The sun4m architecture is dead dead dead. You're talking about machines that last shipped 10 years ago. The u1 had very serious and real issues operating in 64-bit mode, that were well-known back in Solaris 8.
They didn't just sort of stop supporting the stuff, they dropped 32-bit support entirely from the S10 kernel. That means the sun4m, sun4c, and early edition of ultrasparc machines are impossible to use. It also means no expensive engineer time spent trying to deal with 32/64 bit issues in existing code.
Good.
I'd rather Sun focus on hardware I actually want to run, not crap I find in the back of my closet.
(TEN... YEARS... that's back to 486 days in the Intel world)
Let them go peacefully into their putty-colored night, for they are no more. Or run OpenBSD on them. OpenBSD loves the 4m.
There are two types of packages on Solaris. "datastream" packages are a single file, and are installed with
pkgadd -d "package"
packages may also be a directory named something like FOOpkg, which are installed as so:
pkgadd -d/path/to/parentdir FOOpkg
The confusing part, of course, is that many people name their datastream packages FOOpkg, making the distinction less than obvious. Datastreams may also contain more than one package, and individual packages may be installed from a datastream by naming them on the command line.
Also, datastream and directory package types can be turned into each other with "pkgtrans". Enjoy.
With the BSD license, code tends to slowly drift into closed projects
Man your ass must really hurt,
as you seem to be pulling some huge generalizations out of there.
Do you have any supporting evidence for that claim? Just a vague feeling that the wicked companies are slowly dragging all that innocent BSD licensed source into their secret dungeons, never to be seen again?
The BSD license is for people that want to have a one-off license solution for a project that they've produced.
What does that mean? How many licenses does a project need? I think one will suffice.
It's not just about optimizing code, it's about understanding what the code is doing.
Whether you wrte the code or not (or even have source), sometimes the results are complicated. Why is this app so slow?
Is it disk I/O wait? You can check that. Is the dumbass programmer opening thousands of little files and throwing them away? Sure, you can check that. Are you paging, but only when some other app starts grinding away?
It's *not* a tool to optimize code. It's a tool to understand the behavior and problems with running programs. The fact that you can use that info to optimize your application code is just a bonus.
These functions can get as complex as you'd like, since DTrace is driven by the 'D' programming language.
Here's a few good examples showing the use of speculations (only tell me when a syscall *doesn't* work, not when it does) and aggregations (summarize the data, so you don't have to).
This sort of built in functionality is what separate DTrace from tools like truss, not to mention the huge number of OS hooks those other tools can't touch.
(Disclaimer: I'm older than I look, just check my Queen and Zeppelin 8-tracks. No, I'm not kidding.)
I'm not saying the recording industry was gung-ho for DAT (in consumer's hands, anyway), but I don't think they killed it singlehandedly, and I really don't see their fingerprints on DDS.
DAT didn't meet it's consumer niche, and that's about all. It was great technology, ahead of its time and well ahead of the price curve. Audiophiles were certainly interested, and were the only consumers. I certainly remember the introduction of DAT decks, and I remember seeing them in consumer hands. In fact, they're still available, but have never come down in price.
The same thing happened with DCC, and to some extent with MiniDisc. The Japanese MiniDisc market is huge, and is the only reason they still trickle over here. (I have a MiniDisc portable I bought a few years ago, when MP3 players were 64/128 MB, and still love it). The technology was not good enough to make it worth upgrading, and labels were slow to commit. Combine that with the heavy format competition (DAT vs DCC vs CD vs MiniDisc), and it's no surprise that only one technology hit the sweet spot of mass consumer acceptance.
(I also remember recieving software releases on DDS-2, so there was some fight in the beast).
I greatly wish I could afford a decent backup solution too, but I only blame the RIAA for shitty music and an outdated business model, not the destruction of a technology.
36 GB is pretty small for a primary backup solution, but DDS-x has never been a top of the line technology. Even back in the 90s, DDS was kid stuff compared to DLT (in quality and price).
I seriously doubt that the RIAA had any real influence over the wide use of DAT drives, and certainly not in the backup field (where DAT = DDS-1, about 1GB per tape). The RIAA is guilty of plenty, but don't hang this one on them.
The problem with DAT was being the digital leader before anyone understood the concept of digital music. High prices, low availability of pre-recorded music, the notion that cassettes were "good enough". The only place DAT found any real saturation was amongst Grateful Dead tapers, and even then it was a pricy event.
And yet I used DDS-2,3,4 tapes on small servers for years and years, and they continue to be available and practical.
Good backup solutions are beyond the reach of most home users (and apparently most Slashdotters at work, which is sad).
Part of the problem is demand, as many home users never do backups at all, and thus aren't in the market for a backup system. This will keep prices high, since only companies who really care about their data will buy the expensive systems.
Sadly, if you want to back 250GB for $200, you're screwed.
Even though I insist on excellent tape systems at work, I can't afford them at home. I burn DVDs, and I leave lots of stuff of them. MP3s? don't back those up. Baby pictures? Back em up.
As a home user, it's all about what your data is worth to you, and it's not all worth the same cash per GB.
It's all stone knives and bearskins at home, and some folks never even get to see the shiny stuff.
Call the vendor's main sales line from your cell phone. Hell, even do it from home.
Ask them for the retail price on the unit you want. This will in *all* cases be more than your company will actually end up spending. When they start talking about discounts, sales reps, whatever, explain that you need baseline pricing for budget reasons, and you just want the street retail price.
Use that as your base pricing information, and you won't have to go back for more.
Make sure to show that your system comes at less than this year's maintenance...
We had a similar problem with a large ADIC unit, where the thing was constantly losing vertical alignment, which ended up panicking the robot and shutting itself down.
Our assigned vendor monkeys adjusted it over and over again, and it continued to fail. After some significant pressure, they flew in a *good* tech, who traced the problem to a 39-cent bolt that was partly stripped, letting the arm to travel a fraction of an inch too far down. After X moves, it was way out of alignment.
The moral is if you've replaced 20 robot arms in one silo, and not in another, the techs aren't fixing the real problem.
Tell StorageTek it's time to stop fucking around and fix it for real, or else replace it.
If you create 250 GB of new data nightly, I'd hope it's worth more than $200 to you.
I'd hope that 250 GB of data is worth more than $200 to you anyway, but maybe not.
A used DLT IV or DDS-4 system might work, but you're looking at 6-10 tapes for a full backup. How much of that 250GB changes nightly? How much can be recovered by re-ripping CDs?
Your choices are really either spend more or back up less...
DAT has long since changed names to DDS (Digital Data Storage), and is still available.
The current rev is DDS-5, at 36GB per tape.
As for Joe, he's still going to be forking over a bunch of cash because no DAT/DDS solution was ever *cheap*, but for small systems, they're much cheaper than LTO/SDLT/AIT.
Furthermore, DAT Audio components are still available, although CD burners have long since eclipsed them for most people.
LTO-2 is 200/400GB, and can be easily found for ~$65 each.
They have a larger form factor than AIT tapes, but most of the other details are the same (little to no cleaning, long life cycles, etc.)
When I was using AIT-2 tapes before this job, my biggest issue was the Sony drives, which had a substantial failure rate. Hopefully they're better now, because we were always replacing drives.
The most impressive thing about the Scalar 1000 isn't the storage space, it's the wasted space in that big honkin' refrigerator cabinet.
Having a glass front looks really cool, but it's a huge waste of space compared to a circular robot design. The back side is all wasted space as well. We pondered mounting a bunch of Sun X1s in the back of our ADIC...
(but your point about tape not going away is completely valid.)
This all may be true, but reinforces his point about customer service.
Best Buy has chosen to make their customer service reps underpaid flailers who have no authority to do *anything* for you. The poor girl is just a face to stand there and recite corporate policy. The "manager" is little better.
If Best Buy (or any store) gave a fuck about customer service, they wouldn't treat it like the bottom-line-sucking expense that they do.
Now everyone on his ADSL provider can send email from his domain as well. Great protection there.
How long before spammers start harvesting SPF records looking for easy targets?
SPF won't protect anyone but large companies with easily-recognizable, often-spoofed names (uh HOTMAIL, YAHOO).
Meanwhile, they make things harder for everyone else, while not actually protecting against spam. All I see in this thread is "Change your service provider!", "Rewrite your mailing list software!", "*You* change this, *You* change that".
Fuck that.
How about not breaking widely used protocols for meaningless bullshit. SMTP is totally b0rked, and everybody knows it, but a half-assed patch is not the answer.
No, Stephenson breaks down on simple storytelling gaffes, rather than computers.
Semi-spoiler below, but the damn book is 5 years old now. Statute of Limitations, or something...
Cryptonomicon began to suck for me when the EMP pulse went off. Sure, this thing knocks out everything in the area, all the chips are fried, man fried. His car doesn't work, his laptop is toast...
And yet he pops the drive out of the laptop and pockets it, then puts it in another machine later! Couldn't Stephenson turn a hard drive over and see all those icky little chips on the back? How were those saved? (Never mind the EMP pulse and the platters, I'll give him that as a gift...)
The entire EMP pulse segment was just a hack writers device to keep Waterhouse from looking at the codes until after he'd met Root on the plane, and required another stupid deus ex machina in the electromagnet doorframe.
All when simply saying "at this moment, his battery died" would have accomplished the same thing.
Slashdot Mirror
The sun4m architecture is dead dead dead. You're talking about machines that last shipped 10 years ago.
The u1 had very serious and real issues operating in 64-bit mode, that were well-known back in Solaris 8.
They didn't just sort of stop supporting the stuff, they dropped 32-bit support entirely from the S10 kernel.
That means the sun4m, sun4c, and early edition of ultrasparc machines are impossible to use. It also means
no expensive engineer time spent trying to deal with 32/64 bit issues in existing code.
Good.
I'd rather Sun focus on hardware I actually want to run, not crap I find in the back of my closet.
(TEN... YEARS... that's back to 486 days in the Intel world)
Let them go peacefully into their putty-colored night, for they are no more.
Or run OpenBSD on them. OpenBSD loves the 4m.
There are two types of packages on Solaris. "datastream" packages are a single file, and are installed with
/path/to/parentdir FOOpkg
pkgadd -d "package"
packages may also be a directory named something like FOOpkg, which are installed as so:
pkgadd -d
The confusing part, of course, is that many people name their datastream packages FOOpkg, making the distinction less than obvious. Datastreams may also contain more than one package, and individual packages may be installed from a datastream by naming them on the command line.
Also, datastream and directory package types can be turned into each other with "pkgtrans". Enjoy.
Well, we can start by shaving your back, Svetlana...
I'll take "Famous Titties" for 100.
That's "Titles", not "Titties"...
With the BSD license, code tends to slowly drift into closed projects
Man your ass must really hurt, as you seem to be pulling some huge generalizations out of there.
Do you have any supporting evidence for that claim? Just a vague feeling that the wicked companies are slowly dragging all that innocent BSD licensed source into their secret dungeons, never to be seen again?
The BSD license is for people that want to have a one-off license solution for a project that they've produced.
What does that mean? How many licenses does a project need? I think one will suffice.
In the world of guitar amps, where at least part of the signal path is often overdriven (nice crunchy sound) your point is completely accurate.
You could have eliminated this confusion by perhaps inlcuding the word "guitar" somewhere in your post.
It's not just about optimizing code, it's about understanding what the code is doing.
Whether you wrte the code or not (or even have source), sometimes the results are complicated. Why is this app so slow?
Is it disk I/O wait? You can check that. Is the dumbass programmer opening thousands of little files and throwing them away? Sure, you can check that. Are you paging, but only when some other app starts grinding away?
It's *not* a tool to optimize code. It's a tool to understand the behavior and problems with running programs. The fact that you can use that info to optimize your application code is just a bonus.
No GUI. (yet)
a ce_for_developers
These functions can get as complex as you'd like, since DTrace is driven by the 'D' programming language.
Here's a few good examples showing the use of speculations (only tell me when a syscall *doesn't* work, not when it does) and aggregations (summarize the data, so you don't have to).
http://blogs.sun.com/roller/page/ahl/20040701#dtr
This sort of built in functionality is what separate DTrace from tools like truss, not to mention the huge number of OS hooks those other tools can't touch.
That's the sort of innovation that comes from typing quickly because you're late for work, but in the middle of a post.
I happily release 'progasmming' and all the other typos in that post into the public domain for unrestricted use.
DTrace is not the SE toolkit. The SE toolkit used a similar method (little langugae scripts) to collect data that was made public by the kernel.
DTrace takes advantage of monitoring hooks placed throughout the kernel and in every system call, and it does things that SE never could.
Furthermore, one look at the DTrace manual and your boss will shit his pants. This is not PHB material by any means.
Chapter 1: The D prograsmming Language...
(Disclaimer: I'm older than I look, just check my Queen and Zeppelin 8-tracks. No, I'm not kidding.)
I'm not saying the recording industry was gung-ho for DAT (in consumer's hands, anyway), but I don't think they killed it singlehandedly, and I really don't see their fingerprints on DDS.
DAT didn't meet it's consumer niche, and that's about all. It was great technology, ahead of its time and well ahead of the price curve. Audiophiles were certainly interested, and were the only consumers. I certainly remember the introduction of DAT decks, and I remember seeing them in consumer hands. In fact, they're still available, but have never come down in price.
The same thing happened with DCC, and to some extent with MiniDisc. The Japanese MiniDisc market is huge, and is the only reason they still trickle over here. (I have a MiniDisc portable I bought a few years ago, when MP3 players were 64/128 MB, and still love it). The technology was not good enough to make it worth upgrading, and labels were slow to commit. Combine that with the heavy format competition (DAT vs DCC vs CD vs MiniDisc), and it's no surprise that only one technology hit the sweet spot of mass consumer acceptance.
(I also remember recieving software releases on DDS-2, so there was some fight in the beast).
I greatly wish I could afford a decent backup solution too, but I only blame the RIAA for shitty music and an outdated business model, not the destruction of a technology.
In short, shop EBay and fuck the RIAA.
36 GB is pretty small for a primary backup solution, but DDS-x has never been a top of the line technology. Even back in the 90s, DDS was kid stuff compared to DLT (in quality and price).
I seriously doubt that the RIAA had any real influence over the wide use of DAT drives, and certainly not in the backup field (where DAT = DDS-1, about 1GB per tape). The RIAA is guilty of plenty, but don't hang this one on them.
The problem with DAT was being the digital leader before anyone understood the concept of digital music. High prices, low availability of pre-recorded music, the notion that cassettes were "good enough". The only place DAT found any real saturation was amongst Grateful Dead tapers, and even then it was a pricy event.
And yet I used DDS-2,3,4 tapes on small servers for years and years, and they continue to be available and practical.
Good backup solutions are beyond the reach of most home users (and apparently most Slashdotters at work, which is sad).
Part of the problem is demand, as many home users never do backups at all, and thus aren't in the market for a backup system. This will keep prices high, since only companies who really care about their data will buy the expensive systems.
Sadly, if you want to back 250GB for $200, you're screwed.
Even though I insist on excellent tape systems at work, I can't afford them at home. I burn DVDs, and I leave lots of stuff of them. MP3s? don't back those up. Baby pictures? Back em up.
As a home user, it's all about what your data is worth to you, and it's not all worth the same cash per GB.
It's all stone knives and bearskins at home, and some folks never even get to see the shiny stuff.
Call the vendor's main sales line from your cell phone. Hell, even do it from home.
Ask them for the retail price on the unit you want.
This will in *all* cases be more than your company will actually end up spending. When they start talking about discounts, sales reps, whatever, explain that you need baseline pricing for budget reasons, and you just want the street retail price.
Use that as your base pricing information, and you won't have to go back for more.
Make sure to show that your system comes at less than this year's maintenance...
Your vendor techs are lazy or stupid.
We had a similar problem with a large ADIC unit, where the thing was constantly losing vertical alignment, which ended up panicking the robot and shutting itself down.
Our assigned vendor monkeys adjusted it over and over again, and it continued to fail. After some significant pressure, they flew in a *good* tech, who traced the problem to a 39-cent bolt that was partly stripped, letting the arm to travel a fraction of an inch too far down. After X moves, it was way out of alignment.
The moral is if you've replaced 20 robot arms in one silo, and not in another, the techs aren't fixing the real problem.
Tell StorageTek it's time to stop fucking around and fix it for real, or else replace it.
If you create 250 GB of new data nightly, I'd hope it's worth more than $200 to you.
I'd hope that 250 GB of data is worth more than $200 to you anyway, but maybe not.
A used DLT IV or DDS-4 system might work, but you're looking at 6-10 tapes for a full backup. How much of that 250GB changes nightly? How much can be recovered by re-ripping CDs?
Your choices are really either spend more or back up less...
What are you talking about?
DAT has long since changed names to DDS (Digital Data Storage), and is still available.
The current rev is DDS-5, at 36GB per tape.
As for Joe, he's still going to be forking over a bunch of cash because no DAT/DDS solution was ever *cheap*, but for small systems, they're much cheaper than LTO/SDLT/AIT.
Furthermore, DAT Audio components are still available, although CD burners have long since eclipsed them for most people.
Amen, brother.
Anybody who says "tape is going away" has never worked in a regulated industry.
Stop working for idiots.
It's much less stressful.
LTO-2 is 200/400GB, and can be easily found for ~$65 each.
They have a larger form factor than AIT tapes, but most of the other details are the same (little to no cleaning, long life cycles, etc.)
When I was using AIT-2 tapes before this job, my biggest issue was the Sony drives, which had a substantial failure rate. Hopefully they're better now, because we were always replacing drives.
The most impressive thing about the Scalar 1000 isn't the storage space, it's the wasted space in that big honkin' refrigerator cabinet.
Having a glass front looks really cool, but it's a huge waste of space compared to a circular robot design. The back side is all wasted space as well. We pondered mounting a bunch of Sun X1s in the back of our ADIC...
(but your point about tape not going away is completely valid.)
This all may be true, but reinforces his point about customer service.
Best Buy has chosen to make their customer service reps underpaid flailers who have no authority to do *anything* for you. The poor girl is just a face to stand there and recite corporate policy. The "manager" is little better.
If Best Buy (or any store) gave a fuck about customer service, they wouldn't treat it like the bottom-line-sucking expense that they do.
And quality control ranks among the most important.
That's the elephant in this room you were trying to ignore with your second line.
Oh good.
Now everyone on his ADSL provider can send email from his domain as well. Great protection there.
How long before spammers start harvesting SPF records looking for easy targets?
SPF won't protect anyone but large companies with easily-recognizable, often-spoofed names (uh HOTMAIL, YAHOO).
Meanwhile, they make things harder for everyone else, while not actually protecting against spam. All I see in this thread is "Change your service provider!", "Rewrite your mailing list software!", "*You* change this, *You* change that".
Fuck that.
How about not breaking widely used protocols for meaningless bullshit. SMTP is totally b0rked, and everybody knows it, but a half-assed patch is not the answer.
No, Stephenson breaks down on simple storytelling gaffes, rather than computers.
Semi-spoiler below, but the damn book is 5 years old now. Statute of Limitations, or something...
Cryptonomicon began to suck for me when the EMP pulse went off. Sure, this thing knocks out everything in the area, all the chips are fried, man fried. His car doesn't work, his laptop is toast...
And yet he pops the drive out of the laptop and pockets it, then puts it in another machine later! Couldn't Stephenson turn a hard drive over and see all those icky little chips on the back?
How were those saved? (Never mind the EMP pulse and the platters, I'll give him that as a gift...)
The entire EMP pulse segment was just a hack writers device to keep Waterhouse from looking at the codes until after he'd met Root on the plane, and required another stupid deus ex machina in the electromagnet doorframe.
All when simply saying "at this moment, his battery died" would have accomplished the same thing.
Credibility shot to hell...