Friday, May 9, 2008
Posted by Jason Dunn in "Thoughts Media Status Updates" @ 11:00 AM
Some blame for that can be shared around, especially with The Planet (who's advising sales person didn't bother to point out that issue to me), but it doesn’t change one fact: the SATA hard drives we have on our current server are incapable of the throughput needed to run full site backups whilst still keeping the sites themselves running. Ultimately the blame for this falls squarely on my shoulders since I'm the one with the credit card renting the server. We’ve had sporadic down-time over the past month for exactly this reason: anything that’s disk intensive has the ugly side-effect of bringing down the sites.
You’d think the solution would be to simply upgrade the hard drives, right? Well, The Planet is unfortunately completely inflexible in this regard, and they say the only solution is to order a new server with the hardware we need, and move everything to that. So that’s what we're doing: our new server is an Intel Kentsfield 2.13 Ghz quad-core machine, the same as our current server, and it also has the same 8 GB of RAM, but we’ve gone from two 500 GB 7200 RPM SATA drives in RAID1 to two 300 GB 15,000 RPM SCSI drives in RAID1 – the fastest and biggest they have. If we run into disk throughput issues after this upgrade...then we have a different kind of problem.
Sometime this evening (Eastern Standard Time), we'll be initiating the move. You can expect all our sites to be down, but we’re hoping this will be a relatively quick process. By Saturday AM everything should be back to normal. And I swear by Thor’s Hammer this will be our last server move for quite a while…I hope. ;-)