Description of Problem: When using the same real server in several virtual servers hosted on the same machine, pulse does not start multiple nannies for this real server, but just one matching the first virtual server. Thus, there is no ipvsadm entry for this real server except for the first declared virtual server. Version-Release number of selected component (if applicable): piranha 0.5.5-8 ipvsadm 1.17 ipvs 0.8.1 How Reproducible: anytime Steps to Reproduce: 1. setup a piranha cluster 2. Use a lvs.conf somewhat like : virtual http1 { active = 1 address = 172.22.48.210 eth0:1 vip_nmask = 255.255.240.0 port = 80 scheduler = wrr protocol = tcp server lilas { address = 172.16.0.101 active = 1 weight = 3 } server pcpgn { address = 172.16.0.102 active = 1 weight = 2 } } virtual http2 { active = 1 address = 172.22.48.209 eth0:2 vip_nmask = 255.255.240.0 port = 80 scheduler = wlc protocol = tcp server lilas { address = 172.16.0.101 active = 1 weight = 1 } server pcpgn { address = 172.16.0.102 active = 1 weight = 1 } } 3. Start pulse, and then notice that there are only two nannies running, instead of four.
Created attachment 41611 [details] Patch to start one nanny for each (VIP,RIP,port,proto) instead of one for each (RIP,port,proto). Also correctly handle SIGHUP in pulse.
*** Bug 15911 has been marked as a duplicate of this bug. ***
Danke,.. applied these patchs to 0.6.0-15 so you should see the results in 0.6.0-16 and higher Cheers Phil =--=