External TCP port monitor 5150

Hi Leo D,

 

 

I am guessing that this will be an HTTP Monitor running a get?

 

 

If so, just input your Send String, Recieve String, and Alais Service Port and it should work.

 

 

Example:

 

Send String: GET /server/f5.html\r\n

 

Send String: 0

 

Alias Port: 5150

 

 

I would suggest creating a dummy pool to put the target server into and then apply the Health Monitor to it for testing. When you verify that it is working properly you can apply the monitor to the Pool (or the Node) and set the Availability Requirement to All.

 

 

Then if either Health Monitor Fails it will remove the node from the pool.

 

 

Hope this helps.

 

Hi Leo,

 

 

You could potentially write an inbuilt TCP monitor for this. What's the string the client needs to send to get the status of the app?

 

 

An inbuilt monitor will be more efficient than an external monitor. See this SOL for details on writing a custom inbuilt TCP monitor:

 

 

SOL4186 - Entering escape sequences in an HTTP health monitor's Send and Receive strings

 

https://support.f5.com/kb/en-us/solutions/public/4000/100/sol4186.html

 

 

Aaron

 

Michael,

 

Thanks for the reply. I did not think of doing it with an HTTP get, I was thinking about using a custom build external TCP monitor. If an HTTP GET will work as in your post that would be great. I just need to determine the TCP response on port 5150. Would the node I monitor need to be running an HTTP server to respond to the HTTP GET? I am thinking it would, currently the server is not running but I am checking to see if it can. I also appreciate the advice on the dummy pool as I do not have a Dev environment for this testing.

 

 

 

Leo

 

Aaron,

 

Thanks for the reply. I just run the following: telnet 10.10.10.10 5150 and after about 5 seconds it returned a 0 and indicates connection is closed. If it returns a "1" then I know that the server has some internal app issues. I would really prefer to use the existing mechanisms if possible. So I would like to maybe Telnet to the server and somehow check the responce.

 

 

Leo

 

 

e.g.

1. find response in hex when server returns 0 and 1. for example, 300a is 0 and 310a is 1 (0x30 is 0, 0x31 is 1, 0x0a is new line). 

 returning 0

[root@ve10:Active] config  telnet 200.200.200.101 5150
Trying 200.200.200.101...
Connected to 200.200.200.101 (200.200.200.101).
Escape character is '^]'.
0
Connection closed by foreign host.

[root@ve10:Active] config  tcpdump -nni 0.0 -s0 port 5150 -X
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on 0.0, link-type EN10MB (Ethernet), capture size 65535 bytes
...snippe...
00:24:49.071099 IP 200.200.200.101.5150 > 200.200.200.10.60808: P 1:3(2) ack 1 win 46  in slot1/tmm0 lis=
        0x0000:  0ffd 0800 4500 0036 9130 4000 4006 8790  ....E..6.0@.@...
        0x0010:  c8c8 c865 c8c8 c80a 141e ed88 6a64 c8c1  ...e........jd..
        0x0020:  305e 94d7 8018 002e 25ef 0000 0101 080a  0^......%.......
        0x0030:  23fc 59f1 5ded 28ad 300a 0114 0001 0000  .Y.].(.0.......
        0x0040:  0000 0000 0000 0000 0000 0000 0000 0000  ................
...snippe...

 returning 1

[root@ve10:Active] config  telnet 200.200.200.101 5150
Trying 200.200.200.101...
Connected to 200.200.200.101 (200.200.200.101).
Escape character is '^]'.
1
Connection closed by foreign host.

[root@ve10:Active] config  tcpdump -nni 0.0 -s0 port 5150 -X
tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on 0.0, link-type EN10MB (Ethernet), capture size 65535 bytes
...snipped...
00:26:19.146301 IP 200.200.200.101.5150 > 200.200.200.10.60825: P 1:3(2) ack 1 win 46  in slot1/tmm0 lis=
        0x0000:  0ffd 0800 4500 0036 4240 4000 4006 d680  ....E..6B@@.@...
        0x0010:  c8c8 c865 c8c8 c80a 141e ed99 f54f fb90  ...e.........O..
        0x0020:  647e 06bf 8018 002e 015f 0000 0101 080a  d~......._......
        0x0030:  23fd b9d0 5dee 8888 310a 0114 0001 0000  ...]...1.......
        0x0040:  0000 0000 0000 0000 0000 0000 0000 0000  ................
...snipped...

2. configure tcp health monitor

[root@ve10:Active] config  b monitor mytcp list
monitor mytcp {
   defaults from tcp
   recv "\x31\x0a"
}

3. assign monitor to pool
[root@ve10:Active] config  b pool foo list
pool foo {
   monitor all mytcp
   members 200.200.200.101:5150 {}
}

4. test

 when returning 1

[root@ve10:Active] config  b pool foo|grep -i pool\ member
+-> POOL MEMBER foo/200.200.200.101:5150   active,up

 when returning 0

[root@ve10:Active] config  b pool foo|grep -i pool\ member
+-> POOL MEMBER foo/200.200.200.101:5150   inactive,down

nitass,

 

 

Thanks, this is a big help. I plan to test this as soon as I can. I will update this post with my results!

 

 

 

Leo