sqlippool and exhausted pool

[Mirko Alberio]

Ok, sorry, but some thing that I assumed were right turned out wrong and 
many doubts pop up,

so ok we found out that the NAS (Mikrotik) did not have the 
interim-update enabled, and so we are working to enable it.

What would be a suggested time interval?

Thanks!

Mirko Alberio - Assistenza tecnica
e-mail: mirko.alberio at telemar.it

Telemar SpA Internet Quality Provider
Via Enrico Fermi, 235 - 36100 Vicenza - Italia
Tel 0444 291302 - Fax 0444 566310 - www.telemar.it
Assistenza tecnica 0444 1420000
Reg. Imp. Di Vicenza /C.F./P.I. 02508710247
Cap. Soc. € 120.000,00 I.V.
R.E.A. VI-236292

Il 19/07/2021 17:13, Alan DeKok ha scritto:
> On Jul 19, 2021, at 10:56 AM, Mirko Alberio <mirko.alberio at telemar.it> wrote:
>> Ok, thanks, I did some packet capture in the problematic NAS, accounting Stop packet are actually being sent, but still the table is not updated. I attach the capture. So the NAS is sending account data!
>    Please read my messages.  The Stop packet isn't what is important.  The interim update messages are important.
>
>    It really helps to pay attention, and to understand what's going on.  If you don't, then you'll be randomly changing irrelevant things, and nothing will get fixes.
>
>> And another quick question: we noticed also some cases where username deleted from the radcheck and radreply tables (dismissed customers) are still present in the radippool table, with past expiry_time: should not be automatically "pruned"?
>    No.  That's not how RADIUS works.
>
>    There is nothing which monitors the SQL database for changes, and kicks off users.  Perhaps someone could write a script and submit it back to us...
>
>> Those customers for example could disconnect the router cable and the NAS is not able to send the Stop packet.
>    That's not how it works.
>
>    If the session is stopped for ANY reason, then the NAS sends a stop packet.  If the NAS doesn't do that, then throw it in the garbage, and buy a NAS which follows the RADIUS specifications.
>
>    Alan DeKok.
>