On Thu, 20 Apr 2000 at 18:45, Herbert Melgar wrote:
>1. The ISP we plan to use says that they will queue mail to and from us
>on their servers using our reserved domain. But what about our internal
>e-mail?
I'm presuming this will be using a catch-all system. Meaning
<[EMAIL PROTECTED]> will be retrieved from one POP3 mailbox on your
ISP. I'm using a setup like this. I set up Fetchmail to periodically
download the mail. It parses the headers for the latest "Received:"
envelope and uses this to find out for which user the mail is for. Benefit
is, you don't have to be connected to the 'Net 24x7 to get e-mail with
your domain. Problem is, this doesn't always work properly, depending on
how your ISP's POP3 server separates messages sent to multiple people on
the same domain. If your ISP will separate them into multiple messages
with distinct Delivered-To or Received headers then you're safe. If it
doesn't, like mine, you'll have to pick whether you want Fetchmail to
check the To/Cc headers or the Received headers. To/Cc checking will mean
you can't get messages from mailing lists properly parsed. Received header
checking will mean a single message to more than one recipient on your
domain will not get properly parsed.
So that should cover incoming e-mail part 1 (from your server to you).
Summary: Fetchmail <http://www.tuxedo.org/~esr/fetchmail/>
>2. Do we need to set up our own internal e-mail server that will
>periodically queue outgoing mail and pull incoming mail and route it to
>the appropriate person? What software do we need to use to perform these
>tasks?
Yes, you will need an internal MTA. I use QMail. I simply define what
domains are local in /var/qmail/control/locals so that a message from
someone on the LAN to <[EMAIL PROTECTED]> doesn't get out
anymore. It stays inside. :-)
I don't queue outgoing mail though. I have QMail send out messages right
away. I have a virtual dedicated link to the Internet via a SLIP interface
of DialD. When any packet needs to go out (like mail), DialD kicks in my
chat script, then pppd, then depending on my set of rules will disconnect
after a given amount of inactivity (after e-mail is sent it disconnects
immediately, when the outgoing packet was HTTP then it gives the link 5
minutes idle time, etc). :-)
>3. Will QMail be okay or do we use sendmail (arghhh) as our mail server?
>Which e-mail server is the fastest/easiest to set up?
I don't know about fastest and easiest, but I'm using QMail. I'd go QMail
or Postfix over Sendmail anytime. :)
>5. We need IP masquerading to allow Web access, right?
Yup! Well ... either that or you set up a proxy server (Squid?). :)
>6. To speed up Web access I'm thinking of using Squid. Will it work
>properly with IP masquerading? How?
With Squid you don't even need IP Masquerading. IP Masquerading will be
useful if you want things that aren't proxied. You'll set up your Linux
box as a gateway. :)
>8. Do we need to set up our own DNS server? What type (caching, etc.)?
I would. Just so your Linux box (I call mine linux-server) is known to the
rest of the clients. :)
>What questions will I need to ask from the ISP?
You might want to ask them about the catch-all system thingie. :-)
Good luck! :)
-+[ Jijo Sevilla ]+-
[EMAIL PROTECTED]
-
Philippine Linux Users Group. Web site and archives at http://plug.linux.org.ph
To leave: send "unsubscribe" in the body to [EMAIL PROTECTED]
