SurgeMail Archive: Re: [SurgeMail List] @me.com and @icloud.com

Subject:	Re: [SurgeMail List] @me.com and @icloud.com
From:	Ed
Date:	10-Sep-2014
Hi Frank, We don't use m$ stuff here, linux has all that wonderfullness built in. I did find something that very likely would explain the errors we are seeing. Thanks for the suggestion. A little primer : me.com and aliases have 2 basic blocks of IP's they advertise the MX's on at least that are apparent to us -- don't know if they're using geo-ip management or not. On one block we can deliver email without issue, the other it fails most times. Good block 17.172.* Bad block 17.158.* I should note that the "bad block" is deliverable sometimes. Also explained by the results. So without further a due I picked 2 IP's from their MX1 advertisement : --- Bad Block!!! Notice the MTU mismatch when it hits 17.1.152.1 .. [ line 6 ] Drops to 1488. That explains the error messages perfectly. And it's all apple. tracepath 17.158.8.67 1?: [LOCALHOST] pmtu 1500 1: gateway.easent.net 1.103ms 1: gateway.easent.net 1.075ms 2: 66.192.139.133 4.138ms 3: mia2-pr1-xe-0-3-0-0.us.twtelecom.net 15.307ms asymm 5 4: 17.1.152.1 12.509ms asymm 5 5: no reply 6: 17.1.152.1 12.855ms pmtu 1488 6: 17.0.150.82 85.399ms asymm 7 7: 17.0.150.64 85.346ms 8: 17.0.150.81 85.913ms asymm 7 9: no reply 10: no reply 11: no reply 12: no reply 13: no reply 14: no reply 15: no reply 16: no reply --- good block !!! No MTU mismatch But yet when routed on the good block side when it hits 17.1.152.1 it doesn't exhibit an MTU change. tracepath 17.172.34.9 1?: [LOCALHOST] pmtu 1500 1: gateway.easent.net 1.150ms 1: gateway.easent.net 1.193ms 2: 66.192.139.133 18.927ms 3: mia2-pr1-xe-0-3-0-0.us.twtelecom.net 18.709ms asymm 5 4: 17.1.152.1 23.239ms asymm 5 5: 17.0.140.114 50.572ms asymm 7 6: no reply 7: no reply 8: no reply 9: no reply 10: no reply 11: no reply So now the great mystery has been solved. Will it ever be resolved ? I seriously doubt it. --Ed On 09/10/2014 12:41 AM, Frank Bulk wrote: > If you ever wanted to chase down the MTU issue from a Win32 perspective you could try this free tool: http://www.elifulkerson.com/projects/mturoute.php > > Pleas share the working and not-so-well-working IPs so we can compare to our own logs. > > Frank > > -----Original Message----- > From: Ed [mailtoHIDDEN@ent.net] > Sent: Monday, September 08, 2014 10:12 PM > To: surgemailHIDDEN@etwinsite.com > Subject: Re: [SurgeMail List]HIDDEN@m and @icloud.com > > Hi, > > PMTU I don't think so. > > We did find that me.com and aliases advertise a whole big pile of MX's. > After some testing we found that some MX's work great every time, some > fail most times -- not all [how odd]. Hmmm ! So luck of the draw ? > Also we found that each of those MX's actually front for multiple > different servers ie load balancing and they overlap too. Oh so cool ;-) > > Knowing the IP of a good one we just added a little configuration for > those 2 domains to route to that IP and life is one again good, at least > for now. > > We will have to get into this a bit deeper. I was actually entertaining > the idea we've got an MTU mismatch someplace upstream, seen that action > before but I'm not really sure of that because those 2 domains are the > only ones giving us a problem. And it's obviously not an RBL issue either. > > Thanks! > --Ed > > On 09/08/2014 10:59 PM, Frank Bulk wrote: >> We've not seen any issues. Anything possibly PTMU related? >> >> Frank >> >> -----Original Message----- >> From: Ed [mailtoHIDDEN@ent.net] >> Sent: Monday, September 08, 2014 11:14 AM >> To: surgemailHIDDEN@etwinsite.com >> Subject: [SurgeMail List]HIDDEN@m and @icloud.com >> >> Has anybody out there had an issue sending email to me.com and >> icloud.com which are really the same place. >> >> This all seems to have started a couple months ago. It's the only place >> we're having an issue with so not a network or server issue because all >> of our servers are experiencing the same issue and have zero delivery >> issues any place else. >> >> We're getting >> after data sent: 421 4.4.2 Timeout while waiting for command. >> and >> Write failed 100 times or 550>540 seconds 550, wrote 37305/1485 left >> (37305) >> >> which is quite bizarre. >> >> If so did you fix it and how. >> >> Thanks! >> -- Ed >> ----------------------------------------------------------- >> EAS Enterprises LLC >> World Class Web and Email Hosting Solutions >> IPv6 ready today for your needs of tomorrow! >> Ask us about dual-stacking your site >> www.easent.net >> >> >> >> > -- ----------------------------------------------------------- EAS Enterprises LLC World Class Web and Email Hosting Solutions IPv6 ready today for your needs of tomorrow! Ask us about dual-stacking your site www.easent.net

Last Message | Next Message


Search website: