 |
 |
>>> On 6/15/2009 at 8:53 AM, Brian Martinez <[log in to unmask]> wrote: > Al Puzzuoli wrote: >> Just curious as to what the story is on this weekend's mail outage? >> I called ATS today because I noticed that I hadn't gotten any >> external email since yesterday afternoon. Case number 38773 was >> created and as of about an hour ago, external mail began to come in >> once again. It would seem that the problem is fixed; but my concern >> is for any mail sent during the trouble period? in the comments for >> the case, I saw reference to MX13 being down. Can we expect messages >> sent yesterday to trickle in as servers retry delivery, or are they >> just gone? >> > > Al, > > Sorry, just getting back in from a long weekend otherwise I would have > followed up on this yesterday. There appears to be one mail server that > was hung, and as you noted it was mx13. I am still investigating what > is/was wrong with mx13, in the meantime it is out of our load balancer. > All mail servers that are RFC-compliant will resend their messages, in > fact I've seen a bunch of messages from list trickle in already. > > Regards, > ./brm Perhaps yes. This brings up a fun topic for debate. What should the time-out be on mail retries? Mail is quickly becoming a victim of it's own success. Everyone (broadly speaking) expects mail to arrive in a near-instantaneous time frame. I have been given pressure in the past to dramatically reduce the retry period so that users become aware if there are any problems with this expectation. It is possible that someone has reduced the timeout period to less than 24 hours. In this case, they will receive a bounce message rather than the system retrying. Long story short--Most likely the mail will trickle in today. Any mail that does not arrive today will need to be resent.