LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

Re: FreeS/WAN Cluster - any experiences?

To: lvs-users@xxxxxxxxxxxxxxxxxxxxxx
Subject: Re: FreeS/WAN Cluster - any experiences?
From: Roberto Nibali <ratz@xxxxxxxxxxxx>
Date: Tue, 19 Feb 2002 21:15:50 +0100
        They are already exported. But you have to teach the WLC
scheduler to which estimator to apply the weights: cps, inpps, outpps,
inbps, outbps. Currently, it is to actconns&inactconns but this is
something like cps. May be you can add additional scheduler, say
ip_vs_wlt.c (Weighted Least Traffic). The problem is that you

BTW, while reading wlc I fixed some typos. Maybe you want to apply this:

diff -ur ipvs-0.9.10/ipvs/Makefile ipvs-0.9.10a/ipvs/Makefile
--- ipvs-0.9.10/ipvs/Makefile   Tue Feb 19 01:46:32 2002
+++ ipvs-0.9.10a/ipvs/Makefile  Tue Feb 19 01:50:28 2002
@@ -1,4 +1,4 @@
-# ipvs: IP Virtaul Server Module for NetFilter framework
+# ipvs: IP Virtual Server Module for NetFilter framework

 # uncomment the following line on an SMP system
 #SMPFLAGS= -D__SMP__
diff -ur ipvs-0.9.10/ipvs/ip_vs_wlc.c ipvs-0.9.10a/ipvs/ip_vs_wlc.c
--- ipvs-0.9.10/ipvs/ip_vs_wlc.c        Fri Oct 19 17:05:17 2001
+++ ipvs-0.9.10a/ipvs/ip_vs_wlc.c       Tue Feb 19 01:52:19 2002
@@ -67,9 +67,9 @@

        /*
         * We think the overhead of processing active connections is fifty
- * times than that of inactive conncetions in average. (This fifty - * times might be not accurate, we will change it later.) We use - * the following formula to estimate the overhead: + * times higher than that of inactive connections in average. (This + * fifty times might not be accurate, we will change it later.) We + * use the following formula to estimate the overhead:
         *                dest->activeconns*50 + dest->inactconns
         * and the load:
         *                (dest overhead) / dest->weight



Cheers,
Roberto Nibali, ratz



<Prev in Thread] Current Thread [Next in Thread>