Synchronization tools
1.Lsyncd - Live Syncing (Mirror) Daemon[Directory level]
2.DRBD.[block device level]
Reference:
-----------------
https://code.google.com/p/lsyncd/
http://configure.systems/glusterfs-and-why-you-should-consider-it/
GlusterFS would actually mitigate and simply so much more of that. There would be no need for a Load Balancer, no need for a special script to promote, demote the content servers, nothing, not even to replicate the data between the servers!
Basically, you can create two or more servers, install GlusterFS on each of the servers, have node all of the nodes probe the master node, then you would create the volume. Easy.
Once that’s done, one your actual web nodes, where you have Apache, PHP, and again Varnish installed, you would install GlusterFS, add the correct line to the /etc/fstab, and you’re set. Within that line, you can even add a failover server in case the primary goes down! Say what? Not only that, when it comes back up, it can self-heal to ensure consistency across all servers again.
Adding more servers to the GlusterFS environment is pretty simple too, couple commands and you’re good to go. All of this could even be automated.
There are some other comparable options but GlusterFS seems to be a very viable option, one that I use on this servers configuration. I’m not a big site, nor do I serve tens of thousands of users. However, I’m completely ready to scale at the drop of a hammer if need be. Both from a saved image and from a completely orchestrated manner with Ansible. The less moving parts, the better. Keep everything dedicated one set of resources and you’ll be building for success.
http://configure.systems/glusterfs-and-why-you-should-consider-it/
GlusterFS would actually mitigate and simply so much more of that. There would be no need for a Load Balancer, no need for a special script to promote, demote the content servers, nothing, not even to replicate the data between the servers!
Basically, you can create two or more servers, install GlusterFS on each of the servers, have node all of the nodes probe the master node, then you would create the volume. Easy.
Once that’s done, one your actual web nodes, where you have Apache, PHP, and again Varnish installed, you would install GlusterFS, add the correct line to the /etc/fstab, and you’re set. Within that line, you can even add a failover server in case the primary goes down! Say what? Not only that, when it comes back up, it can self-heal to ensure consistency across all servers again.
Adding more servers to the GlusterFS environment is pretty simple too, couple commands and you’re good to go. All of this could even be automated.
There are some other comparable options but GlusterFS seems to be a very viable option, one that I use on this servers configuration. I’m not a big site, nor do I serve tens of thousands of users. However, I’m completely ready to scale at the drop of a hammer if need be. Both from a saved image and from a completely orchestrated manner with Ansible. The less moving parts, the better. Keep everything dedicated one set of resources and you’ll be building for success.
Comments
Post a Comment