Re: RemoteDelivery Gateway not reachable - Only 2 hours retry period - How to increase?

Hello Dominik,

for a "dumb" internal-to-external relay I use sendmail with its 
nullclient config.

My setup looks like this:

OS: opensuse 15
local sendmail: sendmail
sendmail config: nullclient(localhost)
proper mail-server: James

In my config I use it to provide local sendmail for logs from cron or 
sending mails via PHP mail() via apache: They use the default sendmail 
binary and sendmail is set up to just forward the mails to be handled by 
James. I have a bit of code in the mailet config so such mails get 
sorted: Local status mails like from cron are redirected to a local 
inbox, anything from apache is send out to its target.

Using James just as a dumb relay seems a bit overkill. I'm sure you have 
your reason for such relay for outgoing mail - but using James for it?

Maybe you can expand a bit on your actual situations what it actually is 
you want to achieve and describe your setup a bit more.

Also: As I used the spring setup myself for a long time - it's subject 
to get removed with James 4. Setting up the preferred guice build is 
just as easy. Give it a shot.

Also also: When a mail can't get delivered it's not lost but gets send 
to the external storage in the <james>/var sub-folder. Look for your 
mails there.


Greetings

Matt

On 06.03.24 21:41, Dominik Katzenmaier wrote:
> Hello,
>
> I am testing Apache James since a few days. I am using the current 
> 3.8.1 Spring-Edition.
> My goal is to use Apache James as an SMTP relay for some of our 
> internal Applications. We don’t need any Users, Domains or Mailboxes 
> on our James, just a simple relay to our external SMTP-Provider. I’ve 
> already successfully configured the RemoteDelivery mailet in the 
> mailetcontainer.xml. James accepts the Emails via SMTP and forwards 
> them to the RemoteDelivery-Gateway.
>
> After this success, I changed the Gateway-Port in my 
> RemoteDelivery-Configuration to a wrong value, to test what happens if 
> the remote gateway is not reachable (to simulate a internet connection 
> failure). I sent an E-Mail to the James-SMTP and waited some hours to 
> see what James did with the E-Mail.
>
> I can see in the log/wrapper.xml that James tried 1x per Minute to 
> send the E-Mail to the RemoteDelivery-Gateway. BUT James stopped the 
> retries after almost 2 hours (so almost 120 tries). James (or ActiveMQ 
> itself??) deleted the E-Mail from the ActiveMQ-Mailqueue too, the Mail 
> is lost.
>
> This is a unacceptable behavior. We need at least retries for 15 
> Hours, more better several days. Where can I configure this?
> I saw in the RemoteDelivery-Section of mailetcontainer.xml the options 
> <delayTime> and <maxRetries>. I didn’t changed the default values, so 
> they are:
>
> <!--  Delivery Schedule based upon RFC 2821, 4.5.4.1  -->
> <!--  5 day retry period, with 4 attempts in the first
>                  hour, two more within the first 6 hours, and then
>                  every 6 hours for the rest of the period.
>           <delayTime>4 * 15 minutes, 2 * 3 hours, 18 * 6 
> hours</delayTime>
>            -->
> <delayTime>5000, 100000, 500000</delayTime>
> <maxRetries>3</maxRetries>
>
> I can't really understand the comment above.
> “5 day retry period, with with 4 attempts in the first hour, two more 
> within the first 6 hours, and then every 6 hours for the rest of the 
> period.” ???
> Sounds Good! I would take that :D . But in my test I see only a 2 
> hours retry period, with 1 try per minute.
>
> Does anyone have an idea how I can increase the retry period?
>
> Thank you for this great open source project and your support!
>
> Best Regards
> Dominik Katzenmaier
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>