Start a Conversation

Unsolved

This post is more than 5 years old

2178

August 8th, 2013 08:00

Using "sendmail" from Avamar

We use the replemail.sh script to report on replication status each morning.  It works fine on several of our grids.  I copied it to a newer single-node system.  The script completes successfully and I can see the replstatus.html file get created but the email is not being received from this system.  Avamar notification alert emails sent from this system do work and are being received.

I've been able to emulate sending a message via sendmail with the following code snippet (modified from replemail.sh):

(cat <

From: avamargrid@domain.com

Subject: Test Email   

To: destination@domain.com

MIME-Version: 1.0

Content-Type: text/html

Content-Disposition: inline

EOCAT

cat $html) | sendmail -f avamargrid@domain.com destination@domain.com

I can run this bit of code (after inserting the proper email addresses) from the Linux command line on our other systems and I receive a blank email (good result).  When running it from the system from which the replication report is not being received I also do not receive a test email.

Is there any configuration that is needed prior to using sendmail on an Avamar system?  I'm not a Linux expert so this very well could be something simple.  Perhaps it isn't using the correct SMTP relay?  I tried doing some Googling regarding sendmail but got lost in the differences between different distributions and how configuration is managed (definitely showing my lack of Linux knowledge here).

Thanks in advance,

Joe

143 Posts

August 8th, 2013 15:00

Not sure if this helps to clarify what the problem may be or not but looking through the Postfix logs, it appears Postfix is attempting to connect to our external SMTP relay (which the Avamar server does not have access to) via the MX records configured for our external domain to deliver the message instead of our internal SMTP bridge server.  The question is how is this behavior configured?  I've compared the /etc/postfix/main.cf files of a server that works with one that does not and both are identical.  So that doesn't seem to be the relevant detail.  I also noticed our internal SMTP bridge server for this site didn't have a reverse DNS record so I added that to see if it might make a difference.  No dice there, either.  I also tried passing "AuthInfo: bridge.localdomain.suffix" with the sample test email code above.  It still seems postfix wants to use our external SMTP servers.  Hmm...

Edit: if I add "relayhost = bridge.localdomain.suffix" to /etc/postfix/main.cf everything works fine.  I'm still a bit curious why this works on the other grids without explicitly specifying the SMTP relay server name.

No Events found!

Top