1776 still seems to be leaking memory in reverse proxy mode

[yusufg@…]

Platform Centos 4.4/x86-64, lighty compiled from svn trunk

Backend, nginx (listening on 127.0.0.1:80) serving a 50M file generated via crfile (from autobench) http://www.xenoclast.org/autobench/

lighty [1776] from svn trunk configured as rproxy to this nginx

server.max-keep-alive-requests = 10
server.max-keep-alive-idle = 5
server.max-read-idle = 60
server.max-write-idle = 360
server.event-handler = "linux-sysepoll"
server.network-backend  = "linux-sendfile"
server.max-fds = 16384
server.modules  += ( "mod_proxy_backend_http" )
$HTTP["url"] =~ "^/" {
    proxy-core.balancer = "sqf" 
    proxy-core.protocol = "http" 
    proxy-core.backends = ( "127.0.0.1:80" )
    proxy-core.max-pool-size = 32
}

Before curl

21768 pts/3    Sl+    0:00      0   192 60067 1428  0.1 ./sbin/lighttpd -D -f ./proxy.conf

After curl -o /dev/null http://ip.add.re.ss/50M

21768 pts/3    Sl+    0:00      0   192 63451 4996  0.4 ./sbin/lighttpd -D -f ./proxy.conf

[yusufg@…]

output from valgrind --leak-check=yes --show-reachable=yes

[jan]

All you see here is the increase of the initialized caches.