cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
333
Views
0
Helpful
1
Replies

CSS 11800 show many server up/down message in log file

vincent1103
Level 1
Level 1

My customer use css 11800 running 6.10 Build 304 software for server load-blance

I config default http keepalive to probe the server status like below.

service WWW2

ip address 163.29.x.x

keepalive type http

active

service WWW3

ip address 163.29.x.x

keepalive type http

active

Since last week,I found many server up/down message in sys.log file

SEP 20 01:22:36 7/1 1145 NETMAN-2: Enterprise:Service Transition:dpsvr2 -> down

SEP 20 01:22:41 7/1 1146 NETMAN-2: Enterprise:Service Transition:amd52 -> down

SEP 20 01:24:26 7/1 1147 NETMAN-5: Enterprise:Service Transition:amd52 -> alive

SEP 20 01:24:56 7/1 1148 NETMAN-5: Enterprise:Service Transition:dpsvr2 -> alive

SEP 20 03:20:06 7/1 1149 NETMAN-2: Enterprise:Service Transition:WWW3 -> down

SEP 20 03:20:16 7/1 1150 NETMAN-5: Enterprise:Service Transition:WWW3 -> alive

SEP 20 07:00:57 7/1 1151 NETMAN-2: Enterprise:Service Transition:www5 -> down

SEP 20 07:01:11 7/1 1152 NETMAN-5: Enterprise:Service Transition:www5 -> alive

The server status change down to up,almost during 20 sec only,and We can access the server without any problem at the same time

Does this is a software bug.If not,How do I trace this problem,thks!!

1 Reply 1

Gilles Dufour
Cisco Employee
Cisco Employee

capture a sniffer trace to see exactly what's going on.

Going to the server with a browser may work because using GET/POST method, but the CSS configured the way you did will use a HEAD method which could be a problem for your server.

The server could be slow to respond as well.

Try to increase the keepalive frequence since the timeout is linked to the frequency.

Finally, you may want to configure a url to poll with the command 'keepalive uri .....'

Once again, by simply sniffing between the CSS and the server, you will know exactly where is the problem.

Gilles.