Ok, I'm new to these lists, I had some previous threads open on
lists.postfix.users for the problems I am having, maybe I submitted to
the wrong list, I haven't gotten any answers, so I am trying this list.

The problem I am having is that during testing using smtp-source, mail
is not being delivered while it is incoming from smtp-source. As soon
smtp-source is done and all the messages have been queued, it finally
starts delivering the mail to our internal servers. If I kick of
another smtp-source while it is delivering this mail, it stops delivery
until all the mail has once again queued and it then start delivering
mail to our internal servers. It appears that mail delivery stops when
mail is incoming.

Here my postconf -n(changing host names and ips for anonymity):

alias_maps = dbm:/etc/postfix/aliases
command_directory = /usr/sbin
config_directory = /etc/postfix
daemon_directory = /usr/libexec/postfix
debug_peer_level = 2
default_destination_concurrency_limit = 100
default_process_limit = 3000
default_rbl_reply = $rbl_code Service unavailable; $rbl_class
[$rbl_what] blocked using $rbl_domain${rbl_reason?; $rbl_reason} - see
http://$rbl_domain.
html_directory = no
in_flow_delay = 0
inet_interfaces = all
initial_destination_concurrency = 40
local_destination_concurrency_limit = 40
mail_owner = postfix
mailq_path = /usr/bin/mailq
manpage_directory = /usr/local/man
mydestination = $myhostname, localhost.$mydomain, localhost, $mydomain
mydomain = domain.tld
myhostname = hostname.domain.tld
mynetworks = x.x.x.x/24
mynetworks_style = host
myorigin = $mydomain
newaliases_path = /usr/bin/newaliases
queue_directory = /var/spool/postfix
readme_directory = no
relay_domains = /etc/postfix/store_and_forward_domains, $transport_maps
sample_directory = /etc/postfix
sendmail_path = /usr/lib/sendmail
setgid_group = postdrop
smtp_destination_concurrency_limit = 100
smtpd_banner = $myhostname ESMTP
smtpd_client_connection_count_limit = 0
smtpd_client_connection_rate_limit = 0
smtpd_client_message_rate_limit = 0
smtpd_client_recipient_rate_limit = 0
smtpd_client_restrictions = dbm:/etc/postfix/access,
reject_rbl_client rbl-plus.mail-abuse.org, reject_rbl_client
sbl.spamhaus.org, reject_rbl_client mail-security.rr.com
smtpd_data_restrictions = reject_unauth_pipelining, permit
smtpd_error_sleep_time = 0
smtpd_hard_error_limit = 10000
smtpd_helo_required = yes
smtpd_helo_restrictions = check_helo_access dbm:/etc/postfix/access,
reject_unknown_hostname, reject_invalid_hostname,
reject_non_fqdn_hostname, permit_mynetworks
smtpd_recipient_restrictions = reject_invalid_hostname,
reject_unknown_client, reject_non_fqdn_sender,
reject_non_fqdn_recipient, reject_unauth_pipelining,
reject_unknown_sender_domain, reject_unknown_recipient_domain,
permit_mynetworks, reject_unauth_destination,
reject_rbl_client rbl-plus.mail-abuse.org, reject_rbl_client
sbl.spamhaus.org, reject_rbl_client mail-security.rr.com,
check_recipient_access dbm:/etc/postfix/access,
check_sender_access dbm:/etc/postfix/access, check_client_access
dbm:/etc/postfix/access, reject_unauth_destination
smtpd_soft_error_limit = 10000
strict_rfc821_envelopes = yes
transport_maps = dbm:/etc/postfix/mailertable
unknown_address_reject_code = 550
unknown_client_reject_code = 550
unknown_hostname_reject_code = 550
unknown_local_recipient_reject_code = 550

Please tell me what I am doing wrong

jlew wrote:
> maybe I submitted to
> the wrong list, I haven't gotten any answers, so I am trying this list.

You probably got no answers because there is no definitive answer to
your question, as it would probably require hands-on system analysis.


> The problem I am having is that during testing using smtp-source, mail
> is not being delivered while it is incoming from smtp-source. As soon
> smtp-source is done and all the messages have been queued, it finally
> starts delivering the mail to our internal servers.

You didn't say how you were using smtp-source. If you are using it to
send thousands of external messages using parallel SMTP delivery, then yeah,
that's gonna have an effect on the delivery of your normal email relaying
to internal servers.

There's a number of system or postfix resource limits that could be reached. If so there
should probably be errors in the system or Postfix syslog files.

Also, if there are lots of external messages in queue, any internal messages
might have to wait a while before the queue manager gets around to delivering them.
When the problem happens, you could try running "tail -f /var/log/maillog" to watch
what's going on at the time. And, the "mailq" command should show you what (if anything)
is sitting in the "active" queue trying to be delivered at the time.

--
Greg

Greg:

Perhaps I'm wrong in my assumptions of the way smtp-source works, my
previous experience with postfix was with far simpler mail system.
With smpt-source, I was trying to determine at what point the system
would max out and get an approximation on where I should set
performance based tunables. Here's an example of how I was running
smtp-source:

time smtp-source -m 100000 -s 1000 -t "<user@domain.tld>"
hostname.domain.tld

I had assumed that as mail was recieved by the server it would deliver
the mail as it came in but that is not what has been happenning. Once
smtp-source generated all the mail and exited, it would then start
delivering the mail. And once it has started delivering the mail, it
still seems as though it's not pushing the mail as fast as it should
be.

Here's a little backgroup on our mail system, the servers I am speaking
of right now are inbound servers for our system. We currently host
mail for thousands of domains and our current inbound servers are at
capacity. Our current mail system consists of 6 inbound sendmail
servers which took in about 147,000,000 messages for the month of
April, delivering only 23,400,000 of those messages due to spam. I am
trying to expand capacity by adding 2 additional servers using postfix,
possibly migrating the other sendmail servers to postfix if there is a
noticeable difference between the platforms.

We are using our inbound servers as antispam gateways, once mail is
taken into our inbound servers it gets transferred to our internal
mta's and then on to our messages stores.

Please let me know if you require additional information.

Thanks,
-Jonathan

Sorry, I forgot to mention, I have been going through the logs, but I
haven't seen any errors or problems.

-Jonathan

jlew wrote:

> I had assumed that as mail was recieved by the server it would deliver
> the mail as it came in but that is not what has been happenning.

On my system, the smtp-source generated email starts getting delivered
right away, before smtp-source finishes.

I don't know much about the Postfix queue manager internals. But I
know the qmgr code was re-written in one of the 2.1 versions, and
there's been some changes since then. Are you running an older version
of Postfix?

Just out of curiosity: If "nobody" points to /dev/null in aliases,
and you send a batch of smtps-source email to nobody, does the email start delivering
right away?

> Our current mail system consists of 6 inbound sendmail
> servers which took in about 147,000,000 messages for the month of
> April, delivering only 23,400,000 of those messages due to spam.

Sounds a whole lot like where I last worked. We had 6 inbound, and
9 outbound Postfix servers. When first installed, 40% was spam. Nowdays
85% spam sounds about right.

--
Greg