LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

[lvs-users] [LVS - NAT] InActConn in Ipvsadm with UDP stop incrementing

To: "LinuxVirtualServer.org users mailing list." <lvs-users@xxxxxxxxxxxxxxxxxxxxxx>
Subject: [lvs-users] [LVS - NAT] InActConn in Ipvsadm with UDP stop incrementing during load test, and message are not correctly routed
From: "pierrick grasland" <pierrick.grasland@xxxxxxxxx>
Date: Wed, 26 Mar 2008 14:28:08 +0100
Hello,

After correcting my hardware problem, I rerun my load test on LVS with a
simple test for UDP.

Actually, my setup is just the following :

Client (CIP : 192.168.1.209) ==> Director  (VIP : 192.168.1.231, DIP:
192.168.2.231) ==> realserver (Only one, RIP : 192.168.2.3), with the
following configuration :

Client : XP, using SIPp in order to generate a lot of SIP messages in
parallel (500 differents sockets open at max). These message are a single
UDP packet, with only 425 Bytes each
Director : debian etch, kernel 2.6.18 + ipvsadm 1.24

ipvsadm setup :

-A -u 192.168.1.231:sip -s rr
-a -u 192.168.1.231:sip -r 192.168.2.3:sip -m
-A -t 192.168.1.231:sip -s rr
-a -t 192.168.1.231:sip -r 192.168.2.3:sip -m
-A -t 192.168.1.231:www -s wlc
-a -t 192.168.1.231:www -r 192.168.2.3:www -m

realserver : the corresponding service, which use cipango, a sip servlet
container to generate valid SIP response (200 OK), with only 300 bytes in
each response.

I realize a first test, using TCP, in order to validate my configuration.
With this test, I'm obtaining the following result : (see attachment,
TCP_test.txt). In this case, LVS show perfect result, even with a high load
rate (4000 msg / seconds).
But, when I'm trying the same test, with UDP, I actually have very bad
result (between 500 and 3000 successful call for test with ten of thousands
calls). Results are UDP_test1.txt, UDP_test2.txt and UDP_test3.txt

When I'm check with ipvsadm, I can view the following (from situation
UDP_test3.txt) :

director:/home/pierrick# ipvsadm
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn
UDP  director.local:sip rr
  -> 192.168.2.3:sip              Masq    1      0          499
TCP  director.local:sip rr
  -> 192.168.2.3:sip              Masq    1      0          0
TCP  director.local:www wlc
  -> 192.168.2.3:www              Masq    1      0          0

It's seems that LVS doesn't increase InActConn anymore, and, in SIPp, only
the first calls are correctly routed, all others finished as timeout.

Did someone encounter the same behavior ?


-- 
Grasland Pierrick
ENSSAT - LSI 3

Attachment: TCP_test.txt
Description: Text document

Attachment: UDP_test1.txt
Description: Text document

Attachment: UDP_test2.txt
Description: Text document

Attachment: UDP_test3.txt
Description: Text document





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