Peter Clahsen
asked on
Exchange 2016 large (10-20Mb) messages stuck in queue
Good day folks, and thank you to anyone that may be able to assist!
I have a rather strange situation - we have an Exchange 2016 machine which happily delivers short e-mails, yet large (10-20Mb) mails will often sit in the queue showing "active".
This behaviour seems a little random, as I have seen some messages of this size get delivered quickly, some will sit in the queue for many hours, and some will bounce back to the sender as undelivered (the sender being one of the staff within the organisation).
We talked to one recipient's IT chap, who was able to show us a log that indicated the mail began sending, opened the data channel, but didn't complete the send - in the end, their mail server gave up, and closed the transmission channel.
We have ensured that all message size limits are well beyond the size of mail we're trying to send - minimum size we have is 35Mb, and the message after encoding we are testing with now is 17Mb.
We have set up a delivery rule for a domain we can use to test with to go via another mail server on which we can see the logs, and when messages do not clear the queue in a timely fashion, we see the same behaviour as was originally reported to us - Exchange is opening the channel, beginning to send by giving the headers, then timing out in actually sending the data.
Resource monitor on the Exchange server shows a machine that is effectively near idle - 2Mb/sec network activity, under 1Mb/sec disk activity, memory (16Gb) is obviously fairly full, as is true on most Exchange servers, CPU is doing next to nothing.
For mail collection, for safety, we have all mail collect at an ISP in POP mailboxes, which we check via MAPILabs POP3 connector - incoming mail is operating like a charm and no issues whatsoever.
The server connects back to the Internet on a 10Mb symmetrical connection.
We have set up an iperf test between the Exchange server and an outside destination, it showed 2Mb/sec throughput.
I am getting to the end of things to try and test here - my gut feeling is that this is a bandwidth issue at the ISP, but not only is this going to be tricky to prove, I am unsure as to how we might definitively test this (the iperf test does not use port 25).
Has anyone any ideas or thoughts they could share please?
I have a rather strange situation - we have an Exchange 2016 machine which happily delivers short e-mails, yet large (10-20Mb) mails will often sit in the queue showing "active".
This behaviour seems a little random, as I have seen some messages of this size get delivered quickly, some will sit in the queue for many hours, and some will bounce back to the sender as undelivered (the sender being one of the staff within the organisation).
We talked to one recipient's IT chap, who was able to show us a log that indicated the mail began sending, opened the data channel, but didn't complete the send - in the end, their mail server gave up, and closed the transmission channel.
We have ensured that all message size limits are well beyond the size of mail we're trying to send - minimum size we have is 35Mb, and the message after encoding we are testing with now is 17Mb.
We have set up a delivery rule for a domain we can use to test with to go via another mail server on which we can see the logs, and when messages do not clear the queue in a timely fashion, we see the same behaviour as was originally reported to us - Exchange is opening the channel, beginning to send by giving the headers, then timing out in actually sending the data.
Resource monitor on the Exchange server shows a machine that is effectively near idle - 2Mb/sec network activity, under 1Mb/sec disk activity, memory (16Gb) is obviously fairly full, as is true on most Exchange servers, CPU is doing next to nothing.
For mail collection, for safety, we have all mail collect at an ISP in POP mailboxes, which we check via MAPILabs POP3 connector - incoming mail is operating like a charm and no issues whatsoever.
The server connects back to the Internet on a 10Mb symmetrical connection.
We have set up an iperf test between the Exchange server and an outside destination, it showed 2Mb/sec throughput.
I am getting to the end of things to try and test here - my gut feeling is that this is a bandwidth issue at the ISP, but not only is this going to be tricky to prove, I am unsure as to how we might definitively test this (the iperf test does not use port 25).
Has anyone any ideas or thoughts they could share please?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
It's OK bro, I just wanted to help you out in case you still having the issue xD
Good to hear you solved it.
Cheers,
Good to hear you solved it.
Cheers,
ASKER
Problem was solved after further investigation
There's certain things that needs to be considered before to move on.
1- How many users do you have in your email server?
2- Do you have an estimate of how much email the company sends and receives during the day?
3- Are you using the same internet connection for the email server and as well for the company traffic?
4- Since you have symmetrical connection 10/10 I'm guessing you are using DSL connection am I right? if it so, then do you know how much bandwidth is you network utilizing?
Regards,