25 Nov
2011
25 Nov
'11
8:58 p.m.
I'm sure this is a newbie-level question, but I can't find it clearly
answered in the docs or example configurations.
Assuming a single instance of Perdition is bound to multiple IP addresses,
can Perdition rewrite requests based solely on what IP address receives a
login request?
I'm replacing an old mail system where each domain has a separate IP
address, and there are multiple instances where folks have the same
username and password on different domains. Their mail clients are often
configure with a user name of 'bob' as opposed to 'bob(a)domain.com'com', so I
really need the IP address in addition to the username/password to figure
out where a connection should go.
I know I could get around this by having multiple instances of Perdition,
each bound to a separate IP address, but there are about 40 domains (and 40
IPs) involved in this process. If there's an easier way to do it, I'd love
to avoid having all those nearly-identical configuration files floating
around, waiting to confuse me.
David Smith
MVN.net
26 Nov
26 Nov
5:17 a.m.
On Fri, Nov 25, 2011 at 02:58:56PM -0600, David E. Smith wrote:
I'm sure this is a newbie-level question, but I
can't find it clearly
answered in the docs or example configurations.
Assuming a single instance of Perdition is bound to multiple IP addresses,
can Perdition rewrite requests based solely on what IP address receives a
login request?
I'm replacing an old mail system where each domain has a separate IP
address, and there are multiple instances where folks have the same
username and password on different domains. Their mail clients are often
configure with a user name of 'bob' as opposed to 'bob(a)domain.com'com', so I
really need the IP address in addition to the username/password to figure
out where a connection should go.
I know I could get around this by having multiple instances of Perdition,
each bound to a separate IP address, but there are about 40 domains (and 40
IPs) involved in this process. If there's an easier way to do it, I'd love
to avoid having all those nearly-identical configuration files floating
around, waiting to confuse me.
Hi David,
I believe that the query-key option is what you are after.
It should allow you to use the IP address that the client connect to
as one of the keys for the popmap lookup.
28 Nov
28 Nov
3:13 p.m.
On Fri, Nov 25, 2011 at 23:17, Simon Horman <horms(a)verge.net.au> wrote:
I believe that the query-key option is what you are
after.
It should allow you to use the IP address that the client connect to
as one of the keys for the popmap lookup.
I did, eventually, see that, and felt silly afterwards for overlooking it. :)
If a user supplies a full email address (user@domain), I want to look
up the domain, and send their POP3 session to a specified server,
regardless of what IP they connected to. Failing that, I'll look up
the IP they connected to, and route their POP3 based on that.
For this, it looks like query_key should be \d,\I . Then, I assume
popmap looks like this:
domain1.com:10.1.1.1
domain2.com:10.1.1.2
192.168.1.1:10.1.1.1
192.168.1.2:10.1.1.2
(Fake IPs all around - the 10. IPs are the "real" back-end mail
servers, pretend the 192.168 IPs are public.)
I'm just having a bit of trouble wrapping my head round the fact that
there's only one popmap table, and I'll want it to do several things.
Thanks for the help!
David Smith
MVN.net
1 Dec
1 Dec
9:36 p.m.
On Mon, Nov 28, 2011 at 09:13, David E. Smith <dave(a)mvn.net> wrote:
domain1.com:10.1.1.1
domain2.com:10.1.1.2
192.168.1.1:10.1.1.1
192.168.1.2:10.1.1.2
I don't think I'm getting what I want out of my popmap with the above - I
need to be able to change the username/domain based on the key. Here's what
I'm trying to do:
* If a user sends a fully-qualified email address as their POP3 login name,
look up the domain provided in that login name, and route based upon that.
(The first entries in the above table seem to do that.)
* If a user sends only a username as the POP3 login name, look up the IP
they connected to, route their connection based on that, and rewrite the
username.
I don't understand the correct syntax to use in the popmap to handle the
domain rewriting part. Should it be something like
192.168.1.2:10.1.1.2@domain.com
or some other, more odd syntax?
David Smith
MVN.net
2 Dec
2 Dec
6:50 a.m.
On Thu, Dec 01, 2011 at 03:36:08PM -0600, David E. Smith wrote:
On Mon, Nov 28, 2011 at 09:13, David E. Smith
<dave(a)mvn.net> wrote:
Hi David,
unfortunately perdition only supports one lookup.
I believe that the code would need to be (slightly)
modified to allow the or logic you are after.
domain1.com:10.1.1.1
domain2.com:10.1.1.2
192.168.1.1:10.1.1.1
192.168.1.2:10.1.1.2
I don't think I'm getting what I want out of my popmap with the above - I
need to be able to change the username/domain based on the key. Here's what
I'm trying to do:
* If a user sends a fully-qualified email address as their POP3 login name,
look up the domain provided in that login name, and route based upon that.
(The first entries in the above table seem to do that.)
* If a user sends only a username as the POP3 login name, look up the IP
they connected to, route their connection based on that, and rewrite the
username.
I don't understand the correct syntax to use in the popmap to handle the
domain rewriting part. Should it be something like
192.168.1.2:10.1.1.2@domain.com
or some other, more odd syntax?
6:19 p.m.
On Fri, Dec 2, 2011 at 00:50, Simon Horman <horms(a)verge.net.au> wrote:
On Thu, Dec 01, 2011 at 03:36:08PM -0600, David E.
Smith wrote:
Actually, I don't think I'll need it. I didn't realize the
"add_domain"
option existed. I can just put the intended destination domains in DNS, and
probably even simplify the popmap that way (since, post-migration, almost
everything will be going to the same destination server anyway).
Looks like you are once again way ahead of me. :)
David Smith
MVN.net
On Mon, Nov 28, 2011 at 09:13, David E. Smith
<dave(a)mvn.net> wrote:
unfortunately perdition only supports one lookup.
I believe that the code would need to be (slightly)
modified to allow the or logic you are after.
4553
days inactive
4560
days old
5 comments
2 participants
participants (2)
-
David E. Smith -
Simon Horman