Guys, don't repeat my stupid mistake!!!

[dlausevic]

Finally got my roommate's old machine rebuilt and ready to install Kubuntu 6.06.

I'll spare the long long story about my issues. When you're installing anything connected directly to the internet, and you can't figure out why it won't see the connection, make ABSOLUTELY SURE the cable you're using isn't a crossover cable. This stupid mistake cost me HOURS of personal time.

[GliderMike]

I think anyone involved in tech has had a similar mistake at one time or another. Whether they'll admit it or not :-)

[Icarus]

Cross-over cable, good one...I usually for get the patch cable at the switches and wonder why it's not working :)

[leonpmu]

Been there, done that, got the T-Shirt, DVD, Video, CD, ogg file, etc... I got caught like that the other day, but the other way around, where the idiot at the store gave me a straight through when I asked for a crossed over cable... loat 2 hours that way!

[Pafnoutios]

Yesterday I wanted to clear the boot sector from one of my harddrives. A few hours later I was surprised to find a corrupted partition table. That's what "dd if=/dev/zero of=/dev/sdb bs=512 count=1" will do for you.

[cybertron]

Way back when I was helping a friend try to get their computer's sound working again, and eventually had to resort to calling the company's tech support. Naturally the first thing he had me do was make sure that all the cable's were plugged in correctly, which I believed they were (I was told sound had been working earlier that day, which turned out to be false...), but I did ask the support guy because one of the plug's color didn't match what it was plugged in to and he told me that was the way it was supposed to be (???). Turns out that wasn't correct. The cable that was supposed to connect to the subwoofer which then distributed the sound to satellite speakers had gotten plugged in in a loop from the computer to the computer.

So if I had just traced the cables like I should have in the first place I would have saved a couple of hours of time.* :o

*In my defense, I did check the cables, but wasn't careful enough because I thought the sound had been working earlier so the cables presumably had been plugged in correctly then. I guess the moral of the story is never trust what the users tell you.:D

[hlrguy]

A couple from my funny files.

A friend insisted that his USB mouse did not work in linux. I talked him through everything I could think of over the phone, then went over to his place. I plugged the USB hub into the computer, and amazingly, it started working.

My parents bought a color USB printer. It worked perfectly out of the box on Mandrake 8.1, except for color. When I went for a visit (they are 1200 miles away), they asked me about it. They said it was supposed to be color, but that had never worked. I checked the printer settings, etc. After a couple of minutes, I pulled the color cartridge, removed the protective plastic tab and put it back in. They got a good laugh out of that.

On the plain old scary side, ever see a person force a phone cable into an ethernet slot, then complain that dialup doesn't work?

hlrguy

[cybertron]

A couple from my funny files.

A friend insisted that his USB mouse did not work in linux. I talked him through everything I could think of over the phone, then went over to his place. I plugged the USB hub into the computer, and amazingly, it started working.

My parents bought a color USB printer. It worked perfectly out of the box on Mandrake 8.1, except for color. When I went for a visit (they are 1200 miles away), they asked me about it. They said it was supposed to be color, but that had never worked. I checked the printer settings, etc. After a couple of minutes, I pulled the color cartridge, removed the protective plastic tab and put it back in. They got a good laugh out of that.

On the plain old scary side, ever see a person force a phone cable into an ethernet slot, then complain that dialup doesn't work?

hlrguy
Heh, I can relate to all of those, especially the printer and the phone cable. I think one time someone may have even tried jamming an ethernet cable into a phone slot. My advice: "If you need a hammer to plug it in you're not doing it right.":)

[irlandes]

Just this week, I spent a couple days fighting cdrecord problems in puppy 2.0x. I wanted to try to modify remasterpup2 to make it run in dao while burning iso's, which is the problem cdrecord has with kernel 2.6. 'find' told me it was only in initrd.gz which is read-only. i wasted two days on this, before I accidentally saw remasterpup2 in /usr/sbin where it can easily be modified if I am smart enough to make it work. Clearly 'find' has problems or is different in puppy.

Please don't answer those issues, I have postings all over the place, and have a solution at hand. I changed burniso2cd to run in dao mode, and it does very well. And, I am going to change remasterpup2 to simply create a iso file, and burn it as a separate action, since dao needs -tsize-nnnns and that is hard to make without bash diffs, which I can install. My point here is to avoid wasting others time on this side issue, just relating my own GRRR!!!!!!!!!! story.

[dlausevic]

Thanks guys for posting your stories. Yall didn't have to. Anyway, I do feel a bit better about it. But I am still embarrassed.

[janne_oksanen]

One of the offices had a blown up fiber/copper converter yesterday and they couldn't get the replacement unit working. So I drove 6 hours just to tell the local techie to use the crossover switch on the converter and that fixed it.

[m3rlin]

i think my biggest flaw as a network administrator happend when i had a power failure and i forgot that my firewall didn't had the UPS, and i was wondering why all services were failing, llool. We learn with mistakes :P

[cybertron]

Thanks guys for posting your stories. Yall didn't have to. Anyway, I do feel a bit better about it. But I am still embarrassed.
We enjoy these types of threads. Check out the "Stupidest thing a CS professor has said in class" one sometime.:)

[leonpmu]

I forgot one, a friend of mine the other day says to me, "hey my printer doesn't work anymore". So I ask him what happened, he said he had unplugged everything to clean the box, and when he plugged it all back in, the printer no longer worked. So I got on my hands and knees, and checked that back of the PC. I had to crawl out from under the table I started laughing so hard.

He had managed to force the USB cable into the ethernet port!! Thankfully the ethernet is still working because I transferred some files to it after that!! :D

[m3rlin]

llool thats funny, that reminds me of a situation that a person called me saying that the usb scanner had unplugged and she plug it again, but the scanner wasn't working, so i checked what was happening on the software, when i realise that the problem was that he wasn't connecting with the scanner, so i checked the cable and i saw that the person pluged where the screw is, lloool

[bwkaz]

Let's see...

(1) When your switches don't have Spanning Tree turned on, don't ever plug a crossover cable into two ports on the same switch, unless each port's "default" VLAN ID is different. Otherwise, at the first broadcast packet (e.g. an ARP request), you'll create a broadcast storm -- all the lights on every switch in your building will light up and stay on, and nobody will be able to talk to anything on the network because of all the repeating traffic.

Different VLANs may be able to be plugged into each other, but we had some issues with that as well. Don't remember exactly what they were, though; IIRC it was something to do with ARP responses or switching decisions or something.

(2) When your computer is plugged into a UPS, and the "normal" shutdown procedure for the computer is to unplug the UPS and let it power-down with apcupsd (or on windows, apcupsd or APC's PowerChute), don't unplug the power cable from the PSU instead of unplugging the UPS from the wall.

:o

(3) When you write a program that uses WinPcap 3.x or earlier to get and set various NDIS OIDs, don't disable (or change any options on) the NIC that your program is bound to without shutting down the program. Otherwise you'll blue-screen. (The problem is a stale handle in the user-mode program. The WinPcap driver has already lost its handle to the lower-layer miniport device, but the handle that the user-mode program was given doesn't get invalidated. So the WinPcap driver blue-screens as it tries to dereference an invalid handle.)

WinPcap 4.0alpha1 and (hopefully) any later version shouldn't have this problem.

[gehidore]

can we make this the [Main] don't do this at home - post your story here
thread?

As for me.... we spent 3 hours trying to figure out why the box we just build wasn't working, push the buttons the fans/cdrom/harddisk all spun up... but no post codes no nothing, turned out the PSU had not been plugged into the mainboard, damn switches attached to the PSU's....

[cybertron]

Let's see...

(1) When your switches don't have Spanning Tree turned on, don't ever plug a crossover cable into two ports on the same switch, unless each port's "default" VLAN ID is different. Otherwise, at the first broadcast packet (e.g. an ARP request), you'll create a broadcast storm -- all the lights on every switch in your building will light up and stay on, and nobody will be able to talk to anything on the network because of all the repeating traffic.

Different VLANs may be able to be plugged into each other, but we had some issues with that as well. Don't remember exactly what they were, though; IIRC it was something to do with ARP responses or switching decisions or something.
On that note, don't ever plug a newly installed Windows machine into your network if it has two network cards and you haven't disabled the MAC bridge that it probably decided to set up for you. It can cause your port to get automatically disabled and your network administrator to run out of his office wondering what you just did. :mad:

Oddly enough, I only plugged the one network card in, but Windows still decided to start sending out packets that tripped the loop detection on the switch.


(3) When you write a program that uses WinPcap 3.x or earlier to get and set various NDIS OIDs, don't disable (or change any options on) the NIC that your program is bound to without shutting down the program. Otherwise you'll blue-screen. (The problem is a stale handle in the user-mode program. The WinPcap driver has already lost its handle to the lower-layer miniport device, but the handle that the user-mode program was given doesn't get invalidated. So the WinPcap driver blue-screens as it tries to dereference an invalid handle.)

WinPcap 4.0alpha1 and (hopefully) any later version shouldn't have this problem.
Alright, see, that just doesn't belong in this thread. If most people can't even understand the problem it wasn't "stupid" in this context.:D

[M_N]

I recovered a backup from an AIX server right over the /etc of my linux box...
When I realized what I had done, I just went for a walk ;)

[Pafnoutios]

can we make this the
thread?


Hmmm, no nested quotes, eh?
I suggest "Don't try this at home, kids."

[acid45]

Mount your partition, not your block device :P

[gehidore]

Mount your partition, not your block device :P
One I'm sure we've all done, fdisk your *disk* not partition.

[Parcival]

Time to dig up this old classic of mine (http://www.justlinux.com/forum/showthread.php?t=112824). (I accidentally hit the main power button on my computer case and came to justlinux desperately crying for help)

[cybertron]

Hah, I did that on my first build. I pulled the power supply out of the box and said to myself "I better not forget to flip that switch" and promptly forgot to flip the switch (although in my defense I had a desperate hunt in between for some screws to attach the motherboard to case because they forgot to include them). I actually took all of the front panel connections off and reattached them before realizing what I had done.

Oh, and later that summer I was helping another friend build a computer, but she was going to install Windows. We got it all put together and she booted the Windows CD. Surprise! It can't find the hard drive because it was SATA and Windows doesn't support that without a floppy for the drivers. Naturally I hadn't considered that because it wasn't a problem for me...we ended up leaving her to sort that out herself because it was late and we needed to leave. I guess the stupidity here though is that Windows still requires a floppy drive for installation (which she had to cannibalize from an older system because naturally she wasn't putting one in the new one).:)

[bwkaz]

Alright, see, that just doesn't belong in this thread. If most people can't even understand the problem it wasn't "stupid" in this context.:D Yeah, well, I needed something to use for number 3, and that was the first thing that came to mind. :p

[gehidore]

Time to dig up this old classic of mine (http://www.justlinux.com/forum/showthread.php?t=112824). (I accidentally hit the main power button on my computer case and came to justlinux desperately crying for help)
See and this whole time I thought you were never a newbie... :p

[leonpmu]

NO!! It can't be!!! Are you saying that Parcival was a newbie once??? It is NOT possible!!! :o :)

[Parcival]

Hello, my name is Parcival, and I have been able to live without newbie for two months now.

[cybertron]

All at once now: "Hi Parcival".:)

[Icarus]

Liar, I saw you newbieing just the other day! Once the group hears about this....

[Wiz]

*In my defense, I did check the cables, but wasn't careful enough because I thought the sound had been working earlier so the cables presumably had been plugged in correctly then. I guess the moral of the story is never trust what the users tell you.:D


For my shame, I am a (Windows-based) corporate bob. I can testify to this.

USERS LIE! Never trust 'em. :cool:

[r0ck]

Finally got my roommate's old machine rebuilt and ready to install Kubuntu 6.06.

I'll spare the long long story about my issues. When you're installing anything connected directly to the internet, and you can't figure out why it won't see the connection, make ABSOLUTELY SURE the cable you're using isn't a crossover cable. This stupid mistake cost me HOURS of personal time.

It all depends where you come from. For me, that would have been probably the first thing to check since I work in labs and we got patch pannels that look like walls of [colorful] spaghetti.

Some of the goofs I did in my days as a lab tech

* Restored the wrong tape on the right machine or vice-versa (solaris) . No biggie, just wasted some time.

* Re-installed the wrong node, simply because my co-worker has a different way to labeling things and I wasn't paying attention (linux). We had a bunch of these nodes and they were all HW identical side by side. That was kind of hard to explain why I made this goof. :D

* Rebooted a production server by typing ctrl-alt-del (linux) cause it was part of a KVM cluster. I though I was connected to a machine right above it. No biggie, but it took something like 7 hours to finish the journaling on a the terabyte raid that was connected to it

* Deleted the wrong partition when doing newfs , instead of doing it on c1 I did it on c0 (solaris) . Not sure what I did to fix this, but I don't remember getting in trouble for it

* Plugged a signalling (ss7) connection into someone elses DIP . Wasn't my fault, I got the definition from the person responsible for assigning me the DIP. It was working fine, but there was already a hardwired connection in the back of the patch pannel and I was patching in the front , which was overwriting the hardwired connection. If anyone worked with ADC panneling , you know what I'm talking about. So I put my connection ... then the link stops working the next day. WTF mate!?? Customer is screaming. I go nuts trying to figure out why its not working. I go check the patch pannel, my cable is disconnected. Someone removed it!! So I plug it back. Again, next day its removed. I plug it back, but this time I punch holes in a note and put the the bantam connector through it and put the message on the DIP. Guy calls me next day, tells me that I'm hijacking his connection which was used in a customer demo. :O Didn't get into trouble for it tho, wasn't my fault

* Not really a goof but a cPCI (CompaqPCI) board in a magazine started smoldering. We have a lab with 300 different machines, and I had to go sniff each one to see where the odor was strongest. We have AC blowing full blast so if there's smoke, its immedially dissipated and the whole labs smells the same. Eventually the board started smoldering more and caught on fire. That in turn tripped the building alarm and annoyed the whole company for about 15 mins till it was disconnected. No damage was done.

Some of the goofs I SEEN in my days as a lab tech

* We have these custom built SUN servers here. The power socket on them is in the front, not back and its a trapezoid shape with a 3 male pinout. Girl unpowered the server, then has to put back the plug and she puts it the wrong way. She obviously couldn't put the connector fully in, but all you need to have is contact between the pins and *BOOM*. Server blows. Smoke. Girl is freaking out. Was funny cause she was this asian girl, totally shy and quiet :D She was some student and I have NO CLUE why she was allowed to even touch the equipment. Well, there goes a 30K machine, but it was under warranty so it was all good. We actually made a suggestion to the manufacturer to change the power plugs and they did on the newer models of the box.

* We have our cabling under the floor tiles, and we got a VESDA fire alarm system with CO2 gaskets/sensors all over the place. In case of fire/smoke the system releases the CO2, but you can trip one of those gaskets manually and release the CO2 by stepping on it. A power cabling contractr opened a bunch of floor tiles, and left the lab for a coffee break or something. Another guy comes in and he's has to pass over these gaps in the floor. I guess he wasn't paying attention and tripped one of the gaskets. Room gets filled with CO2. Im in charge of the labs and was just in a lab beside the one that just got fumigated. I go there real quick to check, and I see the guy that tripped the gasket like some lost sheep. I had to scream at him but eventually he left. Costed a few Ks to replace the CO2 talk and calibrate the system.

* We got a node that was poorly assembled and tested. That was back in 1999 when we didn't subcontact the work. A piece of metal got loose and caused a short circuit. The node was operating, but it you could see from the LEDs in the front that something was seriously wrong. These 2 "engineers" wanted to do some testing, so they started plugging/unplugging power cables in the back, heh, one of the guys fried a whole magazine. 10 boards @ few K's each + magazine. Heh.


There are other stories a lot more... ;)

[hlrguy]

I work telecom too, and before the lab tech got fired, being lazy and instead of running new power taps to new frames, heck, no one will turn all the nodes/frames on at once, so why not connect a 140 amp total draw up to a 100 amp feed. What can it hurt?! Amazingly enough, everyone wanted to use their equipment at the same time. It was a breaker flippin mess. Just as I would get set up to run some automated tests, poof, a colleage would power on his boxes where the breaker mysteriously tripped and mine would go. The lab tech was fired for this, and because 100% of the time, his labelling was never 100% right.

I can remember days spent trying to get all SS7 links up and running because 7 would go to the test equipment, 3 went to a different HLR entirely, 4 were terminated at a patch panel without connections on the back and 2 were connected to each other.

My biggest peeve with him was him using flanged side cutters to cut his tie wraps off, leaving just enough of a stub to successfully scape your hands everywhere you reach. In telecom you need flush cut side cutters. (well, anywere miles of cable is tie wrapped)

Oh yeah, this guy would hide the floor pullers (lift the raised subfloor tiles) so that everything had to go through him (empire building). He even came to my cube once and demanded the one I had back. I told him loud enough for several to hear that if he was not so useless and I didn't need to fix everything I ever worked on, he could have it back. Obviously, he liked me a lot.

When we hired the new lab manager, he spent 9 months and rewired, re-ran where required, and re-labelled at least 200 frames and all wirling in the floor and ceiling in our lab.

hlrguy

[r0ck]

OMG! I work with HLR also! Small world. :D

Bro. I have had the EXACT same issues as you describe. Maybe not to the letter, but we had power issues very much to what you describe. We had a powerbar fully loaded with SUN equipment, and a plastic plug with a note saying "MAXED". Some twit removed the plug and connected his machine cause he was too lazy to move it to another rack. And you know what happened next. Everything went down.

Flush cutters should be used and some idiot is not, then you put your hand underneath a tile to pass a cable or something and you know that when you'll get your arm out, its gonna be all scraped cause the guy didn't use flush cutters.

[cybertron]

Well, I never caused thousands of dollars of damage (yet;)), but one time I did break a toner cartridge worth a couple of hundred dollars. The problem was that on this one particular brand of toner cartridge for one fairly uncommon type of printer, they put a sticker on the guard that protects the roller when it's not in use. If you don't remove the sticker then when you put the cartridge in the guard doesn't get opened like it should and the cartridge ends up completely stuck. Since the sticker was on the bottom I didn't see it when I put the cartridge in. In the process of trying to get it out I broke off a piece of the guard arm and made the toner pretty much worthless. Since then I made it a point to tell every new employee that I trained about that type of toner cartridge so they didn't duplicate my mistake.:rolleyes:

Apparently I wasn't the only one who did this either. Our inventory person gave us a little talking to one day because 4 or 5 toner cartridges had been broken that year alone. :eek: