06-16-2011 02:19 AM
Hi,
I have created a TCP probe port 9099, but I have always have a failure
probe tcp PR_SmartBike-Antwerp_9099
port 9099
interval 10
faildetect 2
passdetect interval 60
receive 3
connection term forced
Is it possible that that the ace is still waiting for an answer from the server? If I read this I gues it is/
Initiates a TCP 3-way handshake (SYN, SYN-ACK, ACK) and expects the server to send a response. By default, a successful response causes the probe to mark the server as passed and send a FIN to end the session. If the response is not valid or if there is no response, the probe marks the server as failed.
This is some data I see
2011-06-06 19:43:12.325702 IP 172.16.56.4.4901 > 172.16.57.11.9099: S 686345311:686345311(0) win 5840 <mss 1460,sackOK,timestamp 395707542 0,nop,wscale 0>
2011-06-06 19:43:12.325841 IP 172.16.57.11.9099 > 172.16.56.4.4901: S 1975372337:1975372337(0) ack 686345312 win 8192 <mss 1460,nop,wscale 8,sackOK,timestamp 1632036 395707542>
2011-06-06 19:43:12.326028 IP 172.16.56.4.4901 > 172.16.57.11.9099: . ack 1 win 5840 <nop,nop,timestamp 395707542 1632036>
2011-06-06 19:43:12.326130 IP 172.16.56.4.4901 > 172.16.57.11.9099: R 1:1(0) ack 1 win 5840 <nop,nop,timestamp 395707542 1632036>
06-16-2011 04:45 AM
Hi Erik,
With the config you have, the ACE should establish a TCP handshake with the destination server (sends SYN, expect SYN-ACK, sends ACK) then will send a reset to the server ( because of the "connection term forced" configured).
I tested to tcpdump a 3WHS on my machine and got something quite similar to wat you had:
11:30:25.874657 IP CDN-WAVE-474-1.cisco.com.49776 > bsns-fs.cisco.com.www: SWE 696841843:696841843(0) win 5840
11:30:25.874930 IP bsns-fs.cisco.com.www > CDN-WAVE-474-1.cisco.com.49776: SE 2006130139:2006130139(0) ack 696841844 win 5792
11:30:25.874973 IP CDN-WAVE-474-1.cisco.com.49776 > bsns-fs.cisco.com.www: . ack 1 win 23
To see the reason why the probe is failing, maybe you could check the output of the "show probe PR_SmartBike-Antwerp_9099 detail"?
This should give you the reason the probe is failing under the "Last disconnect err" part of the output.
If it doesn't, maybe you could take a pcap capture of the communication between the ACE and the backend server to see what is happening at packet level?
Regards,
Nicolas
06-16-2011 01:24 PM
from the failed ACE view, is your rserver behind another ACE, or some kind of applicationn FW that creates a reverse proxy?
Telnet from the ace to your real server over the probe port an post the pcap file plz. with the caps taken from de rserver or from a monitor session over the switch connecting to this rsever.
The-great-l0k1
mx
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide