My home server is a late 2012 Mac mini server running Mavericks on 2 1TB drives and 8GB RAM. It runs OS X Server with Time Machine, Caching and File Sharing services enabled.
Combined with a fixed IP it allows me to backup to my server from virtually anywhere where I’m online, and to get data from my Data HD via AFP or WebDAV over VPN.
It’s a stable setup, that thanks to Apple’s Server.app doesn’t require any real maintenance.
Last summer when Apple released Mavericks, I upgraded an old 2008 mini with this new model, and things ran smooth ever since.
It’s my media center so it has to stream and convert pleny of data. It’s my development server, so it has plenty of python and php scripts running.
But the Mac mini is only part of the story. Next to it are two Drobo’s, first generation devices released and active since 2008. I bought them cheap from the company I work for and gradually updated them from their original 4x250GB drives to 4x2TB drives. This gives me around 5TB of actual data on each Drobo.
I run a CarbonCopyClone across both Drobo’s on a nightly basis, interrupting the process when more than 5% of data will be affected. (This to prevent any serious dataloss).
I had a couple of dead drives througout the years, but basically I never had any problems with these two beasts.
That is, up to these last couple of weeks.
The main Drobo started acting up, and it required reboots of the entire server on a weekly basis in order to get everything running smoothly again.
Today I got a couple of alert emails throughout the afternoon, and came home to a server running blazingly hot and unresponsive.
Three hours later and a couple of reboots, PRAM-zaps and safe boots the server is up and running again, with both Drobo’s fully operational. It required a few hard resets of all devices, uncoupling drives, testing if the Drobo would mount over USB or Firewire, and finally jumpstarting the volume mounting via Terminal with a Disk Repair. Everything is online again, and no data lost (although I’m currently running a comparison between Data and Clone to verify if the backup is 100% OK.
Three hours of me troubleshooting isn’t worthy of my time. I already do these kinds of things at the office, so at home I want a working setup without any hassle. An evening with my Drobo was not really what I planned doing.
Considering what three hours of consultancy costs for a client, if I count all the hours I spend keeping these relics running, it probably would’ve been cheaper of replacing them from the moment they started acting up.
So tomorrow I’ll order a new Promise Pegasus2 RAID with Thunderbolt with 4x2TB drives. Similar capacity but waaaay faster. I’ll keep the old Drobo running as a mirrored backup, and use the other one for spare parts.
Should’ve done this a long time ago.