Basically order confirmation emails are not being sent from uCommerce, all settings are correct and in place with regards to smtp setup, it worked on our supplier's site, now it's migrated to our environment the config files have been updated and now it doesn't work.
With a bit of investigation the issue appears to be a problem with our load balancer (88.x.x.x) and we don't know how to fix it; essentially, from the web server itself we can not use the internal IP address (10.x.x.x) or the DNS entry itself (www.website.co.uk) otherwise the request goes back up to the load balancer (88.x.x.x) and causes a continuous loop, thus the web server can never generate its own emails to send the email...
Our setup is as follows:
Baracuda load balancer - (88.x.x.x)
Web server - (10.x.x.x)
DNS entry is poiting at the Load balancer i.e. www.website.co.uk --> (88.x.x.x)
So in other words any incoming requests to www.website.co.uk go to the load balancer first and then are forwarded on to the webserver on port 81 (10.x.x.x:81), the web server is then trying to access http://www.website.co.uk/order/emails/order-confirmation.aspx?orderNumber=TEST-82 i.e. trying to access itself, which in turn is going back to the load balancer again and repeating the above causing a loop!
Does this make sense?
I'm not sure if this is a bug or something to do with our setup but if someone could help with either it would be much appreciated!
I've given this some thought and asked around and apparently this type of problem is pretty common with load balancing scenarios.
The fix you can try is to edit the hosts found in /windows/system32/drivers/etc file on each of the servers and redirect www.website.co.uk to the internal IP of the server thus circumventing the load balancer for requests made from the servers themselves.
Email problem
Ok bit of a wierd one so bare with me....
Basically order confirmation emails are not being sent from uCommerce, all settings are correct and in place with regards to smtp setup, it worked on our supplier's site, now it's migrated to our environment the config files have been updated and now it doesn't work.
With a bit of investigation the issue appears to be a problem with our load balancer (88.x.x.x) and we don't know how to fix it; essentially, from the web server itself we can not use the internal IP address (10.x.x.x) or the DNS entry itself (www.website.co.uk) otherwise the request goes back up to the load balancer (88.x.x.x) and causes a continuous loop, thus the web server can never generate its own emails to send the email...
Our setup is as follows:
So in other words any incoming requests to www.website.co.uk go to the load balancer first and then are forwarded on to the webserver on port 81 (10.x.x.x:81), the web server is then trying to access http://www.website.co.uk/order/emails/order-confirmation.aspx?orderNumber=TEST-82 i.e. trying to access itself, which in turn is going back to the load balancer again and repeating the above causing a loop!
Does this make sense?
I'm not sure if this is a bug or something to do with our setup but if someone could help with either it would be much appreciated!
Cheers
Ben
Hi Ben,
I've given this some thought and asked around and apparently this type of problem is pretty common with load balancing scenarios.
The fix you can try is to edit the hosts found in /windows/system32/drivers/etc file on each of the servers and redirect www.website.co.uk to the internal IP of the server thus circumventing the load balancer for requests made from the servers themselves.
Hope this helps.
is working on a reply...