RE: Re[6]: [Declude.JunkMail] DNS Changes
Thanks Kevin. I will be sure to pass this info on to the customers that i speak with. From: "Kevin Bilbee" <[EMAIL PROTECTED]> Sent: Thursday, October 09, 2008 2:19 PM To: declude.junkmail@declude.com Subject: RE: Re[6]: [Declude.JunkMail] DNS Changes I have a suggestion since DNS is so critical to Declude. A secure recursive bind implementation can be setup in less than 5 minutes. How is this difficult? There is no learning DNS. Steps - Tested with Bind 9.5.0-P2 Windows XP/2003/2008 Binary Kit 1) Download BIND http://www.isc.org/index.pl?/sw/bind/index.php 2) Install Bind a. Unzip and run BINDInstall.exe b. Give the service account a strong password c. Do not start bind service after install 3) Configure Bind a. Go to c:\windows\system32\dns and give the service account "named" read\write permissions to the etc folder b. Place the attached files into the etc folder. c. Open a command propmt i. CD to c:\windows\system32\dns\bin ii. Run rndc-confgen -a d. Change the first line of the named.conf file to your IP range that needs recursive DNS. I would use the ip address of the mail server and 12.0.0.1. change the 169.14.238.0/24 to the IP of your mail server. acl "dmz" { 169.14.238.0/24; 127.0.0.1; }; 4) Start bind and make some queries This process takes less than 5 minutes. You now have a reliable easily upgradeable recursive DNS dedicated to your mail server. Kevin Bilbee From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Linda Pagillo Sent: Thursday, October 09, 2008 8:47 AM To: declude.junkmail@declude.com Subject: re: Re[6]: [Declude.JunkMail] DNS Changes Sandy, Yes, things may be slowed down a bit by using a DNS server over a > WAN, how many people i speak with who do not have the recursive option > set on their DNS servers... ... even more so, they are using their ISP's DNS server and the ISP > does not allow recursive lookups because of the high traffic. We have no bearing on how people choose to run their business or > educate their employees. I will work on getting a few articles together next week. If you > would like to contribute your extensive knowledge of DNS, shoot me > an email at [EMAIL PROTECTED] and i will glady add your > information. Sent: Thursday, October 09, 2008 4:44 AM To: "Linda Pagillo" Subject: Re[6]: [Declude.JunkMail] DNS Changes > In a perfect world this would be correct, but as you already know > from working in the IT profession, no server, DNS or otherwise has > an uptime of 100%. A single physical "DNS server" may go down, sure, whatever. The DNS config (redundant DNS servers or load-balanced on a virtual IP) used by a mail infrastructure _must_ be 100% as available as the mailservers themselves. I'm certain that everybody on this list who runs a hosting provider or supports a large company completely agrees and has built their infrastructure accordingly. My clients always have DNS resolution -- yes, _100% of the time that they are connected to the internet_ -- as is commonplace in enterprise-class IT (if not in all "enterprise" IT). It is not so in SMB IT, to be sure, but for your (presumably) SMB clients, we are likely talking about making DNS _as available as a single-point-of-failure MX_. That can mean running caching DNS on the same box. If an admin can't keep a modern DNS daemon running on the mailserver, then their mail should be outsourced. Period. > Yes, things may be slowed down a bit by using a DNS server over a > WAN, Will certainly be slowed down, no "may", let's please be clear about this. > but in my experience, it's more reliable to use the OpenDNS servers > with Declude because they are configured properly for use of the RBL > tests. An OpenDNS server is not "more reliable" for RBL lookups than local recursive DNS servers. It is "more reliable" than overloaded ISP DNS servers. That is not the same statement. > You'd be suprised how many people i talk to in a week who have very > little understanding about the role DNS plays in having these tests > work properly. I wouldn't be surprised at all... and I wouldn't be surprised if, nnn months after they magically switch to OpenDNS, they _still_ have very little understanding of DNS and how to troubleshoot SMTP sending and receiving problems. Because you've patched the problem, but you haven't educated them one bit by telling them that DNS -- rather than being the mail-critical, distributed, scaleable, high-performance, learnable, fairly brilliant protocol that it is -- is something they should get from a free pro
re: Re[6]: [Declude.JunkMail] DNS Changes
Sandy, Yes, things may be slowed down a bit by using a DNS server over a > WAN, how many people i speak with who do not have the recursive option > set on their DNS servers... ... even more so, they are using their ISP's DNS server and the ISP > does not allow recursive lookups because of the high traffic. We have no bearing on how people choose to run their business or > educate their employees. I will work on getting a few articles together next week. If you > would like to contribute your extensive knowledge of DNS, shoot me > an email at [EMAIL PROTECTED] and i will glady add your > information. Sent: Thursday, October 09, 2008 4:44 AM To: "Linda Pagillo" Subject: Re[6]: [Declude.JunkMail] DNS Changes > In a perfect world this would be correct, but as you already know > from working in the IT profession, no server, DNS or otherwise has > an uptime of 100%. A single physical "DNS server" may go down, sure, whatever. The DNS config (redundant DNS servers or load-balanced on a virtual IP) used by a mail infrastructure _must_ be 100% as available as the mailservers themselves. I'm certain that everybody on this list who runs a hosting provider or supports a large company completely agrees and has built their infrastructure accordingly. My clients always have DNS resolution -- yes, _100% of the time that they are connected to the internet_ -- as is commonplace in enterprise-class IT (if not in all "enterprise" IT). It is not so in SMB IT, to be sure, but for your (presumably) SMB clients, we are likely talking about making DNS _as available as a single-point-of-failure MX_. That can mean running caching DNS on the same box. If an admin can't keep a modern DNS daemon running on the mailserver, then their mail should be outsourced. Period. > Yes, things may be slowed down a bit by using a DNS server over a > WAN, Will certainly be slowed down, no "may", let's please be clear about this. > but in my experience, it's more reliable to use the OpenDNS servers > with Declude because they are configured properly for use of the RBL > tests. An OpenDNS server is not "more reliable" for RBL lookups than local recursive DNS servers. It is "more reliable" than overloaded ISP DNS servers. That is not the same statement. > You'd be suprised how many people i talk to in a week who have very > little understanding about the role DNS plays in having these tests > work properly. I wouldn't be surprised at all... and I wouldn't be surprised if, nnn months after they magically switch to OpenDNS, they _still_ have very little understanding of DNS and how to troubleshoot SMTP sending and receiving problems. Because you've patched the problem, but you haven't educated them one bit by telling them that DNS -- rather than being the mail-critical, distributed, scaleable, high-performance, learnable, fairly brilliant protocol that it is -- is something they should get from a free provider over the WAN. By the way, I completely support shops that outsource their anti-spam/anti-virus + their mailboxes (and just about everything else) using OpenDNS for web browsing, since otherwise they would have to support their first reliable, recursive DNS server(s). But if you are capable of supporting your own anti-abuse and mailbox servers, _you are capable of supporting a recursive DNS server_. Or you lied about the first part. > I don't consider the questions that are asked by our customers as > "stupid stuff that is not our fault", especially the questions about > how DNS plays an important role in our product. But you know very well what I mean by "stupid stuff...". These are the issues you have to deal with that cause collateral damage to the reputation of your product or service, even though you have no direct control over the problem area. In my password example, people with bad memories or unstuck post-it notes are not your fault. But you don't yell at them, and you don't tell them to rely on somebody else's account. You do the smart thing and reset their password. Likewise for people that can't open their corporate e-mail account because they forgot to plug in their LAN cable when they came back from a trip. You don't hang up on them, and you don't tell to go down to the local coffee shop and use their GMail account. You tell them how to deal with the problem, not how to avoid it. > When a customer comes to me in a panic about their mail backing up > and causing delays, they are quite happy when we diagnose, fix and > educate them about the issue, DNS related or otherwise. I do not see > that as "bad" service. We provide some of the best support > available. If you would like to see the thank you letters and cards > that i receive each year, i will gladly show them to you. I'm not debating whether people are pleased with your service. I am sure they are pleased as punch to have avoided learning something new and nonetheless brought their mailserver back to life (albeit at lower performance). That does not chang
Re: Re[6]: [Declude.JunkMail] DNS Changes
There is also the question of loss of connectivity from point A to OpenDNS server #1 which is all that you have if you setup Declude to use a Single Source DNS server. If anywhere in that path there is an outage you will have no DNS. Far better to learn a little about DNS and run your own. Then you can at least use several other sources even if you care to set it up to use OpenDNS. As for $0.01 I have changed that as of recent events to $0.005 as your 401K has probably gone down that much in recent months. At 09:01 AM 10/9/2008 -0400, Darin Cox wrote: >1. The customer has no control over its availability. >My $0.01. (decreased due to inflation and other financial considerations, >plus being mostly a reiteration of points already made) - This information is intended only for the use of the individual or entity named above. If you are not the intended recipient, you are hereby notified that any disclosure, copying, distribution, or action taken in reliance on the contents of these documents is strictly prohibited. If you have received this information in error, please notify the sender immediately and arrange for the return or destruction of the document(s). Warning: All e-mail sent to or from this address will be received or otherwise recorded by the Corporate e-mail system and is subject to archival, monitoring or review by, and/or disclosure to, someone other than the recipient. --- This E-mail came from the Declude.JunkMail mailing list. To unsubscribe, just send an E-mail to [EMAIL PROTECTED], and type "unsubscribe Declude.JunkMail". The archives can be found at http://www.mail-archive.com.
Re: Re[6]: [Declude.JunkMail] DNS Changes
I have to say I also agree with Sandy. While recommending a free external DNS solution like OpenDNS is an easy fix for many less technical customers, as Sandy has pointed out it is not the best solution. 1. The customer has no control over its availability. With a free external DNS solution there is no guarantee it will be available in the future. This is why an internal or pay-for solution is generally a better choice, especially for something as critical as business mail services. 2. There is a performance hit from using external DNS for mail processing. So again, while recommending it may be an easy fix, and may get you many thanks, the above points should always be discussed so the customer understands the implications of using a solution like OpenDNS. While there is a full range of customer knowledge levels and desired depth/control of a technical solution, I would have to agree that running mail servers and use of a technical solution like Declude should require a background knowledge in DNS and SMTP. I would think that being halfway up-to-speed with the technical background necessary is a much worse and dangerous place to be in running these services than either outsourcing or having a deep enough understanding to do something as simple as set up multiple internal DNS servers with recursion turned on. My $0.01. (decreased due to inflation and other financial considerations, plus being mostly a reiteration of points already made) Darin. - Original Message - From: "Sanford Whiteman" <[EMAIL PROTECTED]> To: "Linda Pagillo" Sent: Thursday, October 09, 2008 4:52 AM Subject: Re[6]: [Declude.JunkMail] DNS Changes > In a perfect world this would be correct, but as you already know > from working in the IT profession, no server, DNS or otherwise has > an uptime of 100%. A single physical "DNS server" may go down, sure, whatever. The DNS config (redundant DNS servers or load-balanced on a virtual IP) used by a mail infrastructure _must_ be 100% as available as the mailservers themselves. I'm certain that everybody on this list who runs a hosting provider or supports a large company completely agrees and has built their infrastructure accordingly. My clients always have DNS resolution -- yes, _100% of the time that they are connected to the internet_ -- as is commonplace in enterprise-class IT (if not in all "enterprise" IT). It is not so in SMB IT, to be sure, but for your (presumably) SMB clients, we are likelytalkingaboutmakingDNS _as available as a single-point-of-failure MX_. That can mean running caching DNS on the same box. If an admin can't keep a modern DNS daemon running on the mailserver, then their mail should be outsourced. Period. > Yes, things may be slowed down a bit by using a DNS server over a > WAN, Will certainly be slowed down, no "may", let's please be clear about this. > but in my experience, it's more reliable to use the OpenDNS servers > with Declude because they are configured properly for use of the RBL > tests. An OpenDNS server is not "more reliable" for RBL lookups than local recursive DNS servers. It is "more reliable" than overloaded ISP DNS servers. That is not the same statement. > You'd be suprised how many people i talk to in a week who have very > little understanding about the role DNS plays in having these tests > work properly. I wouldn't be surprised at all... and I wouldn't be surprised if, nnn months after they magically switch to OpenDNS, they _still_ have very little understanding of DNS and how to troubleshoot SMTP sending and receiving problems. Because you've patched the problem, but you haven't educated them one bit by telling them that DNS -- rather than being the mail-critical, distributed, scaleable, high-performance, learnable, fairly brilliant protocol that it is -- is something they should get from a free provider over the WAN. By the way, I completely support shops that outsource their anti-spam/anti-virus + their mailboxes (and just about everything else) using OpenDNS for web browsing, since otherwise they would have to support their first reliable, recursive DNS server(s). But if you are capable of supporting your own anti-abuse and mailbox servers, _you are capable of supporting a recursive DNS server_. Or you lied about the first part. > I don't consider the questions that are asked by our customers as > "stupid stuff that is not our fault", especially the questions about > how DNS plays an important role in our product. But you know very well what I mean by "stupid stuff...". These are the issues you have to deal with that cause collateral damage to the reputation of your product or service, even though you have no direct control over the problem area. In my password example, people with bad memories or unstuck post-it notes are not your fault. But you don't
RE: Re[6]: [Declude.JunkMail] DNS Changes
Sandy, I agree with you. While recommending OpenDNS is certainly painless and easy for the support desk, it is certainly not the best solution for an "in-house" mail server - especially those running anti-spam products. Just my .02 ~Patrick -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Sanford Whiteman Sent: Thursday, October 09, 2008 4:52 AM To: Linda Pagillo Subject: Re[6]: [Declude.JunkMail] DNS Changes > In a perfect world this would be correct, but as you already know from > working in the IT profession, no server, DNS or otherwise has an > uptime of 100%. A single physical "DNS server" may go down, sure, whatever. The DNS config (redundant DNS servers or load-balanced on a virtual IP) used by a mail infrastructure _must_ be 100% as available as the mailservers themselves. I'm certain that everybody on this list who runs a hosting provider or supports a large company completely agrees and has built their infrastructure accordingly. My clients always have DNS resolution -- yes, _100% of the time that they are connected to the internet_ -- as is commonplace in enterprise-class IT (if not in all "enterprise" IT). It is not so in SMB IT, to be sure, but for your (presumably) SMB clients, we are likelytalkingaboutmakingDNS _as available as a single-point-of-failure MX_. That can mean running caching DNS on the same box. If an admin can't keep a modern DNS daemon running on the mailserver, then their mail should be outsourced. Period. > Yes, things may be slowed down a bit by using a DNS server over a > WAN, Will certainly be slowed down, no "may", let's please be clear about this. > but in my experience, it's more reliable to use the OpenDNS servers > with Declude because they are configured properly for use of the RBL > tests. An OpenDNS server is not "more reliable" for RBL lookups than local recursive DNS servers. It is "more reliable" than overloaded ISP DNS servers. That is not the same statement. > You'd be suprised how many people i talk to in a week who have very > little understanding about the role DNS plays in having these tests > work properly. I wouldn't be surprised at all... and I wouldn't be surprised if, nnn months after they magically switch to OpenDNS, they _still_ have very little understanding of DNS and how to troubleshoot SMTP sending and receiving problems. Because you've patched the problem, but you haven't educated them one bit by telling them that DNS -- rather than being the mail-critical, distributed, scaleable, high-performance, learnable, fairly brilliant protocol that it is -- is something they should get from a free provider over the WAN. By the way, I completely support shops that outsource their anti-spam/anti-virus + their mailboxes (and just about everything else) using OpenDNS for web browsing, since otherwise they would have to support their first reliable, recursive DNS server(s). But if you are capable of supporting your own anti-abuse and mailbox servers, _you are capable of supporting a recursive DNS server_. Or you lied about the first part. > I don't consider the questions that are asked by our customers as > "stupid stuff that is not our fault", especially the questions about > how DNS plays an important role in our product. But you know very well what I mean by "stupid stuff...". These are the issues you have to deal with that cause collateral damage to the reputation of your product or service, even though you have no direct control over the problem area. In my password example, people with bad memories or unstuck post-it notes are not your fault. But you don't yell at them, and you don't tell them to rely on somebody else's account. You do the smart thing and reset their password. Likewise for people that can't open their corporate e-mail account because they forgot to plug in their LAN cable when they came back from a trip. You don't hang up on them, and you don't tell to go down to the local coffee shop and use their GMail account. You tell them how to deal with the problem, not how to avoid it. > When a customer comes to me in a panic about their mail backing up > and causing delays, they are quite happy when we diagnose, fix and > educate them about the issue, DNS related or otherwise. I do not see > that as "bad" service. We provide some of the best support > available. If you would like to see the thank you letters and cards > that i receive each year, i will gladly show them to you. I'm not debating whether people are pleased with your service. I am sure they are pleased as punch to have avoided learning something new and nonetheless brought their mailserver back to life (albeit at lower performance). That does not change the fact that by suggesting that the "right" thing to do for DNS is use a