Hi! On 2009/12/15, at 16:05, Henrique Gomes wrote:
Just two small notes and I'll stop talking about SSD before somebody thinks I want to convince anyone instead of just making a suggestion:
Please keep talking, it's the best way to learn and share knowledge! :)
1-No system will be 100% corruption free. Yes even fancy caches with battery. We are talking about probabilities here. And the case of a SSD losing power while erasing a page to write a block is a possibility but with low probability. In fact, that operation on the SSD firmware can be logged or be done on a COW mode.
Sure it can. SSD firmwares could be good and reliable also. I'm just saying some of them are not. :) And unfortunately, they are made as black boxes, so there's no way to go around that.
(I really hope next generation filesystems ALL implement checksumming like ZFS, btrfs etc... that's the way to go)
They all will, eventually, and they all will support native access to raw devices, like SSD, optimize their writing operations according to the cell capacities, etc.
2- I know that's not the case here, but I always cringe when I see the faith some people put in RAID. It's not a magic bullet. Mechanical drives have a higher probability of failing than the rest of the system so it makes sense, specially to avoid down time.
You are being a little hard on hard drives. Hard drives today are very reliable, most of them work for 3 years without a glitch (there is a Google-authored paper about this, and I think they have a very large sample ;) ). I also do assure you that, in average, hard drives last longer that a lot of computers, specially Apple crappy laptops. All the people I know, but one, with plastic MacBooks had a failure on the machine that was related to some electronics part, and not the drive. Only one of them burned the drive before the Mac itself.
Ask around how reliable are SSDs compared to that, and you will be very surprised. I'm not talking about data corruption only, but the drive itself to completely fail. If you compare the return rate of the same model and brand of a laptop, with two different drive options (one with HD, other with SSD), the SSD return rate is much higher.
Data safety is guaranteed by backups. Only by backups.
It's all probabilities. What if your backup fails? :) But yes, backups are essential, whatever you use as main storage.
On that note, I don't see the point of mirroring or striping SSD. Are they really that much unreliable than the computer itself?
As I said, yes. A lot, really.
The power supply?
God, yes. I deal with dozens of Apple machines, and so far, only one power supply failed (ironically, the xServe one).
The operator of the computer? :-)
That fails a lot. :)
My only doubt for a database server would be if the workload implied a lot of *sustained* writes. Thats a point where SSDs just don't cut it.
And not just that. Intel SSDs, for instance, start to get very very slow at some point, due to the way the data is being organized, shuffled, stored, etc, inside the drive to avoid hitting the same cells over and over. The only way to fix it is to perform a low level format.
Don't get me wrong, I can hardly wait the day where SSDs will be reliable enough to be used seriously, killing seek times and making some stuff (like databases) faster by several orders of magnitude. I want my kids to look at hard drives the way I look at data tapes or punched cards (ok, I really don't want to have kids, but that's beside the point!). But as I said, our industry (as every other in this damn planet) is in the "Ship now, fix later" mode, which gives people the illusion the technology is evolving faster than ever, although, in reality, it just keeps f*****g everything up, and destroying what could be a great idea, for years, or for ever. I believe "SSD" is included in the "for years" category, but those years still did not went by. I'll wait for that moment, patiently spinning at 7200rpm. :)
Yours Miguel Arroz
smime.p7s
Description: S/MIME cryptographic signature
_______________________________________________ Do not post admin requests to the list. They will be ignored. Webobjects-deploy mailing list ([email protected]) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/webobjects-deploy/archive%40mail-archive.com This email sent to [email protected]
