2y ago

Peak engineering

73 comments

Hate to be that guy, but for something bad like this to happen, it's never one person's fault. Like the engineer who nuked the gitlab backup by mistake while production had been deleted. He didn't lose his job and rightfully so, there were a thousand other issues that led to that.
- Recently, YouTube started adding a tracking parameter to their share URLs, when using the "share" button on a video. With this, they can track who is sharing videos with who, and under some circumstances even how they are shared. The tracker starts with the question mark in the link you posted and the link works perfectly fine without that part.
  
  Thanks for pointing this out, I think it's not only youtube that does this and the solution is to edit such link before sharing them, right?
- https://www.latimes.com/business/autos/la-fi-hy-vw-hearing-20151009-story.html
  Or when VW management tossed some employees under the non emissions compliant bus.
- There aren't bad employees. Only bad processes.
  
  I struggle with your statement. I've worked with inept people, but they weren't malicious. In one instance the inept person was the DBA. That one guy made the whole team's life miserable. He was a significant reason I quit a job.
  I don't know what framework you could put on a DBA to make them not royalty mess up a system.
  
  You never worked anywhere?
  
  And managers
- I love the video
- Don't hate to be that guy. You're completely right

That wasn't the design engineer's fault. It was the design engineer's fault and the QA tester's fault and management's fault.
- Also...I blame society.
  
  I blame reddit
  
  Plus the government
- It's funny to me when people act like this is some weird take but at the same time call every layer of management above the workers "leaders". If leaders aren't responsible for anything then what purpose do they serve?
  
  I once got a funny look at a job interview when I implied managers should do their job. It was a pretty clear indication that I don't want to work there and I think my response also crossed me off their list.
- And marketing, and sales. Tons of people would have pointed this out.
  Some small committee of managers would have come up with some reason to dismiss all of these complaints.
  Also there’s a very simple workaround for this that doesn’t require a full recall.
  And what is going on 4 ports to the right? Seems like a similar problem.
  
  And marketing, and sales. Tons of people would have pointed this out.
  Maybe, but (unlike QA and management) I wouldn't expect them to notice the problem or hold them responsible for failing to do so.
  
  And what is going on 4 ports to the right? Seems like a similar problem.
  The mini-USB console port? Yeah, that could be a similar issue, but I've never had to use that port while a device is in active production. If I can't access the device via IP on our management fabric, the device is probably in a broken-enough state that I can probably unplug a cable or two to attach a laptop and troubleshoot.
- I guarantee management was rushing this product out the door to meet deadlines without adequate testing and without running a pilot program. That's the only way this could realistically happen.
  I suspect this one falls squarely on management. But I bet they didn't take the blame.
  
  Incan imagine many reasons to have done this on purpose
  how many bad raises to the center of a tootsie pop

Right up there with the classic Macintoshes with unshielded speakers nested right up against the hard drive and would periodically emit a tone that would reboot the computer.
- My personal favorite was the early-90s Macs that didn't have an eject button for the floppy drive, but did have a pushbutton power switch ... directly above the floppy drive. It took me weeks to stop powering off the computer every time I wanted to eject the floppy. Silly me, not picking up on the oh-so-very-intuitive practice of dragging the floppy icon over to the trash can in order to eject it.
  
  Also extra fun was if the computer was non-functional and had a floppy disk in it, since it required working software in order to eject the disk, you had to do some disassembly in order to retrieve the disk.
- Was that the same mac that had an officially sanctioned maintenance drop of 5cm to combat socket creep?
  
  That would be the Apple III
  
  Wait you had to drop it 5 cm so it would knock something back into place?
  
  I know what all of those words are but I don't understand what this is referring to and I can't find anything in a web search. Can you elaborate?
- That sounds like planned obsolescence as the speaker slowly rots the bits spinning nearby

How the actual fuck did this get through QA and production?
- QA budget is real low. They can only afford the ones that are bare copper stuffed into a RJ45.
  If they're lucky a DIY job with no exposed pairs outside the RJ45

How did the design make it past quality control, though? Sounds like a few balls were dropped.
- QC probably tested with normal cables, not the protective jacket ones.
  
  That's pretty shitty QC, but perfectly believable.
- They're is no QC that's how.

Just to clarify something... they say it "resets the switch" but some people may not realize in Cisco parlance, that means factory reset, as in wipe it completely and start with a fresh config. It was WAAAY worse than just rebooting it.
When Express Setup is inadvertently invoked by the protective boot of the cable, these messages are seen in the syslog:
%SYS-7-NV_BLOCK_INIT: Initialized the geometry of nvram
%EXPRESS_SETUP-6-CONFIG_IS_RESET: The configuration is reset and the system will now reboot
%SYS-5-RELOAD: Reload requested by NGWC led process. Reload Reason: Reload command.
%STACKMGR-1-RELOAD_REQUEST: 1 stack-mgr: Received reload request for all switches, reason Reload command
%STACKMGR-1-RELOAD: 1 stack-mgr: Reloading due to reason Reload command.
After this occurs, the device resets. The startup configuration is erased once the device enters Express Setup.
- I almost found this out the hard way. I think on the Cisco equipment, it's something like: Hold for 2 seconds to cycle power. Hold for 5 seconds to wipe config.
  Our IT guy nearly had a heart attack when, over the phone, I asked if I should press the little "Reset" on the back.

That couldn't have been an accident, they wanted people to suffer.
- there's a saying, never attribute malice when incompetence would do. I don't think an engineer would do this on purpose, but it really should have been caught by QA.

Found one of these at work one day. It's equally hilarious in person
- I would love to see one of those IRL

And that design engineer's name? Pagliacci
- But doctor... I am the ethernet cable.
  
  Doctor? I thought you were this ship's doctor!

how common are those specific types of cables? The ones with that specific “protective boot”
- Really common with quality premade cables.
  You know, the ones used in datacenters
  
  Yeah every one of my Ethernet cables at home have that.
- Almost every Ethernet cable has this, you can search for RJ-45 cables on Amazon and you will basically always see something like this.
- Very. While the specific length and position of the protective varies between brands, the concept is very common in high-end premade cables. All of the premade RJ45s I use at work have it. The purpose is so that you can pull the cable in one end without the plastic clip snagging in some other cable and breaking.
- I've seen some really cheap ones that don't have it. But the vast majority of cables like that I've seen have the protective boot.
- Basically all of them
- I believe it's part of the standard so, all of them. Unless you get your cables from some cheap Chinese knockoff brand, but I don't imagine that any business would do that. Not worth the risk.
- Pretty charitable calling that fucking flipper a 'boot', I must say.

Haha!

73 comments