There are three new boxes, all running as Scoop servers, as follows:
There is also another one like the above that is having some issues, but should come up in the next couple days.
- Dual P3 1Ghz
- 3 Gb RAM
- 9 Gb SCSI
- Slackware 8.0
- Kernel 2.4.10-ac12
So right now, we have the three machines above, plus a dual P3/700 all running Scoop, and venerable old hex (dual P3/700 as well) running the database.
When the last Scoop box is live, we'll move the oddball 700 machine back a layer, where it will mirror the database in real-time, and take over automatically if hex fails. And at that point, we will have acheived full redundancy. To interrupt service, at least two different machines will have to fail. Not that that's not possible, with our luck.
So, if you notice that things are faster today, that's why. If not, then something's probably wrong. Don't tell me though, I'm asleep. Oh, and if anything seems to be behaving oddly, post a comment below.