iRule logging [LB::server pool] crashes tmm on 11.4.1 HF3

Hi Josh,

 

it´s a common approach to use [LB::server pool] in the context of

CLIENT_ACCEPTED

to store the virtual server´s default pool to a variable.

 

By now I haven´t seen this function causing TMM restarts.

 

Did you check /var/core/ for core dumps?

 

I would recommend to make the behavior reproducible, i.e. by reducing to a single event with i.e.

 

when CLIENT_ACCEPTED {
   log local0.debug "TEST: start pool in rule CLIENT_ACCEPT [LB::server pool]"
}

Does it still cause TMM restarts after establishing a new connection to the virtual server?

 

Perhaps it´s not the LB-command but the usage of procedures?

 

As you are using the

HTTP_REQUEST

context I guess you have an http-profile or a fasthttp-profile assigned.

 

To parse http, it would be required to terminate SSL, in case you are using encrypted traffic (only available in virtual server standard mode operation).

 

What type of virtual server are you using, what profiles are assigned (

tmsh list ltm virtual

)?

 

Do you have a default pool configured?

 

Did you try iHealth to validate your configuration?

 

Thanks, Stephan

 

PS: Formatting a post can be frustrating, I agree. For long posts I used to put everything into a plaintext editor (no formatting, no indent for first and last line of code sections, no text bullets etc.). Now paste in and format each section separately.

 

Sometimes indented code or paragraphs seem to be formatted automatically. Please remove and re-insert allowing you to format manually.