I'd agree with everyone else and say that you're wagging the dog, but for the fact that several of the systems I've pushed past their rated limits have taken a few days to become stable.
One in particular (2x PPro-180 256k running 233) illustrates the effect. It'd seem fine for the first half hour or so, but I kept getting really odd problems after that. Temps were stable and well within limits, RAM etc tested out elsewhere, so I left it running multiple memtest processes on 512MB for about a month. With no configuration changes, just rebooting it when it died. The first few days it crashed, had processes vanish, etc. every few hours, but after a week or so it'd go longer periods without problems.
It ran the last 12 days or so error free, and has been nearly rock solid (sandstone?) for a year or more now as an everyday workstation. Policy boots keep me from bragging on uptime scores but it's never off. Any problems it's had since have been as easily blamed on multiple causes, and not severe enough to nail down.
So I dunno. It has been said: "Electrons are timid little things, and notional; you have to show them who is in charge." I don't know enough to offer any rational explaination of it; for all I can tell the grooves the bits travel in had to get re-worn... Ya know, like a well banked curve on a back road: when they change the pavement the ruts soon drift to a new spot depending on the new max speed for the curve.
Boy hound don't read that fast, but I think he's got the idea now... He's trembling and backing away from the monitor; so I'd say the reaction is not as enthusiastic amonst their target demographic as they might hype.
Back when, there was the PC. Think VW bug (the older one); simple, effective, minimalist. Then came the XT and the AT, which added a V6 and V8 engine on that bug. I'd say the 386 machines by way of comparison were a V12 or so.
Nowadays, we've got ultra high RPM jet engies; big tires, a towed trailer full of fuel, still welded on to that VW frame that never did get rustproofed.
The 64bit machines will probably continue the tradition. Any really new architechture is going to have to be self-evidently and wildy superior to what exists now to have a hope, and what we've got for all its flaws works pretty well most places.
Find your local RFC mirror. One that has the entire directory, not just rfc*.txt files. Look therein for the iana/assignments/one-letter-domains file; which I will enclose entire here:
The IANA has reserved the one letter and number domains in COM, NET,
EDU and ORG, in case they are needed to support one proposed scheme to
distribute the load on the COM (and other top level domains). It has
not been decided if the scheme these names were reserved for will be
used or not, and it probably won't be decided for several months.
A few of the one letter codes had been allocated under the normal
procedures before the IANA had the idea to reserve all of them. These
allocations were allowed to stand.
There have quite a few inquiries about allocating these one letter
domains to various companies. We haven't been keeping a list of those
that have asked about them. At this point it would be unfair to
notify some people and not others, if and when they are made
available. So, if this ever happens, it will have to be done with a
very public announcement (at least an "Informational" RFC).
splicing and terminating fiber a cheap after a certain point; how am I as an individual to replace the SC connector the puppy just ate off my 1/4 mile of fiber to my home? The connector is about $5, but the tools to apply the connector are over $500.
hard drives capable of 20MB/sec sustained shouldn't be hard to find; I have several. So, in order to get 1,000 frames of 250kb each on disk in a second, we only need 12.5 drives in parallel. Since slower drives are cheaper, let's just go for the case, and put 24 IDE drives behind some nasty little multiplexer/RAID controller deal. After all, portability isn't really the primary goal here...
I started writing that to be a smartass, but having given it a thought I bet it could be done, and cheaply enough to be feasible. Priced against RAM in those quantities, lots of stuff is cheap.
Re:ORBS claim they are OPT-IN also...
on
MAPS Sued Again
·
· Score: 1
I don't care for ORBS testing either, but if those tests alone can down yer box you are in serious need of a new mail server regardless. They do nothing that spammers won't also try on you, as far as I know. In fact, their tests the last time I saw 'em were far more polite than most of the spam relay attempts I've seen; ORBS doesn't attempt delivery to over 2,000 random addresses to mask their activity whilst probing for relays, the spammers do that and worse.
It can be static if you want it to be. When I evaluated mod_ssl and Apache-SSL I was sold by the effort spent on documenting mod_ssl and making it clear and hackable by those who probly aren't qualified to be doing that.
I'm waiting for the apache 1.3.13 release later this week (according to the CVS tarball STATUS from a few days ago); the bit where you can tell it a directory instead of a config file, and it'll parse the files in that directory as that config, that sounds like it'll make some real fun possible.
[ever more offtopic] I saw something in those same files about GDBM, I presume for auth_dbm replacement; I've got a simple hack that does that but not the time to spray it with the substances called for in the official patch submission rituals. Anybody wants it send me an email...
While they've been having their merry contests, the rest of the world has utterly failed to learn the lesson they were trying to promote and now calls what they do "programming".
that nobody has yet mentioned the location of my favorite kernel mirror; http://ftp-mirror.internap.com/. They have other interesting stuff there too, are about to upgrade their servers and are supposed to be looking for suggestions as to what else they mirror.
I'm kinda backwoods in terms of routing distance from anywhere, and that site's always responsive, it seems.
I'm still looking for something to replace my 430HX chipset machines. The PPro FX chipsets weren't bad by comparison to the later stuff, but my BX chipset machines are all non-server worthy.
Look through the older stuff being sold on ebay; the pentium 200's that are going for more than modern machines are often actually worth the money.
The motherboard makers are screwed by the chipset makers; why shouldn't they throw marketing trash like built in audio on; they board's never going to be worthwhile 'cuz the chipset sucks and there's very little they can do about it.
Midiman may make good stuff, but avoid their lower end. The DMAN PCI card is a Crystal 4614 (same thing as in the new IBM A20's) but they've managed to screw it somehow. Mine lived 4 days. I'd never seen a sound card initialization lockup the box at a hardware level before, but Windoze 95, 98, NT4, and Linux all freeze when they touch that damned card now, on multiple motherboards, too.
I'm not terribly familiar with other implementations so I won't speak of those; but in the case of Linux, you're wrong. The NAT that linux 2.2+ does behaves as stated. See the iproute2 IP Command Reference (the link is to the NAT section).
Right; they are not the same thing. Masquerading is closer to what the patent covers than NAT, I think... Oh, and Linux does do real NAT as well as masquerade, I dunno if the BSD's have a masquerade like facility.
The difference is that Masq makes an internal network appear as one address to the world; NAT takes internal addresses and translates them to external addresses one for one.
Eric Raymond and several other bright folks have had a project related to this going for about two years now, called Trove.
Trove is a next-generation Internet software archiving facility, intended to supplant the classical FTP-tree-with-decorations model.
There's even real code available, in Python, which I confess I haven't looked at, so I'm vauge on what it does or soesn't do yet. I suspect there's that which is worth a look.
My favorite source for those numbers is Chris Hare's CPU Electrical specs page. Looks like AMD's 65W maximum power figure is the top right now.
Me, I want simple instructions on how to measure the power dissipation of the CPU's on *my* board. According to the dead 400W power supply, the draw is a bit more than it's supposed to be.
It must be a question of early training. I've got that one and the one that came with the 420's (can't be bothered to go look at the numbers). Can't stand 'em. The layout is squirreled, the feel is wonky; I will admit they're well built.
On topic; I got those, a bunch of old IBM keyboards, and other assorted misc from a school. I could tell from the crud where the keyboard had been used; the students lives as seen from keycrud accumulation are much more interesting than the staff's.
Slashdot Mirror
One in particular (2x PPro-180 256k running 233) illustrates the effect. It'd seem fine for the first half hour or so, but I kept getting really odd problems after that. Temps were stable and well within limits, RAM etc tested out elsewhere, so I left it running multiple memtest processes on 512MB for about a month. With no configuration changes, just rebooting it when it died. The first few days it crashed, had processes vanish, etc. every few hours, but after a week or so it'd go longer periods without problems.
It ran the last 12 days or so error free, and has been nearly rock solid (sandstone?) for a year or more now as an everyday workstation. Policy boots keep me from bragging on uptime scores but it's never off. Any problems it's had since have been as easily blamed on multiple causes, and not severe enough to nail down.
So I dunno. It has been said: "Electrons are timid little things, and notional; you have to show them who is in charge." I don't know enough to offer any rational explaination of it; for all I can tell the grooves the bits travel in had to get re-worn... Ya know, like a well banked curve on a back road: when they change the pavement the ruts soon drift to a new spot depending on the new max speed for the curve.
Note the press quotes page.
Not a guru, but have an opinion:
Back when, there was the PC. Think VW bug (the older one); simple, effective, minimalist. Then came the XT and the AT, which added a V6 and V8 engine on that bug. I'd say the 386 machines by way of comparison were a V12 or so.
Nowadays, we've got ultra high RPM jet engies; big tires, a towed trailer full of fuel, still welded on to that VW frame that never did get rustproofed.
The 64bit machines will probably continue the tradition. Any really new architechture is going to have to be self-evidently and wildy superior to what exists now to have a hope, and what we've got for all its flaws works pretty well most places.
splicing and terminating fiber a cheap after a certain point; how am I as an individual to replace the SC connector the puppy just ate off my 1/4 mile of fiber to my home? The connector is about $5, but the tools to apply the connector are over $500.
I started writing that to be a smartass, but having given it a thought I bet it could be done, and cheaply enough to be feasible. Priced against RAM in those quantities, lots of stuff is cheap.
I don't care for ORBS testing either, but if those tests alone can down yer box you are in serious need of a new mail server regardless. They do nothing that spammers won't also try on you, as far as I know. In fact, their tests the last time I saw 'em were far more polite than most of the spam relay attempts I've seen; ORBS doesn't attempt delivery to over 2,000 random addresses to mask their activity whilst probing for relays, the spammers do that and worse.
Is SMBfs related to samba at all? Even sharing developers doesn't mean that SMBfs and SAMBA are related in any mangerial way.
I'm waiting for the apache 1.3.13 release later this week (according to the CVS tarball STATUS from a few days ago); the bit where you can tell it a directory instead of a config file, and it'll parse the files in that directory as that config, that sounds like it'll make some real fun possible.
[ever more offtopic] I saw something in those same files about GDBM, I presume for auth_dbm replacement; I've got a simple hack that does that but not the time to spray it with the substances called for in the official patch submission rituals. Anybody wants it send me an email...
While they've been having their merry contests, the rest of the world has utterly failed to learn the lesson they were trying to promote and now calls what they do "programming".
I beleive the -pre18's have a workaround included now.
I hereby offer my services as a test rat...
mount, touch, make, date, swapon and swapoff, I'm sure there's more....
I'm kinda backwoods in terms of routing distance from anywhere, and that site's always responsive, it seems.
...but the name of that option should be "CONFIG_AWW_YOU_BASTARD_BIGMEM"
"bzImage" stands for Big Compressed (z) image. The actual compression used is vanilla zlib I think.
Thanks for the info. Saw a new, "open box" DIO 2496 for less than $200 the other day, I knew I should have snapped that up.
Look through the older stuff being sold on ebay; the pentium 200's that are going for more than modern machines are often actually worth the money.
The motherboard makers are screwed by the chipset makers; why shouldn't they throw marketing trash like built in audio on; they board's never going to be worthwhile 'cuz the chipset sucks and there's very little they can do about it.
Care to reccomend decent external DAC/ADC's?
I'm not terribly familiar with other implementations so I won't speak of those; but in the case of Linux, you're wrong. The NAT that linux 2.2+ does behaves as stated. See the iproute2 IP Command Reference (the link is to the NAT section).
The difference is that Masq makes an internal network appear as one address to the world; NAT takes internal addresses and translates them to external addresses one for one.
There's even real code available, in Python, which I confess I haven't looked at, so I'm vauge on what it does or soesn't do yet. I suspect there's that which is worth a look.
Me, I want simple instructions on how to measure the power dissipation of the CPU's on *my* board. According to the dead 400W power supply, the draw is a bit more than it's supposed to be.
On topic; I got those, a bunch of old IBM keyboards, and other assorted misc from a school. I could tell from the crud where the keyboard had been used; the students lives as seen from keycrud accumulation are much more interesting than the staff's.
GNU Zebra.