The irony of your calling someone else clueless...
Drives do indeed have checksums on their blocks. That does not prevent them from sometimes feeding you back garbage anyway -- see misdirected and phantom reads and writes. Since ZFS uses a self-validating merkle tree, whereas disk checksums live in the same block as the data, ZFS is largely immune to this problem.
If you've worked with disks any length of time, as in actually trying to write a robust filesystem, you'd know that disks sometimes lie. They usually work but every now and then they do the most ridiculous things, due to mechanical, electrical or firmware problems. That's why filesystems like ZFS were created (what, you thought Sun spent man-decades of expert time on it for giggles?). kthreadd is correct.
Please just stay away from storage. The topic is much more complicated than you make it out to be.
It was called Slowlaris in large part because of STREAMS support, which was mandatory in SUS v1 and v2. I find it a bit sad that Solaris historically was ragged upon because it basically tried to follow de jure standards to a T, and backwards-compatibility was a high priority.
But that's a different discussion. Nowadays both Linux and Solaris derivatives have similar computational overheads: sometimes Linux comes out ahead, and sometimes Solaris. I think system administration is completely in favour of Solaris, but we obviously have differing opinions on this.
Solaris was abandoned for reasons completely unrelated to SMF and its various technologies. If anything, Solaris 10 was a decade ahead of its time (DTrace, ZFS, zones, crossbow).
It's just that by the time Solaris was opened, the mindshare and momentum was vastly with Linux. It never stood a chance, completely unrelated to its technical merits, and Oracle closing it again just sealed the deal.
Having said that, illumos (an OSS version of Solaris that forked a few years back) is awesome and actively developed by several companies making money with its technologies.
I hate this stuff, none of my admins remember where the damn log files are because they play with it so rarely.
If you use illumos or Solaris 11.2, svcs -L gives you the path to the service log. I use "grep foo `svcs -L bar`" all the time.
I use SMF a lot, and I've rarely had trouble with it. svcs -vx shows you which services are down (as well as the log path, hello), and you take a look at the log to see what your application is complaining about. I have consistently found it easier to use than init scripts, despite me usually having a very conservative OS bent.
While Solaris itself is no longer relevant outside of some enterprise niches, it has an actively-developed OSS fork named "illumos", developed by former-Sun hackers working at several different private companies. There are several distributions -- I use SmartOS in particular, and OmniTI's OmniOS is also excellent.
Anybody who has run software on a non-trivial scale knows how important changelogs are.
They give you some idea of what to expect, but more importantly let you know whether a problem you're having now has been fixed in the upgrade. Although developers would like everyone to run the newest version of software, in practice you don't touch production systems without good reason. Fixed pain points, and maybe security (depending on isolation) are valid reasons. "Because it's there" is not.
Elimination is a stupid move. It's a triumph of marketing at the cost of we who must run this shit.
If you're willing to look beyond AWS, there's something called Manta out there (http://www.joyent.com/products/manta). The data rests on some servers, and you submit UNIX map/reduce jobs. The jobs are run on the nodes where the data is resting, you get a full UNIX environment, and you only get charged as you'd expect (compute time, combined with the cheaper at-rest time). It might be a better fit for what you're doing than your proposal, plus it'll likely be faster too due to reduced data movement.
Backups are necessary. But, you know, it's nice to avoid restoring a backup if you don't need to.
Furthermore, backups don't help with data that flips bits on disk without being noticed, then consumed by the system in some calculation. A lot of filesystems do not notice non-metadata bits flipping, so you could end up end up with false suddenly becoming true in your database, affecting future calculations.
Once you add RAID, failure conditions become yet more complicated, although some of them become less likely. I don't even know where to begin on that one, except that it isn't a panacea.
Except that Ext and ZFS are in different classes. Ext provides few of the protections that ZFS does.
I'm only aware of two filesystems in the same class: ZFS and Btrfs. I'm looking forward to the day I can use Btrfs on production, but until then there's just ZFS. Ext is a non-contender here.
Most of OpenSolaris was under the CDDL, which provides protection from patent claims from Sun (now Oracle). So if you used OpenSolaris, they wouldn't have a case through copyright infringement -- it's an approved open-source license -- or through patents they hold. Reality is complicated, so it's always a good idea to read the license code is released under: http://www.opensource.org/licenses/cddl1.php
In other words: your concern about OpenSolaris specifically is unfounded. DalvikVM wasn't make by Sun and released under the CDDL, so there was no patent protection. This will still have a chilling effect on the Java ecosystem, of course.
In practice I would use Solaris for databases and storing other critical data. Linux has a long way to go before it has something as mature as ZFS, and I wouldn't trust important data on anything less. DTrace adds introspection that is wonderful on a live database as well. Operating systems are tools, so use them for what they're good at.
Not to put too fine a point on it, but libertarians live in a fantasy land. They talk theory, when hard data has been available for years. Empiricism > Wankery.
In other words, the danger lies in your concentration not being on the road. If you're writing text, you're not concentrating on the road. Therefore I'd be surprised if texting didn't have similar risks.
Please, do the right thing. It won't kill you to wait till you're at a destination to text. It might if you do (and others, unfortunately).
Slashdot Mirror
The irony of your calling someone else clueless...
Drives do indeed have checksums on their blocks. That does not prevent them from sometimes feeding you back garbage anyway -- see misdirected and phantom reads and writes. Since ZFS uses a self-validating merkle tree, whereas disk checksums live in the same block as the data, ZFS is largely immune to this problem.
If you've worked with disks any length of time, as in actually trying to write a robust filesystem, you'd know that disks sometimes lie. They usually work but every now and then they do the most ridiculous things, due to mechanical, electrical or firmware problems. That's why filesystems like ZFS were created (what, you thought Sun spent man-decades of expert time on it for giggles?). kthreadd is correct.
Please just stay away from storage. The topic is much more complicated than you make it out to be.
Use apachectl.
It was called Slowlaris in large part because of STREAMS support, which was mandatory in SUS v1 and v2. I find it a bit sad that Solaris historically was ragged upon because it basically tried to follow de jure standards to a T, and backwards-compatibility was a high priority.
But that's a different discussion. Nowadays both Linux and Solaris derivatives have similar computational overheads: sometimes Linux comes out ahead, and sometimes Solaris. I think system administration is completely in favour of Solaris, but we obviously have differing opinions on this.
Solaris was abandoned for reasons completely unrelated to SMF and its various technologies. If anything, Solaris 10 was a decade ahead of its time (DTrace, ZFS, zones, crossbow).
It's just that by the time Solaris was opened, the mindshare and momentum was vastly with Linux. It never stood a chance, completely unrelated to its technical merits, and Oracle closing it again just sealed the deal.
Having said that, illumos (an OSS version of Solaris that forked a few years back) is awesome and actively developed by several companies making money with its technologies.
I hate this stuff, none of my admins remember where the damn log files are because they play with it so rarely.
If you use illumos or Solaris 11.2, svcs -L gives you the path to the service log. I use "grep foo `svcs -L bar`" all the time.
I use SMF a lot, and I've rarely had trouble with it. svcs -vx shows you which services are down (as well as the log path, hello), and you take a look at the log to see what your application is complaining about. I have consistently found it easier to use than init scripts, despite me usually having a very conservative OS bent.
The only thing I don't like about SMF is the XML.
On the other hand, I think Arma 3 was and is one of the best games I've ever spent money on. It's flexible, capable and beautiful.
You're not one of those Arma 2 guys who thinks Arma 3 is the worst thing ever because it doesn't have 15 cut'n'paste Humvee variants, are you?
While Solaris itself is no longer relevant outside of some enterprise niches, it has an actively-developed OSS fork named "illumos", developed by former-Sun hackers working at several different private companies. There are several distributions -- I use SmartOS in particular, and OmniTI's OmniOS is also excellent.
Anybody who has run software on a non-trivial scale knows how important changelogs are.
They give you some idea of what to expect, but more importantly let you know whether a problem you're having now has been fixed in the upgrade. Although developers would like everyone to run the newest version of software, in practice you don't touch production systems without good reason. Fixed pain points, and maybe security (depending on isolation) are valid reasons. "Because it's there" is not.
Elimination is a stupid move. It's a triumph of marketing at the cost of we who must run this shit.
If you're willing to look beyond AWS, there's something called Manta out there (http://www.joyent.com/products/manta). The data rests on some servers, and you submit UNIX map/reduce jobs. The jobs are run on the nodes where the data is resting, you get a full UNIX environment, and you only get charged as you'd expect (compute time, combined with the cheaper at-rest time). It might be a better fit for what you're doing than your proposal, plus it'll likely be faster too due to reduced data movement.
That's mighty odd. The place I work has hundreds of OSOL servers. I've seen ZFS only flake out once.
Use illumos or one of its derivatives instead.
Odd. The company I work for uses ZFS on many thousands of disks, we don't pay Oracle a dime, and we shovel code back to illumos.
Most of the top Solaris talent jumped the Oracle ship long ago. A lot of them are committing code to illumos as part of the jobs.
I'll take a look when I have a moment. Thanks. :)
Backups are necessary. But, you know, it's nice to avoid restoring a backup if you don't need to.
Furthermore, backups don't help with data that flips bits on disk without being noticed, then consumed by the system in some calculation. A lot of filesystems do not notice non-metadata bits flipping, so you could end up end up with false suddenly becoming true in your database, affecting future calculations.
Once you add RAID, failure conditions become yet more complicated, although some of them become less likely. I don't even know where to begin on that one, except that it isn't a panacea.
It's easy to go "F***" if you're clueless.
Except that Ext and ZFS are in different classes. Ext provides few of the protections that ZFS does.
I'm only aware of two filesystems in the same class: ZFS and Btrfs. I'm looking forward to the day I can use Btrfs on production, but until then there's just ZFS. Ext is a non-contender here.
Most of OpenSolaris was under the CDDL, which provides protection from patent claims from Sun (now Oracle). So if you used OpenSolaris, they wouldn't have a case through copyright infringement -- it's an approved open-source license -- or through patents they hold. Reality is complicated, so it's always a good idea to read the license code is released under: http://www.opensource.org/licenses/cddl1.php
In other words: your concern about OpenSolaris specifically is unfounded. DalvikVM wasn't make by Sun and released under the CDDL, so there was no patent protection. This will still have a chilling effect on the Java ecosystem, of course.
In practice I would use Solaris for databases and storing other critical data. Linux has a long way to go before it has something as mature as ZFS, and I wouldn't trust important data on anything less. DTrace adds introspection that is wonderful on a live database as well. Operating systems are tools, so use them for what they're good at.
OECD data that shows countries with universal healthcare spend less and have better outcomes in quite a few important metrics.
Here's a summary for the US: http://www.oecd.org/dataoecd/46/2/38980580.pdf
Not to put too fine a point on it, but libertarians live in a fantasy land. They talk theory, when hard data has been available for years. Empiricism > Wankery.
It's called "forum shopping", and all litigants with a clue take advantage of it.
Whether your case is morally right or wrong, attempting to stack the case in your own favour is a rational action; nobody attempts to lose a lawsuit.
I, on the other hand, would like a stylus. I'd like to draw sketches during note taking. More importantly, I like drawing.
Stating that styluses suck only means you're not in the arts. This is pretty obvious, yo.
IIRC, Opera was actually the first browser to release a (alpha?) version of their browser that supported <video>.
For whatever reason they haven't released one since.
That was Beethoven, and for only part of his life.
Could you at least check facts you're not sure of?
Ah, the classic "I'm special, why are you so blind?" retort. Keep patting yourself on the back.
And I do understand why it's forbidden.
But you're special and won't screw up, right?
It's amazing the rationalizations that people go through. Stop coming up with excuses and pay attention to the road.
You're not special, and you're threatening people's lives with your selfish stupidity.
Talking at all on a cell-phone is a bad idea; studies have found it's similar to DUI. Knock yourself out: http://scholar.google.ca/scholar?hl=en&q=cell+phone+driving+accident
Talking on a phone is equivalent to DUI: http://hfs.sagepub.com/cgi/content/abstract/48/2/381
It doesn't matter if you're using a hand-held or hands-free phone: http://hfs.sagepub.com/cgi/content/abstract/48/1/196
In other words, the danger lies in your concentration not being on the road. If you're writing text, you're not concentrating on the road. Therefore I'd be surprised if texting didn't have similar risks.
Please, do the right thing. It won't kill you to wait till you're at a destination to text. It might if you do (and others, unfortunately).