Inconsistent between ipvsadm stats & realserver's

To: "' users mailing list.'" <lvs-users@xxxxxxxxxxxxxxxxxxxxxx>
Subject: Inconsistent between ipvsadm stats & realserver's
From: "Son Nguyen" <trungson@xxxxxxxxx>
Date: Wed, 13 Jul 2005 23:30:30 -0700

Why do I experience inconsistency between the ActiveConn between the stats
by ipvsadm and the stats by the real servers? There is only one application
using the VIP, each request will be saved as one row into the local (on the
real) database. Ipvsadm reports that RIP#1 is the most active but actually
RIP#2 has the most entries in the local database. Can you think of any clue
or test I could do to figure out the problem? Also why the InActConn for
RIP2/RIP3 are much higher than RIP1 while it's less active (based on the

[root@d2 root]# ipvsadm
IP Virtual Server version 1.0.8 (size=65536)
Prot LocalAddress:Port Scheduler Flags
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  VIP:http wlc
  -> RIP1:http          Tunnel  4      15         36
  -> RIP2:http          Tunnel  1      3          395
  -> RIP3:http          Tunnel  2      6          507

Database Stats (actual usage)
RIP1: 2,429 rows
RIP2: 5,052 rows -> most active 
RIP3: 2,680 rows

Best regards,
Son Nguyen

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