Sendmail stops sending outbound e-mail

[ScalixSupport]

I just want to verify that you did restart sendmail, right? If you didn't, the local-host-names file hasn't been read yet. When you say the problem still persists, do you mean that when you come in in the morning and do a "ps -aef|grep sendmail" you see a whole slew of connections to ryans.com? If you have a huge number of messages queued up for sendmail to reject, it will take a while for it to clear those, especially if a lot of them are spam and are going back to non-existent domains.

Regarding the POP problem, what version of Scalix are you using? What does it say in the fatal log regarding the POP server?

Thanks,
Rachel

[RyHD]

sorry, that was a typo. What i meant to say was, I added the ryans.com entry to /etc/mail/local-host-names and see if the problem still persists....

As far as the pop3 interface, nothing is referenced in the fatal log for the pop server.

on the client side, it will pull down between 25-50 e-mails, timeout, and then not delete the recieved e-mails off of the server. In essence this is causing some users to not be able to recieve all of their e-mail. Whe they close their mail clients and try to download the rest of their messages, it just redownloads the first batch and hangs again....

[RyHD]

Suse Linux Enterprise 9.3, Scalix 9.4.2 (10.0 gave us ical issues, which some of our users use, so we had to roll back to 9.4.2, which involved reloading the server and recreating ALL of the mail accounts.)

[ScalixSupport]

Hi. That's actually the nature of POP. If it can't verify that message were successfully downloaded, it won't delete them.

As far as why it's hanging during the download, have you tried it from another PC or at least a different POP client? We need to isolate it down to either something on the client or something in the user's mailbox. If it's latter, can the user logon with IMAP and see all of his/her messages? If POP stops at the same place everytime, he/she should be able to see what the next message is via IMAP. Is there anything odd about that message?

Thanks,
Rachel

[RyHD]

adding ryans.com entry to /etc/mail/local-host-names caused all e-mails outgoing to have the fqdn applied to the address, i.e. instead of joeblow@ryans.com, they were going out as joeblow@mail.ryans.com.

[ScalixSupport]

There's a couple of ways to handle that. The easiest is to change the first email address for your users to name@ryans.com rather than name@mail.ryans.com. The second method is to have sendmail rewrite your outbound addresses from mail.ryans.com to ryans.com. In order to do that, edit your sendmail.mc file and add:

Code: Select all

MASQUERADE_AS(`ryans.com')dnl
MASQUERADE_DOMAIN(mail.ryans.com)dnl

then rebuild your sendmail.cf file. Lastly, run omsendin to re-apply the Scalix specific sendmail.cf entries.

Thanks,
Rachel

[RyHD]

there does not seem to be a sendmail.mc file anywhere on the system. Also, all the users currently have a @ryans.com address, and not the latter.....



Thanks,
RyHD

[ScalixSupport]

I went back and looked at your sendmail.cf and you're already masquerading:

Code: Select all

# who I masquerade as (null for no masquerading) (see also $=M)
DMmail.ryans.com


You should change that to be ryans.com. In other words:

Code: Select all

# who I masquerade as (null for no masquerading) (see also $=M)
DMryans.com


Then restart sendmail.

Thanks,
Rachel

[RyHD]

This morning we encountered the following, again....

00:00:14 sendmail: rejecting connections on daemon Daemon0: 15 children, max 15

what if we upped the value of MaxDaemonChildren in sendmail.cf?


Thanks,
RyHD

[ScalixSupport]

You could certainly increase that value, but it may just temporarily mask the problem. Did you get the local-host-names and marquerarding issue sorted out?

Thanks,
Rachel

[RyHD]

After making the recommended changes, we were still having the issue. I cleaned out the /var/spool/mqueue/ folder and adjusted the spam tolerance threshold on our spam firewall, and so far so good. Apparently, the firewall was set to be less restrictive, and was allowing alot through, that would hit inside, and bounced around indefinitely, tying up a childdaemon in sendmail. We will know by friday if this theory is correct.


Thanks All,
RyHD

[RyHD]

Everything seems to be flowing great!! Thanks All!