Re: Er.... for Sendmail users......enough

On Tue, 29 Aug 2006, Matthew Schumacher wrote:

> Adam Bultman wrote:
>>
>> Re: schu lending a hand: I may cash in on that. Cyrus, if it's not nfs
>> safe, is a real bummer - since I need NFS-safeness.
>
> I honestly think you are better off to get rid of nfs. Most mail
> systems that are NFS safe trade off a lot of performance for it.
> Performance comes in small indexed binary files which generally don't
> work with NFS. Can't you setup your netapp to serve an iscsi target?

Eek. how do I handle multiple mail servers? Mounting my mail via a
netapp isn't a bad idea - a lot of people do it. You need a bit more CPU
horsepower to handle it, but I know of at least a few people who do it.
AFAIK, it's the only graceful way to handle mail clusters of indeterminate
size.

>>
>> One thing I DO need is a milter to let me block invalid users. We use
>> postini to filter *some* users mail, but postini happily forwards any mail
>> on to our servers - which we then have to bounce.
>
> Where are you users stored? LDAP?
>

>>
>> Sadly, this blows greylisting out of the water (since the only mail we
>> accept is via postini, and postini won't give up after one try) and I
>> might need a hand inserting the milter code. But I have to get it built
>> first (I have to build from source, and solaris is missing a few key
>> packages required. I'm currently trying to build glibc.)
>
> Why solaris? BTW, you don't need glibc to use sendmail.

Solaris is what is used here. It's not my choice. And I know sendmail
doesn't require glibc - after all, it's running as we speak ;). However,
in order to build the milter, I need glibc - which doesn't exist. Im'
still working on it.

> > schu
>
---------
To unsubscribe, send email to <aklug-request@aklug.org>
with 'unsubscribe' in the message body.
Received on Tue Aug 29 09:43:14 2006