Hello,
I have a question about "mixing" Postfix lookup table types :
For a transport lookup table i wish to set up an MySQL lookup table (as
primary lookup) and a standard .db lookup table (the text and .db file
are periodically generated by a shell script querying the database) as a
rescue table as secondary lookup for the case of having MySQL server
down or unreacheable.
in main.cf i set :
transport_maps = mysql:/etc/postfix/mysql_transport.cf,
hash:/etc/postfix/transport_rescue
cat transport_rescue :
...
...
...
[EMAIL PROTECTED] relay:[x.y.z.t]
I voluntarily stop the MYSQL database to test the rescue transport .db
file and send a mail to [EMAIL PROTECTED] . Despite this "rescue"
transport, the mail goes to deferred queue ("address resolver failure")
until i start the MySQL database again. I wonder if it's a
normal/logical behaviour of postfix lookup when associating lookup with
MySQL database and MySQL server is down or if there is a way to have a
rescue transport (i probably miss something...)
I know i can have several mysql servers for high availability lookups
but i wonder if this idea of having a primary mysql lookup and a
secondary rescue .db file, could work and how.
my postfix version : 2.3.8 (Debian Etch package)
postfix log after stopping MySQL Database :
Aug 12 16:58:25 smtp1 postfix/trivial-rewrite[23873]: warning:
transport_maps lookup failure
...
...
Aug 12 16:58:26 smtp1 postfix/qmgr[23863]: ECABF38EB:
to=<[EMAIL PROTECTED]>, relay=none, delay=0.14, delays=0.1/0.04/0/0,
dsn=4.3.0, status=deferred (address resolver failure)
Note : i also have set a relayhost to a default server (MX) for mail not
matching addresses in the transport table. With mysql server down a mail
to an outside domain is also placed in the deferred queue and so not
relayed to the mx server.
Thank you for your help / explanation
