On 5/14/19 12:25 AM, johnr wrote: > If in the context of this directive became idle means "done processing the > previous request" then how is the pconn_lifetime directive different than > the client_idle_pconn_timeout and server_idle_pconn_timeout (other than > affecting both at the same time)? The former directive limits the lifetime of a connection. The latter directives limit the time a connection can stay in the idle pool. In other words, the clock for the first directive starts ticking when the connection is created while the clock for the other two directives starts ticking when the connection is placed in the idle connection pool. > If my question doesn't make sense, then > perhaps what I'm asking is better stated as: how do the pconn_lifetime and > client_idle_pconn_timeout interact? There should be virtually no interaction: The former limit is checked just when a connection becomes idle and Squid decides whether to pool the connection or close it. The latter timeout is checked for the already pooled connections. GitHub pull requests clarifying documentation are welcomed. Alex. _______________________________________________ squid-users mailing list squid-users@xxxxxxxxxxxxxxxxxxxxx http://lists.squid-cache.org/listinfo/squid-users
