TLM configuration for t3 protocol

Forums Technical Forums
,
1

Hi Experts,

we are configuring the TLM for the loadbalancing weblogic servers. we have the Javaclient that require to communicate with the weblogic through t3 protocol. hence i would like to know how to configure the TLM with t3 protocol.

thanks

Arun

2

Hi Arun,

If you search the forums for “T3” you should find a few related threads. If you try something and get stuck, reply back with what you’ve tested and the symptoms of the issue.

Aaron

3

Hi Aaron,

i have gone through all forums,but couldnt able to resolve the issue.

Config:

============

we have 2 weblogic servers and those deatils are updated on the TLM . when i access the Agile Webclient it is passing through the TLM successfully(it is not using the t3 protocol),but when i access through the Javaclient- JNLP (it uses t3 to communicate with weblogic). Below is the configuration on the TLM

virtual server - aglprd.vf.com with port 80

type - Perforamnce layer 4

pools - aglprd with weblogic members

Issue:

============

when i access the Javacleint through the LTM http://aglprd.vf.com:80/JavaClient/start.html ,i am getting the below error on the Javaconsole

“javax.naming.CommunicationException [Root exception is java.net.ConnectException: t3://10.180.159.15:80: Bootstrap to: 10.180.159.15/10.180.159.15:80’ over: ‘t3’ got an error or timed out]”

and i could see below error on the weblogic logs

Jan 19, 2011 4:39:38 AM PST Error RJVM vfilvaglwl1.verifone.com vfilvaglwl1.verifone.com-ManagedServer1 ExecuteThread: ‘1’ for queue: ‘weblogic.socket.Muxer’ WLS Kernel 1295440778225 BEA-000572 The server rejected a connection attempt JVMMessage from: ‘-1231415493576212652C:192.30.209.111R:-3942184265094917794S:vfilvaglwl1.verifone.com:vfilvaglwl1.verifone.com:agileDomain:vfilvaglwl1.verifone.com-ManagedServer1’ to: ‘0B:10.70.49.35:[80,-1,-1,-1,-1,-1,-1]’ cmd: ‘CMD_IDENTIFY_REQUEST’, QOS: ‘101’, responseId: ‘-1’, invokableId: ‘-1’, flags: ‘JVMIDs Sent, TX Context Not Sent, 0x1’, abbrev offset: ‘106’ probably due to an incorrect firewall configuration or admin command.

thanks for your response.

Arun

4

Hi Arun,

The forum code here strips out text between angle brackets. Can you remove the angle brackets or put spaces between them like this < text_with_space_padding >?

Thanks, Aaron

5

Hi Aaron,

Thanks for the comment. i have updated with log details now.

thanks

Arun

6

Are the pool members also listening on port 80? If not, do they use their own port in references to themselves instead of port 80? If you run a tcpdump on LTM looking for the test client IP and the server IP, what do you see? You can use syntax like this:

tcpdump -nni 0.0 -s0 host CLIENT_IP or host POOL_MEMBER_IP

Here is a recent post on T3:

http://devcentral.f5.com/Community/GroupDetails/tabid/1082223/asg/52/afv/topic/aft/1174020/aff/2247/showtab/groupforums/Default.aspx

Aaron

7

Hi,

I will check on the tcp details and update .

we have 2 members for the pool , which is having 2 different ports. From the link you have given i could see port 7001 is being used & is there any mandatory that we need to go for port 7001 for t3 protocol including memers of the pool ?

thanks

Arun

8

You might try testing with the virtual server defined on the same port as the pool members to see if that works. If that does work, but it fails once you change the VS or pool member ports to be different from each other, than it’s likely an issue with the ports being hard coded in response headers and/or payload. You could either try changing the server config to use the VS port or use an iRule to rewrite the port references.

Reply back once you have more info and we can give you more detailed suggestions.

Aaron

9

Hi Arun,

We are facing same issue.

Configured VS listening on 80 with type as performance layer 4 and pool members are configured on Tcp port 7002 .

Is the probelm mentioned earlier resolved by configuring the VS Port and pool members as same ?

Please provide me with the recommended configuration if your issue has been resolved.

Thanks

Raj

10

Hi Raj,

Can you try testing the suggestion and letting us know what you see?

Aaron

11

Hi Aaron,

The problem was resolved after changing the virtual server port same as pool member port.

Thanks for you thoughts

Also - I am facing other issue with the persistence using the jsession id from the http path .

Could you please look in to ths issue as well.

http://devcentral.f5.com/Community/GroupDetails/tabid/1082223/aff/5/afv/topic/aft/1179064/asg/50/Default.aspx

Thanks

Raj

12

Hi Raj,

Thanks for confirming the change worked.

Aaron

13

that is behaviour of the weblogic not issue on the f5

“By default, WebLogic expects any remote T3 access to have been initiated by the client, using the same port number that the proxied request actually hits a port of the WebLogic server on”

on weblogic side need to run with this parameter:
-Dweblogic.rjvm.enableprotocolswitch=true

here’s a good read:

pauldone.blogspot. com/2013/04/load-balancing-t3-initial-context.html

14

you replied to a post from 15 years back by the way, not sure if he will be seeing that, just saying.