Regarding drains, I've had that several times. I think the cause is that every shiny new washing machine has a much more powerful and effective pump than the last one, thus always being able to test any partially-blocked pipe -- or even an unblocked but slightly too small, or just not sloped enough -- slightly more than slightly past its flow limit....
Bugs that won't die are a side effect of Sexy Project Syndrome. The people in charge assign themselves the juicy bits (optimize the hash algorithm! serialize the class hierarchy!) leaving the fundamentals to the Low Man On The Totem Pole, i.e. the new guy or the intern. So build management, version control, installer management etc. wind up being slapped together to the detriment of the schedule and the product. Only occasionally does a wise person(*) come along and fix things so that what once took weeks and was unreliable, suddenly happens in a few hours totally reliably with actual reports generated. And still nobody cares....
It's not just IT, even the slightest notion in your head about replacing your car means bits start falling off the old one IME.
Love that Gary Numan song as well, I fondly recall the bewildered looks on TOTP when he performed it from people expecting electro-pop. Metallica were also on that episode, it was a shock to the system for a lot of people.
Had expected this to be a tale about how some half-baked automation that had over-ridden the language files overnight, but only after something else had triggered the scheduled task … i.e., Colin from accounts had changed a spreadsheet, which triggered a cascading chain of activities until the language file was deposited over the top of the recently updated language file i.e., they’d only ever needed to update the language file the first time because Colin had merged something Spanish into the spreadsheet.
I once spent weeks of my life trying to stop people using multiple object names in a firewall system, resulting in multiple host objects like this all pointing to the same thing.
Hostname
Hostname-InterfaceName
HostnameIPAddress
Hostname-IPAddress
Hostname-32
Hostname/32
Each department (Competency - there’s the Oxymoron in this tale) had their Specialist write their own script to generate firewall rule content from their own communication flow template result in chaos trying to work out why 6 different host objects could be in the same firewall group.
Then the new boys and girls would be given access to the online folder where the scripts resided and somehow get instructed to use the old script and off we’d go again.
Still, five years on, and despite being asked each year if I still need those access account - to which I say “No, remove them” - there’s an automated email asking me if I still need the access accounts I last used in 2018.
Trying, between family emergencies - an entire house, home to six people, burned to the ground - to roll back the firmware update that I foolishly allowed to proceed on our Brother colour laser. The previously faultless machine now cannot be printed to expect from one PC that lives in the same room as the printer. Any other PC, anywhere else, can't reach it, even though this lone machine finds it easily via WIFI.
And, to add to the stress level at a time when we actually need to print a LOT of stuff, I see in the Internet that we likely now can only use expensive Brother branded toner, not the perfectly fine Amazon ones.
There's a related phenomenon I've experienced more than once:
Process isn't working:
Investigate and locate problem: n hours
Fix problem and verify process works now: 1 hour
Wonder how the process ever worked, which it clearly had been doing until yesterday: 10 minutes
Investigate how the process ever worked, and in so doing uncover the bit you'd forgotten about even though you designed and built it in the first place: 3n to 20n hours.
Bonus points if it turns out the bit you fixed is not used, but either the fix coincidentally also fixed the real process, or it fixed itself without your intervention.
Of course the opposite of this effect is when a colleague/customer/family member has a problem with their computer, it disappears as soon as you walk near it. Most assume you have done some weird form of remote voodoo to cause the computer to suddenly comply but we all know the likelihood is PEBCAK
I suspect it's not so much that your backup solution starts looking shaky when the primary fails, but that when the primary goes down you actually look at the backup for the first time in ages and suddenly see how rickety it has in fact been for the last several months without anyone noticing.
That faulty UI message you fixed? It's come back and no-one knows why
The 11am prostrate, presumably the pub is open by then ...
I am always annoyed by the astoundingly stupid who insist on "correcting" the correction and putting spurious apostrophes back.
I'm sure I heard of somewhere that had a superhero who would go round in the middle of the night and fix such errors on shop fronts!
Pfft .... obviously the fixes all failed because the child element was not linked to its parent.
Regarding drains, I've had that several times. I think the cause is that every shiny new washing machine has a much more powerful and effective pump than the last one, thus always being able to test any partially-blocked pipe -- or even an unblocked but slightly too small, or just not sloped enough -- slightly more than slightly past its flow limit....
Bugs that won't die are a side effect of Sexy Project Syndrome. The people in charge assign themselves the juicy bits (optimize the hash algorithm! serialize the class hierarchy!) leaving the fundamentals to the Low Man On The Totem Pole, i.e. the new guy or the intern. So build management, version control, installer management etc. wind up being slapped together to the detriment of the schedule and the product. Only occasionally does a wise person(*) come along and fix things so that what once took weeks and was unreliable, suddenly happens in a few hours totally reliably with actual reports generated. And still nobody cares....
(*) That would have been me, once upon a time.
It's not just IT, even the slightest notion in your head about replacing your car means bits start falling off the old one IME.
Love that Gary Numan song as well, I fondly recall the bewildered looks on TOTP when he performed it from people expecting electro-pop. Metallica were also on that episode, it was a shock to the system for a lot of people.
Enquiring minds want to know - what was it's problem? :)
Had expected this to be a tale about how some half-baked automation that had over-ridden the language files overnight, but only after something else had triggered the scheduled task … i.e., Colin from accounts had changed a spreadsheet, which triggered a cascading chain of activities until the language file was deposited over the top of the recently updated language file i.e., they’d only ever needed to update the language file the first time because Colin had merged something Spanish into the spreadsheet.
I once spent weeks of my life trying to stop people using multiple object names in a firewall system, resulting in multiple host objects like this all pointing to the same thing.
Hostname
Hostname-InterfaceName
HostnameIPAddress
Hostname-IPAddress
Hostname-32
Hostname/32
Each department (Competency - there’s the Oxymoron in this tale) had their Specialist write their own script to generate firewall rule content from their own communication flow template result in chaos trying to work out why 6 different host objects could be in the same firewall group.
Then the new boys and girls would be given access to the online folder where the scripts resided and somehow get instructed to use the old script and off we’d go again.
Still, five years on, and despite being asked each year if I still need those access account - to which I say “No, remove them” - there’s an automated email asking me if I still need the access accounts I last used in 2018.
Trying, between family emergencies - an entire house, home to six people, burned to the ground - to roll back the firmware update that I foolishly allowed to proceed on our Brother colour laser. The previously faultless machine now cannot be printed to expect from one PC that lives in the same room as the printer. Any other PC, anywhere else, can't reach it, even though this lone machine finds it easily via WIFI.
And, to add to the stress level at a time when we actually need to print a LOT of stuff, I see in the Internet that we likely now can only use expensive Brother branded toner, not the perfectly fine Amazon ones.
There's a related phenomenon I've experienced more than once:
Process isn't working:
Investigate and locate problem: n hours
Fix problem and verify process works now: 1 hour
Wonder how the process ever worked, which it clearly had been doing until yesterday: 10 minutes
Investigate how the process ever worked, and in so doing uncover the bit you'd forgotten about even though you designed and built it in the first place: 3n to 20n hours.
Bonus points if it turns out the bit you fixed is not used, but either the fix coincidentally also fixed the real process, or it fixed itself without your intervention.
Of course the opposite of this effect is when a colleague/customer/family member has a problem with their computer, it disappears as soon as you walk near it. Most assume you have done some weird form of remote voodoo to cause the computer to suddenly comply but we all know the likelihood is PEBCAK
I suspect it's not so much that your backup solution starts looking shaky when the primary fails, but that when the primary goes down you actually look at the backup for the first time in ages and suddenly see how rickety it has in fact been for the last several months without anyone noticing.