EC2 has only recently got an SLA. I haven't had time to investigate the SLA document thoroughly yet, but even so I'm happy to stick apps on EC2 because nothing I currently work on is mission critical. In the context we were discussing I'd be a bit more skeptical.
S3 has recently been launched in Europe, although TBH it still seems slow. On some applications S3 delivery is so slow that we've had to use Akamai. The big problem with that is the massive price difference between S3 and Akamai bandwidth. The Amazon services look great, but they need a little more lovin' yet. On 31 Jan 2009, at 22:31, John L. Singleton wrote: > >> >> On Jan 31, 2009, at 4:34 PM, Lee Bolding wrote: >>> Last year (or it may be the year before now...) there were several >>> cases of hosting centres loosing power, and the backup generators >>> failing because they were never tested. >> >> >> The data center I was at was one of those. They lost power, the UPS >> took over like it was supposed to. The generator started and took >> over like it was supposed to. When it was time to switch back to >> mains power; *pop* the main breaker popped and power was lost. >> AFAIK, >> this was the first time the generator was tested under load. >> >> They also had the power company come out to switch meters. The meter >> *exploded* and sent a guy to the hospital (he was OK). Again, all >> servers in the center went down hard. >> >> Another data center I've used switches to the generator *every* >> Friday. Which is good. Bad part is they don't have enough UPS >> capacity for the entire data center - you have to provide your own or >> just accept that your servers will reboot every Friday at the same >> time. >> > > Ditto that. I had a project at Alchemy (where MySpace hosts here in > LA) and there was a massive power outage. The generators kicked in, > but someone borked something with the fuses that handled the circuit > to our cage. Fuse gone. Site down. No spare fuses on hand. UPS > drained. Bad times. You get the picture. Though almost killing someone > is a little more dramatic ;) > > While we are talking about it, I thought I'd throw in that Amazon's > EC2 is a great environment for clustering. Reason being that for > comparatively little money you get access to a huge infrastructure > that would cost many many dollars to build yourself. As a simple > example, EC2 allows you to create snapshots of your XFS volumes (such > as a mysql database) and persist the deltas redundantly to S3. > Considering how important backups are that's a good reason to consider > EC2. Rolling something like that yourself would cost a fortune. Ie, > you make backups to a different machine, fine, but what if that > machine fails? What's backing up your backup? As Jacob pointed out, > there are two reasons to backup, really. The first is just protecting > against failures leading to data loss. Backups with things like RAID > help with this. The second is to protect against, you know, malicious > or erroneous things, like hackers, or accidently running a rm -f * on > your data directory. RAID ain't gonna help you here. But again, I > don't know the level of redundancy, liability protection, and traffic > your site requires. Something to think about... > > > > On Jan 31, 2009, at 2:11 PM, Jacob Coby wrote: > >> >> >> On Jan 31, 2009, at 4:34 PM, Lee Bolding wrote: >>> Last year (or it may be the year before now...) there were several >>> cases of hosting centres loosing power, and the backup generators >>> failing because they were never tested. >> >> >> The data center I was at was one of those. They lost power, the UPS >> took over like it was supposed to. The generator started and took >> over like it was supposed to. When it was time to switch back to >> mains power; *pop* the main breaker popped and power was lost. >> AFAIK, >> this was the first time the generator was tested under load. >> >> They also had the power company come out to switch meters. The meter >> *exploded* and sent a guy to the hospital (he was OK). Again, all >> servers in the center went down hard. >> >> Another data center I've used switches to the generator *every* >> Friday. Which is good. Bad part is they don't have enough UPS >> capacity for the entire data center - you have to provide your own or >> just accept that your servers will reboot every Friday at the same >> time. >> >> -- >> Jacob Coby >> >> >> >> >> >> >> >>> > > > > --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "symfony users" group. To post to this group, send email to symfony-users@googlegroups.com To unsubscribe from this group, send email to symfony-users+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/symfony-users?hl=en -~----------~----~----~----~------~----~------~--~---
