smtpclient.send Taking way too long

Z

zero

Perch
I have quite a few sites that send out email from the web server. We have experienced large delays on page loads for pages that send mail.

The strange part of it is that once the page sends an email, if you immediately request another send (ex. from our mail test page), it executes immediately. The first time (in the span of 5 minutes or so) it attempts to send the email, it takes the smptclient.send method invocation at least 22 seconds to send the email, per our detailed logging.

Several of our ecommerce sites are experiencing this issue. When a user submits an order, the site sends out emails (one to customer, one to admin). Due to the delay, customers are sometimes clicking the submit button again. Granted, there are other ways around this issue, but I'm not here to debate the developer's code design. I need to find out why it's taking over 22 seconds to send an email.

Server: hwwin1
Mail Server: mail9

The emails send fine, so I don't think it's an issue on our end.

Ticket submitted, but no help from support so far. RS #CKT-86064-408


thanks,
-brit
 
The can happen if using port 25 due to rDNS lookup, try port 587.

I may have to enable you access to connect on alternative ports, if you get an error let me know.
 
Thanks for the reply, Stephen, but we have been using 587 for a few years now. ;-)

No errors so far, just the delay. The ticket has all of the code we're using for this and the link to the mail test page if it helps.


-brit
 
zero said:
Thanks for the reply, Stephen, but we have been using 587 for a few years now. ;-)

No errors so far, just the delay. The ticket has all of the code we're using for this and the link to the mail test page if it helps.


-brit
Click to expand...

ok we've had a few servers still doing rDNS check on 587, I will have it checked. This sounds exactly like the delays in reply for rDNS, that would cause this.
 
Thanks for checking into it, but I have not noticed any difference throughout the day, except that the site was down for quite a while with a web.config error displayed.

Were you able to find any issues?


thanks,
-brit
 
This is the issue, however the config looks good and should be ignoring that, on 587, we are checking other settings and going to unix manager as soon as he is in.
 
Stephen said:
We did a temp fix on it and instant now.
Click to expand...

Wow, you're absolutely right. That was definitely the fix.

I have a few other sites experiencing the same issue. Can that fix be ported for those sites as well?

Huge thanks, again.


-brit
 
zero said:
Wow, you're absolutely right. That was definitely the fix.

I have a few other sites experiencing the same issue. Can that fix be ported for those sites as well?

Huge thanks, again.


-brit
Click to expand...
I will see, what we did is made an rDNS entry for it.
 
Would like to ask another simple question where did you apply the reverse DNS, was it on the web server or on the mail server.
 
It was on the web server, all the mail servers already have rDNS.
The slowness at the time was being caused by the local mail server forcing a rDNS lookup and waiting on a timeout for that process before it would accept mail. It was set to ignore that but not using those settings, to my knowledge we got this resolved the day after in settings.
 
You must log in or register to reply here.
Back
Top