Jaime Gago Condensing Information Systems From the Vapor Of Data

Maximizing Os X Server Uptime

This is the first of 3 posts about the design I have implemented in my current position. Please note if you plan to use this as a reference that this setup was strictly meant for this particular environment so you *will* have to adapt it.

When it comes to uptime one of the way to go is to look at your design from Murphy’s perspective and come with the appropriate counter measure (e.g N+1 redundancy) for each different issue your can think of.

Here is my approach for an X Serve running Open Directory and AFP with a Vtrak Promise for User Data.

1)Hardware

a)Hard Drives Failures: Everything is cool as Fonzie, the X Serve has 2 drives configured using Apple (software) RAID 1 and the Vtrak is a native RAID system ( RAID 5 as per Apple Recommended Configuration Script)

b)Random Component Failures: The Promise is fully redundant (e.g. 2 RAID controllers), for the X Serve I have an older X serve running Open Directory in Replica mode + AFP.

c)Power Outage: Fully redundant power supplies and UPS. We also have a diesel generator that can fire up, this is San Francisco so “the next big one“ is part of the local culture.

2)Software

a)Operating System: custom backup process running 3 times a month

b)Users Data: custom backup process running once a week

I don't have an external backup but I am looking into Amazon S3.