Forum Discussion

gneville's avatar
Icon for Nimbostratus rankNimbostratus
Feb 07, 2024

tmsh to show pool members connection stats for FQDN member?

Hi there,

I have a pool that contains 2 members and 2 auto-generated FQDN members; e.g.

Where _auto_104.1.1.1 is linked to and the other is linked to 02. 

If I want to see the connection stats for node - I notice the stats are only connected against _auto_104.1.1.1.

So how can I use TMSH to query Server01 connection stats?

I hope that makes sense!  Let me know if I'm misunderstanding any of this, as it would really help.

Thank you in advance.

4 Replies

  • gneville I don't believe you can gather data for because that is just the FQDN that is used to resolve a DNS query. In this particular case _auto_104.1.1.1 is the connection stats for so if you could see them they would be one in the same. You will always see connection stats down to the pool member that is auto generated based on the FQDN being used. So if resolved to 4 A records then you would see the stats based on each _auto_ so you would see individual stats rather than a grouped stat from the FQDN.

    • gneville's avatar
      Icon for Nimbostratus rankNimbostratus

      Thanks for the reply. 

      So there is no way to use TMSH (or any other command line / scripting approach) to query the name of a node in a pool, if it's been added as a 'FQDN node' - and get it's connection stats?

      This seems like it should be not a difficult thing to do;  because surely anyone is going to want to use the hostname of a node e.g. and not _auto_104.1.1.1 in any sort of scripting or automation.

  • Perhaps I should add - what I'm trying to do is script a 'drain stop' type thing; i.e. set a member to disabled (rather than force offline) and loop until the # of connections is 0.

    • Paulius's avatar
      Icon for MVP rankMVP

      gneville If your intent is to drain the connections your script would need to reference the node nested in the FQDN pool member. On top of this, an issue that you might run into is you cannot disable an auto pool member nested under an FQDN. Your best bet in this instance would be to create an HTTP health monitor, if this is HTTP traffic, and use the receive string down so that the auto pool member can pull itself out of rotation rather than manually disabling pool members with a script. I recommend that whenever possible these application level health monitors should be used to disable the pool member so that you don't have to change anything on the F5 side of things. Having this in place allows the server admins to not be dependent on the F5 team to disable pool members.

      Because of the shortcomings of FQDN pool members I never recommend using them because it causes more headaches than it solves for. I could be wrong on the part of disabling the auto pool members but last I checked you could not disable them without disabling the FQDN pool member.