Apache runs in worker mod. Multiple vhosts are configured. On of the vhosts runs in a loop, spawning childs. I want to kill this one while keeping the other good vhosts running. An apache2ctl restart is graceful and thus the looping requests of the bad vhost are not killed. apache2ctl stop stops the bad requests of the badly configured vhost, but also the other good vhosts.
Is there a way to just stop the bad vhost (or the bad requests of it)?
There are not workers for specific vhosts. Any worker can handle any request for any vhost.
But if you are having a problem with a vhost, you can certainly remove its configuration and then run apachectl graceful (or apache2ctl on Debian systems). No need to do a full restart. But requests for that virtual host would then be processed by the default virtual host, which might not be what you want.
Instead of reconfiguring Apache you should resolve the underlying problem (which you didn't really describe).
proxy_util.c ... initialized pool in child $FQDNproxy_util.c ... initializing worker $URLorproxy_util.c ... initialized single connection worker in child $FQDN. These are not requests but some kind of preconfiguring of workers. The $FQDN resp. $URL is taken fromProxyPassor/andProxyPassReverseof theVirtualHostconf. If one wants to stop seeing this logs one has to remove the vhost or outcomment theProxyPassdirectives before restarting. Don't know if this question is valid, though, I will leave it open.