Due to hardware problems, I had to move our "graveyard mailout" to
another server. So I copied the relevant configuration files and then
moved the mailqueue to the new server using rsync. After that, I
executed
postmulti -i postfix-gy -x postsuper -s
I expected a change of queue IDs, but this didn't happen, not even
for a single mail. The filesystem is a standard ext3, both servers are
running CentOS 5.5 x86_64 and Postfix 2.7.1. The output of
postmulti -i postfix-gy -x postconf -n
is:
alias_database =
alias_maps =
authorized_submit_users = root
config_directory = /etc/postfix-gy
data_directory = /var/lib/postfix-gy
local_header_rewrite_clients =
local_recipient_maps =
local_transport = error:5.1.1 Mailbox unavailable
master_service_disable =
multi_instance_enable = yes
multi_instance_group = mta
multi_instance_name = postfix-gy
mydestination =
mydomain = example.com
myhostname = gy.mail.example.com
mynetworks = 127.0.0.0/8, 192.168.12.0/24
myorigin = $mydomain
parent_domain_matches_subdomains =
queue_directory = /var/spool/postfix-gy
receive_override_options = no_unknown_recipient_checks
recipient_delimiter = +
relay_destination_concurrency_limit = 8
smtp_helo_name = mail.example.com
smtp_tls_CAfile = /etc/ssl/certs/ca-certificates.crt
smtp_tls_security_level = may
smtp_tls_session_cache_database = btree:${data_directory}/smtp_scache
smtp_tls_session_cache_timeout = 24h
smtpd_client_connection_count_limit = 0
smtpd_recipient_restrictions = permit_mynetworks, reject
Questions:
1. Did I execute "postsuper -s" the right way?
2. I was verifiyng queue IDs remained unchanged by comparing the
output of
postmulti -i postfix-gy -x postqueue -p
befor and after the call to "postsuper -s". Is that methodology
flawed?
3. I verified that the "postfix-gy" instance is working properly by
destroying it, recreating it and feeding it some mails via SMTP. I can
easily write a few lines of Perl to feed the backup of the old mail
queue to the new instance that way - would that be a viable
workaround?
Thanks
Stefan
