Dumb things to do with your FreeBSD

[PMc]

You know how you're not supposed to have food or drink in the computer room?

Really?

At one of my customers the computer room was the smokers' refuge, and another it was the beer store (double floor, cooled). Both are successful shops, not sure if due to that.

 

[chrislongros]

Not having system backups .. Checking out to the false git branch, then compiling world and kernel and noticing it after a couple hours

 

[Jose]

At one of my customers the computer room was the smokers' refuge, and another it was the beer store (double floor, cooled). Both are successful shops, not sure if due to that.

It was my nap room at one of my early jobs. But I didn't eat or drink in there.

 

[bgavin]

My Dad had a sign on his gear: "Danger: 1 Million Ohms - Do Not Touch"

 

[PMc]

It was nap room at one of my early jobs. But I didn't eat or drink in there.

Well, I would think it is too noisy and too cold for either recreation. But back in these times it was very much a male refuge, outside of those do's and dont's of ordinary life. That feeling seems to have gone away a lot with computer becoming a commodity. That's one reason why I would so much love to do all these things rather on a proper ship on interplanetary course - that would likely bring back a lot of the good feeling of being concerned with things that do really matter.

 

[lgrant]

Really?

At one of my customers the computer room was the smokers' refuge, and another it was the beer store (double floor, cooled). Both are successful shops, not sure if due to that.

While we are talking about dumb things to do with your raised computer-room floor...

The IBM mainframe shop I worked at in the '70s had IBM 3270 green-screen terminals, and each one required a run of 93-ohm RG62 coaxial cable back to the computer room. In order to plan for expansion, they would run a few extra cables, and they would come out under the computer room floor. So far, so good.

But they didn't label the cables!

So when they needed to add a terminal, they had to run a brand-new cable, because they had hundreds of unused cable ends under the floor, and no way of knowing which one was the other end of the cable they wanted to use.

I always thought they should apply a high-voltage, low-current to the cable, then look under the floor, where it was dark, and see which cable was sparking. But I could never get anyone interested in trying this, and since I was a systems programmer, not a cable installer, they did it the way they did it.

RG62 is rated for a peak voltage of 1.1 KV. If what I am reading is correct to, get an arc in air requires about 3.4 KV per millimeter, more than three times the operating voltage of the cable. Perhaps it would be better to gently spritz the ends of the cables with water and see which one arcs.

Or they could just label the cables when they installed them.

But although the folks running the shop were usually very sharp, occasionally they weren't. For example, the IBM mainframes back then ran on 400 Hz, 3-phase power. They had these devices called Franklin Power Converters that converted the 60 Hz power to 400 Hz. They had a battery input, sort of like a mainframe UPS, but the company didn't buy batteries, because they were too expensive. Now, that could be a valid business decision, except for one thing: back then, if the IBM mainframes lost power, something would get damaged, and you had to get an IBM engineer out to fix it. So losing power was a non-trivial thing, and could delay getting the business going again. (Later IBM mainframes were much more robust; one time a broken pipe dumped a bunch of water into the mainframe. They just turned it off, let it dry out, and wrung out the filters, and it was fine. But the early ones were pretty touchy.)

In any case, the data center I worked in was torn down and turned into a parking lot about 40 years ago, so whatever good and bad ideas they had are of no consequence, except for being interesting memories.

 

[lgrant]

Well, I would think it is too noisy and too cold for either recreation. But back in these times it was very much a male refuge, outside of those do's and dont's of ordinary life. That feeling seems to have gone away a lot with computer becoming a commodity. That's one reason why I would so much love to do all these things rather on a proper ship on interplanetary course - that would likely bring back a lot of the good feeling of being concerned with things that do really matter.

A funny thing about early computer rooms. IBM used to run ads, and they frequently showed female computer operators sitting at the console, wearing very short skirts. Now, women wearing very short skirts at work in the '70s was not at all unusual (and I worked at stodgy General Motors, so if they were doing it there, they were probably doing it everywhere).

But one place women did not wear short skirts ever was in the computer room, because it was *cold*. We had several female operators where I worked, and they all wore long pants.

(Another interesting thing: there is nothing that seems as quiet as a computer room during a power failure. Normally you have all kinds of equipment making noise, and although the ambient noise level is pretty high, you get used to it and don't notice it. But when it all suddenly stops, it is breathtaking.)

 

[Jose]

Well, I would think it is too noisy and too cold for either recreation.

Not if you're young and hung over enough. Also, obsolete computer manuals make decent blankets if you open them in half and place them all over yourself. Good times.

The IBM mainframe shop I worked at in the '70s had IBM 3270 green-screen terminals, and each one required a run of 93-ohm RG62 coaxial cable back to the computer room. In order to plan for expansion, they would run a few extra cables, and they would come out under the computer room floor. So far, so good.

But they didn't label the cables!

So when they needed to add a terminal, they had to run a brand-new cable, because they had hundreds of unused cable ends under the floor, and no way of knowing which one was the other end of the cable they wanted to use.

No tone generators back then? One of my first jobs was as a phone man working on analog PBXes. We were a fly-by-night operation, so I had a toner, but no wand. I had to ground one terminal of the buttset, grab the other in my hand, and run my finger down the type 66 blocks to find the line I'd thrown a tone on. Nice thing was, analog ring voltage is a cool 100 volts. It definitely made an impression when you hit a line that was ringing. Good times.

I splurged on one of these a few years ago, when I could afford it. Hey! I trace a cable at least once a year! Totally worth it. (Maybe it's more like once every two years.)

There were not one but two dinosaur pens at a later (and much better) job. My kingdom was the smaller one, given the creative name of "backup computer room." Unfortunately they had skimped on the A/C setup, probably because of the "backup" in the name. The flaky units kept the room plenty cold, but would leak condensation water under the floor. Yeah, where all the electrical cables were. Maybe it was a fuse panel stress test. Good times.

 

[lgrant]

That's a good question. Would a toner work with shielded cable (i.e., coax)?

Maybe they didn't think of that, just like they didn't think of labeling the cables.

 

[Jose]

That's a good question. Would a toner work with shielded cable (i.e., coax)?

I just installed a TV antenna on my roof, and my toner worked over that coax. I believe it's RG-62 also?

Maybe they didn't think of that, just like they didn't think of labeling the cables.

It was probably incompetence. Though the inductive amplifiers (wands) were super expensive when I was doing this kind of work in the early '90s. Maybe they had a cheap boss like I did? I'm guessing the customer, and not the installer paid for the additional wires.

 

[DonK]

But one place women did not wear short skirts ever was in the computer room, because it was *cold*. We had several female operators where I worked, and they all wore long pants.

This is the Sun mainframe room circa 2001 at a mine in Wyoming. It was kept uncomfortably cold during Summer. And its cooler kept running during the Winter, when outside temperatures often dropped to below zero. You had to wear outdoor clothing in it, including warm gloves.

I tried my best to always use a terminal elsewhere in the building to avoid the room at all costs. Unfortunately, as shown in the photo, the console sat in the room. And to the left of the console sat a pair of stacked SCSI storage enclosures.

Typically they called me out to the site in the middle of Winter to replace a crashed drive. They weren't hot swap drives. No, you had to disassemble the enclosures to access the drives. As you can probably imagine, it's impossible to wear warm gloves when you work with tiny screws.

Addendum:

My own dumb FreeBSD moves usually involve forcing a newer package to either be installed or updated on an older host. The OS tries to intervene to prevent impending disaster with messages along the line of: "Do you really want to remove two hundred other packages?" Unfortunately when such messages are ignored for the sake of expediency, the end result is often a broken host.

Never enough time to do it right in the first place. More than enough time to fix it later.

On the positive side, such scenarios exercise my OS installation skills.

 

[lgrant]

This is the Sun mainframe room circa 2001 at a mine in Wyoming. It was kept uncomfortably cold during Summer. And its cooler kept running during the Winter, when outside temperatures often dropped to below zero. You had to wear outdoor clothing in it, including warm gloves.

I tried my best to always use a terminal elsewhere in the building to avoid the room at all costs. Unfortunately, as shown in the photo, the console sat in the room. And to the left of the console sat a pair of stacked SCSI storage enclosures.

Typically they called me out to the site in the middle of Winter to replace a crashed drive. They weren't hot swap drives. No, you had to disassemble the enclosures to access the drives. As you can probably imagine, it's impossible to wear warm gloves when you work with tiny screws.

Next to the printer room, but still behind the electric doors, they had a narrow room with four or five IBM 3270 terminals, for people who had quick changes to make. (Kind of like the emergency keypunches they used to have back in the day.) The printer room floor was raised; the terminal room floor was not. And they did not have very good baffling at the edge of the printer room floor. So when you sat down at the terminal, the cold air would run right up the legs of your pants, like they were a couple of ducts. No one every spent very long working at those terminals.

 

[Crivens]

Since we left the FreeBSD turf...
Some guy I knew from school wanted to burn a DVD, and the burn command was rejected by the shell because "target in use".
So he went to root and "-f"ed it, and being a cool haxor, he did type the /dev/xxx name of the scsi target instead of /dev/dvd.
You see where this is going, yes?
The burn progressed amazingly fast, and the burner stayed amazingly quiet.
His main SCSI disc, however, was pretty busy, as was his vocabulary upon the system comming down like a pile of bricks - refusing to come up again because of "no boot loader" and "file system not recognized".

And what did I do?
Removing all the back up files by doing "rm *~". The ~ is right next to the enter key, so...

 

[cracauer@]

This is the Sun mainframe room circa 2001 at a mine in Wyoming. It was kept uncomfortably cold during Summer. And its cooler kept running during the Winter, when outside temperatures often dropped to below zero. You had to wear outdoor clothing in it, including warm gloves.

I tried my best to always use a terminal elsewhere in the building to avoid the room at all costs. Unfortunately, as shown in the photo, the console sat in the room. And to the left of the console sat a pair of stacked SCSI storage enclosures.

Typically they called me out to the site in the middle of Winter to replace a crashed drive. They weren't hot swap drives. No, you had to disassemble the enclosures to access the drives. As you can probably imagine, it's impossible to wear warm gloves when you work with tiny screws.

I spent a winter in an office where I needed one of those terrarium heat lamps on my hands to be able to type. No excuse for that.

 

[fmc000]

I haven't done anything stupid so far with FreeBSD in the four and a half months I've been using it; in my 30+ years career as a unix sysadm I can remember two very idiotic things I've done:

• I mistakenly typed the deadly "rm -rf /" instead of "rm -rf *" on a database server (luckily for me SunOS decided to start removing from a /store2 partition which contained "mostly" old stuff and I could press ctrl+C in time before real damage was done);
• while setting up a new Sun 450 in the data centre I mistakenly assigned the GW address to it instead of the correct IP address and left without going back to the office. The next morning I found the server turned off and when I asked why I was told that a colleague had to understand why everything stopped working at the same time and that they just yanked the power cord from the poor 450 when they noticed that the serial terminal was continuing printing some along the lines of "duplicate IP address found"...

 

[Crivens]

I'm not sure if I can share all of this, as it wasn't me - so I will keep it short.
Professionals have a shadow system parallel to the production system. This is where updates are tested and verified.
Good professionals have a tight lock on the production and schedule downtime of everything for updates and re-tooling the production.
Sometimes professionals hand out the card with the access codes to the production system instead of the shadow system (how did that escape the locked safe, I wonder)
Suddenly some poor contractor sod finds himself in the middle of the 15 minute window with the production system not comming up again after the update, and the knowledge that some very expensive suits will come yelling at him. Luckily he did the roll back in time and had a deal with the admins to the understanding of everybody being quiet about things.

 

[BaronBS]

and had a deal with the admins to the understanding of everybody being quiet about things.

"What happens in the IT basement, stays in the IT basement."

 

[wayes99]

The most dumb thing to do with it is screwing up the system while playing around. I once did with NetBSD, it fricking literally uninstalled the bin directory

 

[Alain De Vos]

I do a backup every 15 minutes. It's difficult to do dumb things ...

 

[Jose]

This is the Sun mainframe room circa 2001 at a mine in Wyoming. It was kept uncomfortably cold during Summer. And its cooler kept running during the Winter, when outside temperatures often dropped to below zero. You had to wear outdoor clothing in it, including warm gloves.

Is that a tape changer under the desk with the Sun workstation on it?

 

[lgrant]

I'm not sure if I can share all of this, as it wasn't me - so I will keep it short.
Professionals have a shadow system parallel to the production system. This is where updates are tested and verified.
Good professionals have a tight lock on the production and schedule downtime of everything for updates and re-tooling the production.
Sometimes professionals hand out the card with the access codes to the production system instead of the shadow system (how did that escape the locked safe, I wonder)
Suddenly some poor contractor sod finds himself in the middle of the 15 minute window with the production system not comming up again after the update, and the knowledge that some very expensive suits will come yelling at him. Luckily he did the roll back in time and had a deal with the admins to the understanding of everybody being quiet about things.

Had a similar thing on the IBM mainframe. Doing some tests of communication between systems, I had the MVS consoles of three virtual machines open in three windows, and the production console, which I only had read-only access to, in a fourth. Went to shut down a virtual machine, but accidentally typed my command in the production window. Production system came down. Operations was cool. They said “I didn’t know you had command access to the production system.” I said “I didn’t, either. Please take it away.”

 

[DonK]

Is that a tape changer under the desk with the Sun workstation on it?

I don't remember. Does anyone recognize the make and model from the design language? There were two identical units in the room on either side of the rack along with other non-Sun hosts on the other side of the rack.

 

[cracauer@]

Went to shut down a virtual machine, but accidentally typed my command in the production window.

On important machines I shadow reboot(8) and shutdown(8) in $PATH with a nag. If I actually want to reboot them I have to type the whole path.

 

[fmc000]

I don't remember. Does anyone recognize the make and model from the design language? There were two identical units in the room on either side of the rack along with other non-Sun hosts on the other side of the rack.

Looks like a tape library indeed. From the color I would say Qualstar or Compaq but I can't find anything similar on the web.

 

[ralphbsz]

while setting up a new Sun 450 in the data centre I mistakenly assigned the GW address to it instead of the correct IP address and left without going back to the office.

I've seen this happen twice. The first time was when Ethernet was just starting to get used at scale, in about 1991. The place where I worked had lots of workstations that were owned by outside parties (it was a research lab where lots of people came from universities and brought their own hardware), and it had a rule that anything connected to the campus-wide ethernet had to be configured by data center staff. But they didn't support MS-DOS (which in those days could use TCP/IP by buying an Ethernet card and installing some bizarre hacky software (either commercial from FTP software or K9something). So one grad student decided to do it himself, and brought the whole network down by using the address of the gateway as his address.

The second time was in the late 90s, and I was working at a company that sold big industrial equipment to semiconductor fabs. One of our machines got shipped to the biggest and most influential chip maker of the time, in spite of having failed QA testing in house (we needed the revenue, and the QA director who wanted to veto sending the machine to a customer was overruled). It brought down the whole network in the fab, and it took <insert famous chip maker here> 3 days to restart chip fabrication. Not going to discuss all the fallout from this, but it was a blood bath.