Quotes of the week

[Posted January 30, 2013 by corbet]

People really ought to be forced to read their code aloud over the phone - that would rapidly improve the choice of identifiers

— Al Viro

Besides, wouldn't it be cool to see a group of rovers chasing each other across Mars, jockeying for the best positioning to reduce speed-of-light delays?

— Paul McKenney

The real problem is, Moore's Law just does not work for spinning disks. Nobody really wants their disk spinning faster than [7200] rpm, or they don't want to pay for it. But density goes up as the square of feature size. So media transfer rate goes up linearly while disk size goes up quadratically. Today, it takes a couple of hours to read each terabyte of disk. Fsck is normally faster than that, because it only reads a portion of the disk, but over time, it breaks in the same way. The bottom line is, full fsck just isn't a viable thing to do on your system as a standard, periodic procedure. There is really not a lot of choice but to move on to incremental and online fsck.

— Daniel Phillips

Quotes of the week

Posted Jan 31, 2013 3:47 UTC (Thu) by alfille (subscriber, #1631) [Link] (34 responses)

72000 seems high.

What's an order of magnitude among friends?

Posted Jan 31, 2013 4:00 UTC (Thu) by corbet (editor, #1) [Link] (33 responses)

In fact, he so clearly meant 7,200 that I didn't even notice the extra zero. I took the liberty of fixing it to avoid further confusion.

That said, it is true that nobody wants a disk spinning faster than 72,000 RPM — for now.

What's an order of magnitude among friends?

Posted Jan 31, 2013 4:09 UTC (Thu) by raven667 (subscriber, #5198) [Link] (29 responses)

I imagine that a magnetic disk traveling at 72,000 RPM would probably work great for a few microseconds before impressively exploding into fantastic little pieces of high-speed shrapnel.

What's an order of magnitude among friends?

Posted Jan 31, 2013 4:56 UTC (Thu) by dlang (guest, #313) [Link] (19 responses)

72K rpm isn't _that_ fast, turbines in jets and generators routinely top this. It takes careful construction, and you _really_ don't want to drop it while it's spinning at full speed, but it could be done.

but the price premium that it would have would not be worth it.

What's an order of magnitude among friends?

Posted Jan 31, 2013 15:29 UTC (Thu) by marduk (subscriber, #3831) [Link] (15 responses)

One wonders what the sound of a hard drive spinning at 72k RPM would be like.

What's an order of magnitude among friends?

Posted Jan 31, 2013 17:05 UTC (Thu) by drag (guest, #31333) [Link] (13 responses)

Currently the fastest spinning object that man has produced is a bullet. A high velocity bullet fired from a barrel with rather rapid rifling will spin at speeds approaching 300,000 rpm. Most bullets, having a light construction, would explode due to the centrifugal force at that speed.

Going from that I figure a 72k rpm disk would make a nice sharp ZZzzz-ng noise as it explodes and shrapnel zips past (hopefully) your head.

What's an order of magnitude among friends?

Posted Jan 31, 2013 20:51 UTC (Thu) by rgmoore (✭ supporter ✭, #75) [Link] (12 responses)

Bullets are not known for their rugged construction; in many cases it's considered desirable if the bullet breaks up on impact so it can dump its kinetic energy more efficiently. Devices that are intended to operate at 72K RPM and above are certainly practical. Many of the machines I work with have turbomolecular pumps that operate in that speed range, and their diameter is larger than a disk drive so the force is larger as well. Similarly, ultracentrifuges operate at substantially higher speed than that- they can go above 100K RPM- without exploding. The thing they have in common is that they operate under vacuum to minimize friction. I suspect an ultra-high speed disk drive would need to do the same. As long as the vacuum is maintained, they would probably be pretty quiet. The motors would tend to give off a ~1200 Hz (e.g. 72K cycle per minute) whine.

What's an order of magnitude among friends?

Posted Feb 2, 2013 18:25 UTC (Sat) by nix (subscriber, #2304) [Link]

The motors would tend to give off a ~1200 Hz (e.g. 72K cycle per minute) whine.

Right around the human hearing frequency optimum, and right in the range critical for comprehension of human speech.

What's an order of magnitude among friends?

Posted Feb 2, 2013 18:27 UTC (Sat) by Tet (guest, #5433) [Link] (1 responses)

they can go above 100K RPM- without exploding. The thing they have in common is that they operate under vacuum to minimize friction

You don't even need that. Most turbos routinely spin in excess of 100K RPM without needing a vacuum, and the Honda CX-500 had a turbo that spun at 200K RPM. Pretty insane, but it worked.

What's an order of magnitude among friends?

Posted Feb 8, 2013 0:16 UTC (Fri) by dfsmith (guest, #20302) [Link]

Did it work at 100kRPM for 5 years continuously?

What's an order of magnitude among friends?

Posted Feb 13, 2013 22:17 UTC (Wed) by Lennie (subscriber, #49641) [Link] (8 responses)

Western Digital says they are going to release Helium-based products 'soon'.

So this isn't an alien concept to producers of harddisks.

What's an order of magnitude among friends?

Posted Feb 14, 2013 17:58 UTC (Thu) by nix (subscriber, #2304) [Link] (7 responses)

Wow. How will they stop it leaking out in short order? (Hey, maybe they prefer that: if it leaks out in three years, you'll be *forced* to buy a new disk!)

What's an order of magnitude among friends?

Posted Feb 14, 2013 19:49 UTC (Thu) by Lennie (subscriber, #49641) [Link] (5 responses)

Arstechnica has an article with information about that:

http://arstechnica.com/information-technology/2012/09/hel...

The register mentions some more about the advantages and specifications:

http://www.theregister.co.uk/2013/01/08/wd_helium/

What's an order of magnitude among friends?

Posted Feb 15, 2013 18:13 UTC (Fri) by nix (subscriber, #2304) [Link] (4 responses)

Hm, they're saying it would *lengthen* drive lifespan. Obviously they think containment isn't that problematic, or is a solved problem, in which case I'm babbling about nothing (as usual).

What's an order of magnitude among friends?

Posted Feb 20, 2013 0:49 UTC (Wed) by man_ls (guest, #15091) [Link] (3 responses)

Helium is the second most inert of all gases so it is not hard to contain. Hydrogen (as said below) or even liquid helium (as you can read on the wikipedia) are harder. With helium gas you just seal the container and it just stays there, without condensing or interacting with the walls or other gases; it shouldn't be harder than creating a vacuum.

What's an order of magnitude among friends?

Posted Feb 21, 2013 21:11 UTC (Thu) by mikewd (subscriber, #46016) [Link] (2 responses)

"Helium is the second most inert of all gases so it is not hard to contain. Hydrogen (as said below) or even liquid helium (as you can read on the wikipedia) are harder. With helium gas you just seal the container and it just stays there, without condensing or interacting with the walls or other gases; it shouldn't be harder than creating a vacuum."

The problem is that the diffusion rate of helium through glass or epoxy and other glues and insulators can be quite high at room temperature (as a low temperature physicist well knows). So the container needs to be all metal with a soldered or welded seal.

Mike

What's an order of magnitude among friends?

Posted Feb 22, 2013 10:28 UTC (Fri) by khim (subscriber, #9252) [Link] (1 responses)

So the container needs to be all metal with a soldered or welded seal.

Well, we are talking HDDs here. They used "all metal" containers for decades now anyway. Usually they had some filters and were not welded shut, but that is minor change IMO.

What's an order of magnitude among friends?

Posted Feb 22, 2013 18:12 UTC (Fri) by magila (guest, #49627) [Link]

There are many openings in a hard drive which are currently sealed with epoxy or plastic/foil adhesives. The biggest is the seal between the top cover and the drive body but there are also many holes used during manufacturing. If you examine a hard drive's casing you can see where these have been sealed with circular stickers. Less visible are the openings for the spindle motor and servo/channel control lines.

Making all of these openings impermeable to He is not a minor change and will likely significantly increase the cost of the drive.

What's an order of magnitude among friends?

Posted Feb 14, 2013 23:43 UTC (Thu) by dlang (guest, #313) [Link]

Helium isn't that bad to contain. You are thinking of how hard it is to contain Hydrogen.

What's an order of magnitude among friends?

Posted Feb 1, 2013 16:49 UTC (Fri) by felixfix (subscriber, #242) [Link]

I "worked" (for various starving student hobby definitions of "work") on a Univac SS-90 which had a 17,000 RPM drum for main storage (50,000 decimal digits!). It took 12 minutes to spin up, and the manual warned that once power was turned off and it started spinning down, you had to wait an hour before turning power back on. (All from memory.) We often wondered what would happen when the bearings seized up; would the entire mainframe (probably a ton or two) twist or thump? One of us later was using it when it did fail, but it failed slowly, just being harder to start or slower to spin up, I forget now, and one day just wouldn't start.

What's an order of magnitude among friends?

Posted Feb 4, 2013 9:45 UTC (Mon) by pr1268 (guest, #24648) [Link] (1 responses)

Indeed. The (now retired) NASA Space Shuttle Main Engine high-pressure oxidizer turbopump spun at 28,120 RPM, and the H.P. fuel turbopump spun at 35,360 RPM. Each of these was around 10 times the dimensions of a typical computer spinning hard drive. Of course, there isn't an actuator arm with a magnetic pickup anywhere near those pumps!

Back to the article, while I do admire Daniel's work on Tux3, I'm wondering if the fast I/O speeds of SSDs will render all the optimization work done to FSCK a moot point?

What's an order of magnitude among friends?

Posted Feb 8, 2013 0:51 UTC (Fri) by daniel (guest, #3181) [Link]

Probably. But not yet. This is just something we need to do because spinning media will still be filling important (though increasingly specialized) roles for at least the next ten years. There are definitely more exciting programming projects around, but without taking care of this Tux3 would be pretty lame.

What's an order of magnitude among friends?

Posted Feb 4, 2013 13:36 UTC (Mon) by man_ls (guest, #15091) [Link]

Well, a 72 KRPM disk might double as a flywheel energy storage and take the place of the UPS too. Handling it in operation would certainly be awkward due to the centrifugal forces inside it -- if you have ever handled a Powerball (rotating gyroscope toy) you know what I mean.

What's an order of magnitude among friends?

Posted Jan 31, 2013 9:58 UTC (Thu) by micka (subscriber, #38720) [Link] (8 responses)

72,000 rpm isn't that fast ! It's just a bit over 1 rotation per second !
Besides, you don't need to write non-significant zeroes !

What's an order of magnitude among friends?

Posted Jan 31, 2013 11:05 UTC (Thu) by mjthayer (guest, #39183) [Link] (3 responses)

micka wrote:
> 72,000 rpm isn't that fast ! It's just a bit over 1 rotation per second !
Besides, you don't need to write non-significant zeroes !
Clarification for monolingual readers: in many languages "." and "," in numbers are reversed relative to their use in English.

What's an order of magnitude among friends?

Posted Jan 31, 2013 12:17 UTC (Thu) by micka (subscriber, #38720) [Link] (2 responses)

Actually, we use "," for comma separator, but we don't use "." for thousands separator. Either we use no thousand separator (for 4 digits) or we use... space (thin insecable space). At least this one is non ambiguous.

What's an order of magnitude among friends?

Posted Jan 31, 2013 12:25 UTC (Thu) by mjthayer (guest, #39183) [Link] (1 responses)

micka wrote:
> Actually, we use "," for comma separator, but we don't use "." for thousands separator.
Yes, and I see I have badly oversimplified[1].

[1] http://docs.oracle.com/cd/E19455-01/806-0169/overview-9/i...

What's an order of magnitude among friends?

Posted Jan 31, 2013 12:44 UTC (Thu) by micka (subscriber, #38720) [Link]

I've even seen other variants, with the use of the dot not on the base line but higher, in the middle of the line, or the use of the apostrophe as thousand separator.

Actually, I use comma as decimal separator but have no problem reading numbers that use dot instead.
On the other hand, I'm lost as soon as some one uses the comma as thousand separator (and when it's mixed with dots, I crash).

What's an order of magnitude among friends?

Posted Jan 31, 2013 14:22 UTC (Thu) by intgr (subscriber, #39733) [Link] (3 responses)

> 72,000 rpm isn't that fast ! It's just a bit over 1 rotation per second !

You mean 1 rotation per millisecond, right?

What's an order of magnitude among friends?

Posted Jan 31, 2013 19:51 UTC (Thu) by dlang (guest, #313) [Link] (2 responses)

no, he's pretending that the comma is a decimal point, which is something that is used in print in parts of Europe.

What's an order of magnitude among friends?

Posted Jan 31, 2013 20:03 UTC (Thu) by intgr (subscriber, #39733) [Link]

D'oh, suddenly that whole thread started to make sense (and I'm European!). :)

What's an order of magnitude among friends?

Posted Jan 31, 2013 21:30 UTC (Thu) by micka (subscriber, #38720) [Link]

Not a decimal point, a decimal comma ;)

> that is used in print

and on screen and in handwriting too.

Yes, that was an uncalled for (and a bit lame) jocke.

What's an order of magnitude among friends?

Posted Jan 31, 2013 23:30 UTC (Thu) by daniel (guest, #3181) [Link] (2 responses)

Thanks Jon, it's what I meant of course. Sigh. Should have done that third proofread...

Mind you, where _is_ my 72K rpm disk? Which might not eat itself for breakfast like a typical flash disk.

What's an order of magnitude among friends?

Posted Jan 31, 2013 23:44 UTC (Thu) by apoelstra (subscriber, #75205) [Link] (1 responses)

> Mind you, where _is_ my 72K rpm disk? Which might not eat itself for breakfast like a typical flash disk.

Are you suggesting that hard drives fail better than flash? In my experience, the failure mode of spinning disk drives tends to be a grinding scream of information being extinguished; the failure mode of flash is the hardware becoming read-only.

Not that I'm disagreeing with you that faster drives would be nice..

What's an order of magnitude among friends?

Posted Feb 1, 2013 0:37 UTC (Fri) by dlang (guest, #313) [Link]

in theory you are right, but in practice I've had too many SSDs and flash devices become completely unreadable to put any faith into that (and I've had enough spinning rust drives that have failed that I'be been able to recover data from to not agree with the other part of the statement.

They both fail in both final and non-final ways

Quotes of the week

Posted Jan 31, 2013 11:34 UTC (Thu) by epa (subscriber, #39769) [Link] (5 responses)

Disk size increases quadratically but the number of files on disk surely does not. More likely there is a linear growth in the number of files and linear growth of their average size. So the complexity of the filesystem is only linear - but the problem is that filesystem structures tend to be scattered over the disk rather than concentrated in one place.

If the size of an allocation unit increases as disk capacity increases then the number of bits to address the disk remains constant. (For example if the filesystem uses 32 bits to identify locations on disk, then a four terabyte disk will have an allocation unit of about one kilobyte, and bigger disks will have bigger chunks.) So the size of the filesystem metadata should grow only with the number of files, provided you're willing to accept wasting more space as the filesystem gets bigger. (If the average file size also grows in proportion with the disk size, then the proportion of disk space wasted on rounding up to a whole allocation unit remains constant.)

So if the filesystem were a bit simpler, keeping all its metadata in a fixed block at the start of the disk and perhaps using larger allocation units for larger filesystems, the time taken to fsck should increase only linearly with the number of files.

Quotes of the week

Posted Jan 31, 2013 18:57 UTC (Thu) by niner (subscriber, #26151) [Link] (1 responses)

Such filesystems would also be painfully slow, so you just traded fs runtime performance for fsck performance. Doesn't sound like a very good deal to me.

Quotes of the week

Posted Jan 31, 2013 20:16 UTC (Thu) by daniel (guest, #3181) [Link]

And apropos to that...

https://www.usenix.org/conference/fast13/birds-feather-se...

Everybody welcome. Incremental and online fsck on the agenda, both for Tux3 specifically and filesystems in general. Contact me for details on attendance if interested, or drop by oftc.net #tux3.

Quotes of the week

Posted Jan 31, 2013 19:54 UTC (Thu) by dlang (guest, #313) [Link] (2 responses)

Actually, lots of files are not getting bigger, while some files are getting huge.

I am in the process of copying data from my home file server, the first ~550K files were about 33G, the next 80K files are the remaining 2.4TB (and it's probably <10K files that account for all of the space)

so if you get inefficient at storing small files, it can have a surprisingly large effect.

Quotes of the week

Posted Jan 31, 2013 21:58 UTC (Thu) by khim (subscriber, #9252) [Link] (1 responses)

so if you get inefficient at storing small files, it can have a surprisingly large effect.

Why? Sure, efficiency for these files will go down, but efficiency for all files will still be pretty good.

Quotes of the week

Posted Jan 31, 2013 22:21 UTC (Thu) by dlang (guest, #313) [Link]

it depends how you measure efficancy.

if you have 500K tiny files take an extra meg of disk space, but save a meg of disk space on 2K giant files, your overall efficiency is not improved

extents work pretty well at consolidating disk allocation units, so the win on modern filesystems would not be very large.