max_requests variable in radiusd.conf
Hello, I'm a new in freeradius and i have a question, what controls are made by the variable max_requests in radiusd.conf? all the established sessions of the various NASs are directly managed by the freeradius daemon or does this control reside in the database? I ask because if the session control is made entirely in the daemon without intervention of the database and it is necessary to work with load balance, the connections that arrive in this 2 freeradius servers must be tied "sticky" to the source IP, if you balance all connections without "sticky" feature this control would be broken , i'm correct? Thanks, Claudio
On Nov 9, 2017, at 6:43 AM, Claudio Carvalho <claudio@nextdigitalservices.com.br> wrote:
I'm a new in freeradius and i have a question, what controls are made by the variable max_requests in radiusd.conf?
The comments in the radiusd.conf file document what out controls: the number of packets *currently being processed* by the server.
all the established sessions of the various NASs are directly managed by the freeradius daemon or does this control reside in the database?
It's all in a DB.
I ask because if the session control is made entirely in the daemon without intervention of the database and it is necessary to work with load balance, the connections that arrive in this 2 freeradius servers must be tied "sticky" to the source IP, if you balance all connections without "sticky" feature this control would be broken , i'm correct?
No. If accounting packets go to two destinations (RADIUS servers or DBs), then those records have to be reconciled. That's it. Alan DeKok.
Hi Alan, Using the phrase "The maximum number of requests which the server keeps track of", my first understanding is that somehow the freeradius daemon is doing let's say in memory a control of the opening and closing of sessions "track of", you are saying this does not happen, everything is done through the information of the database enabling the use of multiple servers, right? "The maximum number of requests which the server keeps track of. This should be 256 multiplied by the number of clients. e.g. With 4 clients, this number should be 1024" Thanks, Claudio. On Thu, Nov 09, 2017 at 11:07am, Alan DeKok <aland@deployingradius.com <mailto:aland@deployingradius.com> > wrote: On Nov 9, 2017, at 6:43 AM, Claudio Carvalho <claudio@nextdigitalservices.com.br> wrote:
I'm a new in freeradius and i have a question, what controls are made by the variable max_requests in radiusd.conf?
The comments in the radiusd.conf file document what out controls: the number of packets *currently being processed* by the server.
all the established sessions of the various NASs are directly managed by the freeradius daemon or does this control reside in the database?
It's all in a DB.
I ask because if the session control is made entirely in the daemon without intervention of the database and it is necessary to work with load balance, the connections that arrive in this 2 freeradius servers must be tied "sticky" to the source IP, if you balance all connections without "sticky" feature this control would be broken , i'm correct?
No. If accounting packets go to two destinations (RADIUS servers or DBs), then those records have to be reconciled. That's it. Alan DeKok. - List info/subscribe/unsubscribe? See http://www.freeradius.org/list/users.html
On Nov 9, 2017, at 8:59 AM, Claudio Carvalho <claudio@nextdigitalservices.com.br> wrote:
Using the phrase "The maximum number of requests which the server keeps track of", my first understanding is that somehow the freeradius daemon is doing let's say in memory a control of the opening and closing of sessions
No. If I had meant too use the word "sessions", I would have used the word "sessions". The server receives *packets*. The "max_requests" refers to *packets*. It does not refer to *sessions*. I tried to make this clear in my previous message, when I said sessions go into the DB. Apparently not enough... Stop trying to force FreeRADIUS to fit your ideas. Your ideas about how it works are wrong. The documentation describes how FreeRADIUS works.
"track of", you are saying this does not happen, everything is done through the information of the database enabling the use of multiple servers, right?
I have no idea what that means. FreeRADIUS stores accounting data (i.e. user sessions) in a database. You can configure two RADIUS servers to talk to the same database. This isn't complex. There's no magic involved.
"The maximum number of requests which the server keeps track of. This should be 256 multiplied by the number of clients. e.g. With 4 clients, this number should be 1024"
Is there any reason you're quoting the config files? Do you think I haven't seen that before? Alan DeKok.
participants (2)
-
Alan DeKok -
Claudio Carvalho