Return-Path: Sender: To: CGatePro Date: Thu, 02 Sep 2010 07:30:05 -0700 Message-ID: X-Original-Return-Path: Received: from mail-qy0-f174.google.com ([209.85.216.174] verified) by mail.stalker.com (CommuniGate Pro SMTP 5.3.9) with ESMTP id 61673157 for CGatePro@mail.stalker.com; Wed, 01 Sep 2010 09:15:41 -0700 Received-SPF: pass receiver=mail.stalker.com; client-ip=209.85.216.174; envelope-from=wsanders@gmail.com Received: by qyk36 with SMTP id 36so827721qyk.12 for ; Wed, 01 Sep 2010 09:15:27 -0700 (PDT) DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:sender:date:x-google-sender-auth:message-id:subject :from:to:content-type; b=iPTdvC5YBOXyOkgK2GqBASN5W7BUf2nJ7Z9Kdur6pRlxDH2RdZfq2oB5gJ7vajKZqb inMDt0XlonmhpP1J7dTCXTEJwsJRbs22uzctgzX0WwPBDwOpBDhZOo9lEu+aFizf1vOI gP/KoZAdhc1DZe2+NwMLDpEn9nLAnZRpRlfJk= MIME-Version: 1.0 Received: by 10.224.89.81 with SMTP id d17mr5392251qam.156.1283357726911; Wed, 01 Sep 2010 09:15:26 -0700 (PDT) X-Original-Sender: wsanders@gmail.com Received: by 10.229.194.199 with HTTP; Wed, 1 Sep 2010 09:15:26 -0700 (PDT) X-Original-Date: Wed, 1 Sep 2010 09:15:26 -0700 X-Google-Sender-Auth: tJyUA6CwuUdG6FlnZAn0PR_B0oA X-Original-Message-ID: Subject: Re: IMAP session time limit From: Wiley Sanders X-Original-To: CommuniGate Pro Discussions Content-Type: text/plain; charset=ISO-8859-1 Linux now defaults to being able to support something like 2.4 million open sockets, so I've decided this is a good thing, since the only resource consumed is an open socket. Our CGPro host is licensed for 2000 users and we actually have more than 2000 IMAP clients at times due to users with mobile devices, etc. The devices that seem to keep sockets open for weeks at a time are almost all Blackberrys.This is necessary to facilitate mail "push". You can limit the number of cached connections, but, below 5, users will be baffled by transient error from Thunderbird and other clients that try to connect with 5 sockets or more by default. One possible resource limit feature request would be to limit the frequency of polling. Some of my users poll multiple times per minute (I don't know how - the lower limit in Thunderbird is 1 min.) But that might not have any effect, since I have no data that a poll induces much system activity, if any.