If host #1 fails, serve from host #2 (emergency only)

[sKunKbad]

I browsed around a little and have seen some solutions, but wondering what the appropriate way to handle a hosting related issue is. The problem is that the host that we have is not from the land of unicorns and rainbows. I really do like the host, and think they are as good as can be expected, but everyone once in a while the site will go down for a few hours because of X, Y, or Z. This is just a glorified or "premium" shared hosting.

 

What I was thinking about doing was to use rsync on host #2 to pull in file changes from host #1 a few times a day. I would also use mysqldump and recreate the database on host #2 a few times a day. If there was a big problem with host #1, we could switch nameservers manually and perhaps the experience would be better than not doing any of this.

 

Does this idea blow? How would you do it?

[QuickOldCar]

If site A goes down people are connecting to site A and it can never go to site B unless you manually set dns and wait it out.

 

I would say dns failover or round robin.

It will need a monitoring service that can determine which dns to send traffic to and a lower TTL

 

The right way would be to use a datacenter and have the same LAN, use HTTP load balancing to handle server failures

 

http://www.haproxy.org/

https://www.digitalocean.com/community/tutorials/how-to-use-haproxy-to-set-up-http-load-balancing-on-an-ubuntu-vps

 

If you ever decided using a datacenter, check out mesosphere

[sKunKbad]

QOC, thanks for the response. Unfortunately, because the website isn't on sophisticated hardware, we really can't use your advice. Perhaps it's time to think about moving the website.