Thanks largely in part to the donations we have received in the last few months, we were able to meet both of the goals we had set for this year.
I will go into more detail about what we bought in a moment – but first, I on behalf of all the OSgrid Admin team & the OSgrid Foundation (OSgrid, Inc.) would like to thank each and every one of you who donated or helped us afford these new purchases. We really do appreciate all the donations we have received – and we’ve got some great ideas for plans for next year to work out where we should grow next.
But, onto the specifics of what we bought. We bought two new identically specified machines – for those curious these are the specifications, and how much we spent:
- 3x Seagate 5900 RPM 2TB Hard Disk ($149.99 each) – For a total of 6TB of Storage Space
- Intel Xeon 3220 – Quad Core 2.4Ghz (inc.)
- Intel Bare-Bones 1RU Server Combo ($549.99 w/ Processor)
- 4x 2GB Kingston DDR2 1667Mhz RAM (8GB in total, $199.98)
The total is approx. $1,218 per server including the shipping & handling. Which nets us a pair of 6TB Disk, 8GB Memory, Quad-Core Xeon servers – what are we planning on putting these to?
Well the first step is we’re going to be setting up a live backup of the current “UGIM” server using replication – right now our backups are being performed as a whole dump on a regular basis – the problem here is that if we suffer a database failure, we lose all the inventory & users which was registered since the dump was taken (usually that dump is done daily now – however it could mean up to 24 hours of lost content.) — with the replicated server, if we suffer a failure, we have a backup that should be within a couple of seconds of the main database.
Having the backup also lets us perform some of the heavier database queries without affecting the performance of the grid – this includes doing our off-site backups, analytics and more. So overall grid performance might go slightly up. This new backup server will also be mirroring the assets server – hence the extended disk on both of the machines — this will allow us to switch to it with a moments notice in the event of either the primary UGIM or assets machines failing.
The second server is going to be used to replace our current assets machine – the problem with the current machine is that it’s somewhat wrongly specified for use as an asset system – it has a fast processor and plenty of memory; but very little disk space – only a single 1TB disk. While that is suiting our use right now, we don’t think it’s going to last in the medium to long term; in addition the lack of RAID is exposing us to long restore times in the event of a failure.
Both of these systems are being shipped to our sponsor & admin Dave Coyle and Knifejaw Systems who apart from being the sponsor of Plaza 04, has volunteered us some space in his server rack at a very deep discount (on top of providing one of the berths & some bandwidth for free); and he’s also volunteered to spend his time assembling & installing these machines. Thank you Dave – your help is much appreciated.
So where are we going next? Well with our 2009 goals met, we want to start both collecting some ideas, and collecting some donations to cover next years expenses – we’ve left enough money in our account to cover our operations for the next few months, but please don’t stop contributing because we met our goals – while we try and run a fairly lean operation & minimise our expenses, running the grid can be an expensive process; and we’re not joking when we say that your donations do keep us on the air. Of course – we also wish to thank everyone who helps make the grid a place to come visit – the artists, users & testers – your contributions are invaluable too.
So on behalf of all the OSgrid Admin Team – we would like to thank all our supporters, sponsors and users and wish you a very happy holidays and safe Christmas season — and thank you everyone who helped make these purchases possible.
The OSgrid Admin Team