 [10653] debug: prefork: child 21075: entering state 1
@4000000041df61e201077f14 [10653] debug: prefork: child reports idle
@4000000041df61e201084264 [10653] debug: prefork: child 21298: entering state 1
@4000000041df61e201099254 [10653] debug: prefork: child reports idle
@4000000041df61e2010b8e24 [10653] debug: prefork: child states: IIBIBIIBIIIBBBI
@4000000041df61e2010d9d7c [10653] debug: prefork: child 21298: entering state 3
@4000000041df61e2010f2034 [10653] debug: prefork: adjust: decreasing, too many
idle children (9 > 5), killed 21298
@4000000041df61e2017c841c [10653] debug: prefork: child 20945: entering state 1
@4000000041df61e2017ef134 [10653] debug: prefork: child reports idle
@4000000041df61e2017ff304 [10653] debug: prefork: child closed connection
@4000000041df61e20185752c [10653] debug: prefork: child states: IIBIBIIIIIIBBB
@4000000041df61e201873e34 [10653] debug: prefork: child 21085: entering state 3
@4000000041df61e201890f0c [10653] debug: prefork: adjust: decreasing, too many
idle children (9 > 5), killed 21085
@4000000041df61e201f895d4 [10653] debug: prefork: child closed connection
@4000000041df61e201fb06d4 [10653] debug: prefork: child states: IIBIBIIIIIBBB
@4000000041df61e201fc3b6c [10653] debug: prefork: child 21081: entering state 3
@4000000041df61e201fd838c [10653] debug: prefork: adjust: decreasing, too many
idle children (8 > 5), killed 21081
@4000000041df61e202cb5adc [10653] debug: prefork: child closed connection
@4000000041df61e20b7403b4 [10653] debug: prefork: child states: IIBIBIIIIKBBB
@4000000041df61e20b741b24 [10653] debug: prefork: child 21075: entering state 3
@4000000041df61e20b7422f4 [10653] debug: prefork: adjust: decreasing, too many
idle children (7 > 5), killed 21075
@4000000041df61e20b743a64 [10653] debug: prefork: ordered 18886 to accept
@4000000041df61e2197aaea4 [10653] debug: prefork: child 18886: entering state 2
@4000000041df61e2197b6e0c [10653] debug: prefork: child closed connection
@4000000041df61e2197d044c [10653] debug: prefork: child 21087: entering state 1
@4000000041df61e2197e3ccc [10653] debug: prefork: child reports idle
@4000000041df61e2197f4a54 [10653] debug: prefork: child 21095: entering state 1
@4000000041df61e2198082d4 [10653] debug: prefork: child reports idle
@4000000041df61e219815d94 [10653] debug: prefork: child 21096: entering state 1
@4000000041df61e21982a99c [10653] debug: prefork: child reports idle
@4000000041df61e2198499b4 [10653] debug: prefork: child states: BIBIBIIIIII
@4000000041df61e21986531c [10653] debug: prefork: child 21096: entering state 3
@4000000041df61e21988089c [10653] debug: prefork: adjust: decreasing, too many
idle children (8 > 5), killed 21096
last debug line..
@4000000041df61e30180e14c [18886] debug: log: [18886] child is processing
message 24 of 25
another time it happend, the last debug line was
@4000000041df028027c4ee5c [4250] debug: log: [4250] child is processing message
3 of 25
so it does the same thing every time it hangs it seems.
is anyone else running this without issues on SMP enabled kernels?
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
