Ten Years Of The Linux Counter

hta writes "In a testament to the fact that Linux has been around for a while, the Linux Counter turned turned ten last month. The site has been counting a few of the users of Linux since September of 1993, and currently lists more than 130.000 names. It's still waiting for Linus Torvalds to claim registration #1, which has been reserved for him for the last ten years. Among the events that have happened in 10 years of running the counter, the Slashdot events are some of the more memorable."

Comments

Uptime?

[Eric_Cartman_South_P]

I wonder if any Linux computers have 10 years of uptime?!?

Re:Uptime?

[suwain_2]

I doubt it. There are Linux boxes that probably could run 10 years, but frankly, I wouldn't want anything to do with them for a few reasons.

First, although it's not nearly as bad as most Microsoft products, there has certainly been more than one kernel-related exploit in the past 10 years. Anyone who's had ten years of uptime is going to have some massive security holes.

The second problem is hardware. What were you using ten years ago? Frankly, I don't even remember what I used. I'll guess rather blindly and say a 66 MHz 486 was cutting-edge ten years ago. My desktop machine has dual processors, each about 25 times faster. A LOT has happened to hardware over ten years; anything that's run for ten years is probably bordering on obsolete.

Finally... Even if Linux was truly perfect, I doubt you can get hardware to run for ten years. If you're using IDE, you've probably gone through a few disks if you run 24/7; unless it's a massive server, you probably ended up shutting down even if it ran SCSI. And although hardware doesn't ordinarily break down out of the blue, I've had bizarre things happen before; earlier this month I swapped out a network card that seems to have been possessed for no particular reason. And even if the hardware was perfect, I'd like to see you get power to it for ten years straight. Even on an enormous UPS that can run for weeks, you're eventually going to have to change batteries.

The highest uptime I've ever seen was about 101 days; oddly all our changes happen right around then, whether it be moving to an entirely new box, building a new kernel, or (most recently) have the network card start acting all weird right after hitting 100 days uptime. People have gone more than a year, but I really have to question their wisdom.

Re:Uptime?

[Xzzy]

> The highest uptime I've ever seen was about 101
> days; oddly all our changes happen right around
> then,

Not going to start a pissing contest for the trolls and be specific, but I've seen numerous machines that exceed that number, by a lot. It seems dependent on what they do and what kind of load they're put under.

Obviously a good UPS helps too.

In my experience, regular downtime is actually a good thing. In EVERY case where I've dealt with a machine that had a 3 digit uptime, as soon as it reboots and/or stays down for more than a few minutes, hardware dies. Considering such outages are often unplanned (and there's often many machines at once).. it makes a stressful situation much more so.

Frequently and routinely shutting machines down allows one to deal with these failures much more smoothly.

Re:Uptime?

[Anguo]

The longest running servers are not Linux but FreeBSD on Apache...

Linux was still in its infancy back then (being just a teen now), so a linux server running for ten years on ten years old hardware wouldn't give much benefit...

Re:Slashdotted Already!

[caluml]

This image shows a well-configured (but underpowered) machine's response to slashdot:

Valiant shouldering of the load
Trouble from an unexpected source
Slow return to normality
In the first minute after the article went up, people arrived.

Soon, the 16 available processes were all busy running my too-heavy Perl scripts, and the new clients were sending SYN packets and waiting. And they kept on doing it.

In fact, so many were doing this that the kernel wondered if there were SYN flood attacks going on. Go figure...

Shortly after midnight, the /var partition, where the HTTP logfiles go, filled up. Apparently something else required access to that partition too - at least registrations weren't successful either.
Luckily, I was online at the time, deleted a couple of files, and watched the counter come back to normal.
After that, it was plain sailing.

16 available processes? :) Need to upgrade that box, and 256kbit line, methinks!

The important parts:

[caferace]

Part I:

Lessons learned

* Slashdotting is good for you.
The counter is now defensively configured, and is able to do something intelligent even under heavy load.
* A 32-Mbyte Pentium can't fill a 256-Kbit link using Perl.
I'm sure management is happy to know that.
* The adrenaline kick of a slashdotting feels real good!
But it does eat time...I spent 4 hours Tuesday night getting the box reconfigured and back on its feet, and then just watching it. Late.
* On Wednesday, 1412 people registered with the counter.

Part II:

Lessons learned

* Slashdotting is still good for you, as an exercise in work under fire.
The counter's configuration has proved that it stays up and running (as long as there is disk space around).
* A well-handled Slashdot load lasts longer and has more people in it than the load left after the server's been down for a while.
* The adrenaline kick of a slashdotting still feels real good!
* On the 25th and 26th together, 2670 people registered with the counter.

In SCO's view,

[overbyj]

the Linux Counter is a who's who list of those that owe $699.

Re:Any one else notice....

[egburr]

Linux is a *multiuser* operating system. One machine could handle tens or even hundreds or users. You're fussing that the number of users registered exceeds the number of machines registered. You should be fussing that the number of users is so low as to almost equal the number of machines. How many of these machines are effectively just single-user machines?

We can only hope . . .

[Idou]

I signed the online petition that asked SCO to please sue me . . . no response yet.

I guess if even the RIAA and MPAA are having trouble using lawsuits to control BLATANT copyright violations, SCO doesn't have much of a chance in hell suing users on completely baseless grounds.

You know, SCO could have really helped their compaign if they had just released a bit of actually infringing code (not all of it, just a taste so that we all would know they aren't full of BS) . . . the fact that they haven't yet, when it is in their best interest, is either a sign of complete incompetence or the nonexistence of such code.