Re: [Declude.JunkMail] OT: DNS Failover advice
Forgot to add the most important part regarding Simple DNS. They have an add-on monitoring piece that will switch DNS records automatically, and this can be used to automatically switch over to the backup. Matt Matt wrote: Rob, As far as DNS goes, the best way to do this is to use Simple DNS Plus with a server in a second location. Simple DNS does full server replication instead of individual secondaries, and if you have a lot of domains, it is nice to just manage one installation. If you have a smaller number of zones, it is easy to just set up secondaries with any software. I don't generally recommend large DNS services because they have been attacked and brought down, and that would be a single point of failure even though the providers claim to be immune from such attacks. Look up the "Blue Security" for one such example. This attack also brought down some of Tucow's systems for over 12 hours, including their E-mail hosting/filtering service. My company just started with VMware's hosting provider program to provide legitimate hosting on VMware ESX (virtual servers). VMware is an enterprise solution unlike most of the others on the market, and they have a lot of very nice features and add-ons for fail-over and replication. If you have multiple servers that could be placed on a big VMware server, you could save a lot of money by going with this approach since the hardware costs are greatly reduced. Administration is also simplified, and restoration or moving of the guest operating systems is a breeze. VMware is the future. As far as regional redundancy goes, you would be best off by moving way outside of Chicago. You likely won't get much more in terms of redundancy by going to Milwaukee than you would by going to another colo in Chicago. You want to be on a different power grid, and you want to be on a completely separate provider's network. If something is big enough to affect all of Chicago, it is big enough to affect Milwakee too. If you are in need of some assistance, feel free to give me a call at (888) 862-9042 x3. My company does do colocation and many other custom solutions for those that prefer choosing experience, knowledge and capabilities over branding and value. In the very least, advice is always free, and it sounds like there are many avenues for you to explore. Matt Robert Grosshandler wrote: Gents and the occasional lady: You all are the smartest network folks I interact with. If you'd be so kind as to give me your opinion / suggestions on the following, I'd be forever grateful. We're trying to increase the level of uptime and redundancy for our service. To that end, we're looking to establish a hot failover site in a location remote from our current colocation facility. We're in Chicago, we're thinking a driveable city on a completely different grid (Milwaukee, probably.) If the entire Midwest gets nuked, nobody is going to be buying much online. We're looking at approaches to achieve that failover automatically. Our budget and technical expertise aren't large (we now can handle BGP internally if we have to, but we don't have any of the necessary infrastructure to do that, and would very much prefer not to invest in that infrastructure.) We rely on our colo facility to provide bandwidth, routing, internal DNS, etc. (they have great bandwidth, routing, seven providers, etc.) but since there are humans involved, they could screw up, too. We rely on Ultradns for external DNS. Once our users actually reach our firewall, we have great redundancy inside our rack. The most promising approach at this time seems to be to use somebody like ultradns or dnsmadeeasy to provide dns failover. That is, they're watching our site, and if we go down, they switch out A records and point traffic to the backup site. If it matters, we run ms sql, mirroring and log shipping. We'd have the mirror db and the witness in the remote location. Thanks for whatever thoughts you can add to this challenge. DNS failover a workable solution? We'll be looking for a colo facility in Milwaukee or Indianapolis with 4U available if somebody wants to point us there. Yours, Rob = www.iGive.com [EMAIL PROTECTED] --- 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. --- 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. --- 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: [Declude.JunkMail] OT: DNS Failover advice
Rob, As far as DNS goes, the best way to do this is to use Simple DNS Plus with a server in a second location. Simple DNS does full server replication instead of individual secondaries, and if you have a lot of domains, it is nice to just manage one installation. If you have a smaller number of zones, it is easy to just set up secondaries with any software. I don't generally recommend large DNS services because they have been attacked and brought down, and that would be a single point of failure even though the providers claim to be immune from such attacks. Look up the "Blue Security" for one such example. This attack also brought down some of Tucow's systems for over 12 hours, including their E-mail hosting/filtering service. My company just started with VMware's hosting provider program to provide legitimate hosting on VMware ESX (virtual servers). VMware is an enterprise solution unlike most of the others on the market, and they have a lot of very nice features and add-ons for fail-over and replication. If you have multiple servers that could be placed on a big VMware server, you could save a lot of money by going with this approach since the hardware costs are greatly reduced. Administration is also simplified, and restoration or moving of the guest operating systems is a breeze. VMware is the future. As far as regional redundancy goes, you would be best off by moving way outside of Chicago. You likely won't get much more in terms of redundancy by going to Milwaukee than you would by going to another colo in Chicago. You want to be on a different power grid, and you want to be on a completely separate provider's network. If something is big enough to affect all of Chicago, it is big enough to affect Milwakee too. If you are in need of some assistance, feel free to give me a call at (888) 862-9042 x3. My company does do colocation and many other custom solutions for those that prefer choosing experience, knowledge and capabilities over branding and value. In the very least, advice is always free, and it sounds like there are many avenues for you to explore. Matt Robert Grosshandler wrote: Gents and the occasional lady: You all are the smartest network folks I interact with. If you'd be so kind as to give me your opinion / suggestions on the following, I'd be forever grateful. We're trying to increase the level of uptime and redundancy for our service. To that end, we're looking to establish a hot failover site in a location remote from our current colocation facility. We're in Chicago, we're thinking a driveable city on a completely different grid (Milwaukee, probably.) If the entire Midwest gets nuked, nobody is going to be buying much online. We're looking at approaches to achieve that failover automatically. Our budget and technical expertise aren't large (we now can handle BGP internally if we have to, but we don't have any of the necessary infrastructure to do that, and would very much prefer not to invest in that infrastructure.) We rely on our colo facility to provide bandwidth, routing, internal DNS, etc. (they have great bandwidth, routing, seven providers, etc.) but since there are humans involved, they could screw up, too. We rely on Ultradns for external DNS. Once our users actually reach our firewall, we have great redundancy inside our rack. The most promising approach at this time seems to be to use somebody like ultradns or dnsmadeeasy to provide dns failover. That is, they're watching our site, and if we go down, they switch out A records and point traffic to the backup site. If it matters, we run ms sql, mirroring and log shipping. We'd have the mirror db and the witness in the remote location. Thanks for whatever thoughts you can add to this challenge. DNS failover a workable solution? We'll be looking for a colo facility in Milwaukee or Indianapolis with 4U available if somebody wants to point us there. Yours, Rob = www.iGive.com [EMAIL PROTECTED] --- 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. --- 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.
[Declude.JunkMail] OT: DNS Failover advice
Gents and the occasional lady: You all are the smartest network folks I interact with. If you'd be so kind as to give me your opinion / suggestions on the following, I'd be forever grateful. We're trying to increase the level of uptime and redundancy for our service. To that end, we're looking to establish a hot failover site in a location remote from our current colocation facility. We're in Chicago, we're thinking a driveable city on a completely different grid (Milwaukee, probably.) If the entire Midwest gets nuked, nobody is going to be buying much online. We're looking at approaches to achieve that failover automatically. Our budget and technical expertise aren't large (we now can handle BGP internally if we have to, but we don't have any of the necessary infrastructure to do that, and would very much prefer not to invest in that infrastructure.) We rely on our colo facility to provide bandwidth, routing, internal DNS, etc. (they have great bandwidth, routing, seven providers, etc.) but since there are humans involved, they could screw up, too. We rely on Ultradns for external DNS. Once our users actually reach our firewall, we have great redundancy inside our rack. The most promising approach at this time seems to be to use somebody like ultradns or dnsmadeeasy to provide dns failover. That is, they're watching our site, and if we go down, they switch out A records and point traffic to the backup site. If it matters, we run ms sql, mirroring and log shipping. We'd have the mirror db and the witness in the remote location. Thanks for whatever thoughts you can add to this challenge. DNS failover a workable solution? We'll be looking for a colo facility in Milwaukee or Indianapolis with 4U available if somebody wants to point us there. Yours, Rob = www.iGive.com [EMAIL PROTECTED] --- 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.
