-
Servers Going Down
I have a web server I host a few sites on. Today something happened... Not sure what, I think the power supply died. Well the web server went down. Since the server is down, the sites that were on there can't be reached.
Now of course I have all the sites backed up on another machine on the network. I can get IIS up on another machine, and set that machine up to use the IP Address that the web server was using. Then setup IIS to serve the web sites.... But all that is a pain in the ass, not to mention I have to be physically in the office to set that up, or use PCAnywhere to get on the machine.
My question is: Is there anyway that if a web server goes down, another server can automatically pick up where it left off? It would have been real convenient that when this power supply died, another server served the sites automatically while we fix up that machine. Any ideas, suggestions, theories that would help avoid this dilema in the future?
Thanks in advance.
Jared
-
Sure, you could be running monitoring scripts on a separate monitoring server that ping the web server every 5 minutes or so. If the web server doesn't answer a ping for say 30 minutes, the script could automatically modify the DNS tables to route traffic to the new server. The major problems with that is that any DNS modifications take time to propagate, and if you're not in control of the DNS server, you can't make the changes. Hmm....
There's gotta be another way that's easier than that. Lemmie ponder for awhile. I'll try to come up with some other solution.
-
roswell1329, thanks for the suggestion... I do have control over the DNS server. I cannot program though, so writing that type of script would be kinda tricky for me. :D
I'm thinking the same thing, there has to be an easier way.
It figures... I make sure to get battery backups, I make sure to setup a file backup for twice a week, and the power supply goes on me when I'm not at the office.... Doesn't go out on the secondary DNS machine... It goes out on the one web server. Ha... Damn Murphy's law!
-
I'm surprised that no one else has responded to this...
Is there no way to ensure that web sites stay up when disaster strikes?
-
I think most people who have a critical web server are likely to have a backup machine ready to go with the same configs as the primary. Using that method, all you would have to do is change the IP number of the secondary system to the primary IP and voila! This process would likely require some human intervention, though. You can be alerted of these failures, though with the monitoring scripts that I mentioned before. You don't have to have the scripts change anything themselves, just alert you of the problem. There are several monitoring packages out there that do exactly that. You don't have to program at all. The best one I've seen so far is Big Brother, and it's released under a fairly relaxed license. Hope this helps some. Other than switching IP's, I couldn't really come up with an automatic way of the redundancy you're looking for.
-
I would go for a cron job powered bash script
but hey, I'm a script-adult..
Find someone to program a small thingy like rosswell adviced.
-
If you have money you can set up a server cluster that way there would be no downtime at all.
Another method would be to use a CISCO Local Director. I am using one of these for a critical web app and it works great. I have two identical web servers running behind the Local Director and if one fails the other one is used with no interuption to the user and no lag time for DNS changes.
-
Quote:
Originally posted here by mmelby
If you have money you can set up a server cluster that way there would be no downtime at all.
Another method would be to use a CISCO Local Director. I am using one of these for a critical web app and it works great. I have two identical web servers running behind the Local Director and if one fails the other one is used with no interuption to the user and no lag time for DNS changes.
There it is, the official "Easier way than that". Thanks mmelby!
