"..someone typing Google into Google caused the whole outage."
Don't be daft. Everyone know that doing that kills the _whole_ internet - this outage only affected O2.
This was the week when O2's network went titsup, leaving thousands of its users and GiffGaff mobile customers without the ability to make calls, write emoticon-laden texts or use 3G. Yes, that's right, they were CUT OFF FROM THE INTERNET. So, they were understandably peeved. As is traditional in these situations, O2 said it …
Don't be daft. Everyone know that doing that kills the _whole_ internet - this outage only affected O2.
Plod: We have records of you accessing this site known for CP - 1000s of times in an hour
Anon: I was trying to destroy the site, really I don't like CP
Plod: You do realize that is illegal as well
Anon: Oh, bugger
There is a very good reason why this is such a laughing stock of a solution, and it is because almost nobody thinks to do it when it is needed. Having worked within the support fields for both IT and Mobile Telecoms, it still astounds me that normal rational sentient human has such trouble trying that first, rather than just ringing the helpline.
It really happens, and in the case of mobile phones, switching them off and on solves so many issues. It is the main reason why customer service tell people that the sim needs to be clean and to take it out and clean it. This forces the phone to be switched off and is not related to the cleanliness of the sim.
I am afraid that people will just not try the obvious first.
The fault was apparently down to phones not being able to register on to the network. Asking a user to turn the phone off and on again is probably easier than explaining that to them and asking them to force their phone to re-register...
Any idea why T-Mobile has told me twice now to just pull the battery out of my Galaxy S2, rather than shutting it down properly and then taking out the battery?
Surely off is off.
I'm sorry SuperTim, but I'm not with you that switching it off and on again is ever an answer to anything. Imagine your doctor saying that you just have to die and be reborn to solve the illness you have. Nor has "turning it off and on again" ever helped me when I couldn't save a document.
There are various systems that are always powered for life and they never have to be switched off and on again to work properly (e.g. the alarm/immobiliser in your car, your RSA security token/cotag access pass, and many other things).
Things should not be designed so that they need to be reset to work properly.
I think the point is that nobody SHOULD need to do that. If your system crashes and can only be fixed by turning everything off and on again, you haven't "fixed" anything. You've just restarted the system. The same problem is still inherent to that system.
It's like saying we could all just have avoided all the Y2K issues by not running the systems during the switchover. Maybe. But that's not the point.
Computers should NEVER reboot. Especially in the modern age of phones being computers, computers having extreme standby and hibernate modes, etc. A reboot is a sign that something, somewhere, is wrong with the system. Reboot "solves" one problem (the user regaining control of the system) but not the cause.
Anything that recommends me to "reboot" something, I immediately sigh and don't do it. If rebooting fixed it, it won't fix the problem, only the symptom. I have Linux servers with 600+ days uptime. Because you don't NEED to reboot just because your ADSL router changes, or your DHCP server goes down for a while, or a runaway script causes problems. You just manage the PC, ask services to reconfigure themselves and carry on - no reboot needed.
Anyone who tells you to "just reboot it" is only interested in getting you off the phone. Not fixing the problem.
I agree with pretty much all of that.
How many users in the real world are anywhere near computer savvy enough* to even be talked through asking a service to reconfigure itself? You or I may be able to fix things with a line or two at the command line but for most users that's a black art that scares them witless. Getting them to restart knowing that it will most likely fix the _symptom_ is by far the easiest and quickest option from a support perspective. Not saying it's right, but that's where we are.
*I thought I'd seen most things, but recently saw someone trying to put an SD card into a slot loading CD/DVD drive on a photo kiosk. The mind boggles...
Switching it off is a software event on almost all mobile/smartphones: there's still power to some functional aspects of the phone, which often still includes the radio section of the device (it goes into listen-only mode in case the network wnats to force a device power-on event...).
You are probably aware that you can put the phone into 'airplane mode' or similar which stops the radio section but still allows the device to function as, say, a media player or a(n off-line) document editor/viewer or both!
THE only certain way to ensure that such devices are truely "off" is to remove the power source.
"Things should not be designed so that they need to be reset to work properly."
As someone who earns a living designing embedded systems, I couldn't agree more with that statement. However, once you start combining purist design goals with real-world project constraints, sometimes compromises end up being made, and if resetting the device is a relatively easy task for the end user to perform, then it may be preferable to rely on that as a final, if all else fails, path to recovering the device, if the alternatives are to increase the product cost to unacceptable levels, or delay release by an unacceptable amount of time whilst you design out every last area of instability.
For the average bit of consumer electronics gear, where the end user is often more concerned about how many features they're getting for how little money, and just how soon can they get their grubby mitts on one, then generally the project manager isn't going to look too favourably on engineering requests to extend the development timescale or add costs to the bill of material any more than is required to get the product out the door as soon as possible and working well enough to satisfy the masses. You do what you can with the resources you've got, and then move onto the next project (which is probably already behind schedule because your engineering team is understaffed and overworked, and so you're under even more pressure to get this next one out the door ASAP...)
"Anyone who tells you to "just reboot it" is only interested in getting you off the phone, *thus* fixing the problem.
There. Fixed that for you...
Unfortunately BristolBachelor, it IS an answer. Due to the shoddy coding, construction or testing of modern stuff, they are all riddled with terrible bugs and inconsistencies. Your hopelessly romantic view of how things should be just does not happen in real life.
I agree stuff should be better, but it isn't and therefore needs to be reset sometimes. As for switching a mobile off and on again, It will often force the phone to reregister with the network, something that may take many days to do if you didn't reset.
Besides, some of the things you mention do actually fail sometimes too you know!
Except it ain't. The days of the "off" switch physically cutting the circuit are long gone. The screen may be turned off, but there's no guarantee the processor has tidied itself up and pulled the plug - especially if some random process has gone tits-up and is refusing to shut itself down. And if the phone is on its way off, and you hit the "on" button during that time, usually it'll bypass all the starting-from-cold stuff and jump straight back to normal running. The only way to *guarantee* starting with nothing else running is to pull the battery.
It's not a solution, it's a workaround.
It'd be lovely if mobile phones and PCs were developed to the same standards as automotive software. I hope you're ready for a basic candybar (without touchscreen) to cost you the fat end of a grand, though.
"Unfortunately BristolBachelor, it IS an answer. Due to the shoddy coding, construction or testing of modern stuff, they are all riddled with terrible bugs and inconsistencies. Your hopelessly romantic view of how things should be just does not happen in real life."
No, it is NOT the answer. the answer is to rewrite better, construct better, test better, fix bugs, and fix inconsistencies. It does happen in real life, just not for consumer crap, and the only reason it DOESN'T happen in consumer fields is consumers don't normally demand it. Even there though, compare a mondern version of windows to one a predating 7. MS had devoted a large effort to makeing sure system do not crash. Security issues may be abound, but I don't think I have ever seen a blue-screen on a Windows 7 machine, except when someone removed the boot disk from a running host, which I'm will to let slide.
"I agree stuff should be better, but it isn't and therefore needs to be reset sometimes. As for switching a mobile off and on again, It will often force the phone to reregister with the network, something that may take many days to do if you didn't reset."
What you describe is a work-around. Calling it a fix does a great injustice to those who DO endever to release bug-free products, and properly correct any bugs which are found in the wild.
"Besides, some of the things you mention do actually fail sometimes too you know!"
Yes they do, BUT noone in these fields considers power cycling a "fix."
"We will soon have the capacity through our technology to repair our bodies, to self-regulate sleep patterns and immune responses, and then extend that to the social networks so that we can improve the wellbeing even of our perceived enemies, so that they are no longer our enemies." - In my life greater capacity through technology leads to me drinking more whiskey and finding new ways to abuse people online.
Didn't I see a Dr. Who episode recently about that? Icon does the rest.
I'm sure the critical functions are tested and retested but not some of the less important stuff though.
A relative of mine had a not very old car (I'll not mention the make or model but its quite popular in the UK) which they purchased new in 2008
After 12 months they noticed the lights on alarm that is supposed to sound when the driver's door is opened with the headlights on stopped working.
They mentioned it during a dealer service.
The solution? They had to sit for two hours while the vehicle's management software was completely reflashed to cure the issue.
As for turning it off and on or pulling the battery this ensures it fully reboots and the memory is clear.
I wonder how many of the commentards banging on about coding being error free actually write code and if they do would they bet their job that all their code was error free first time around?
systemd'oh! DNS lib underscore bug bites everyone's favorite init tool, blanks Netflix
Biting the hand that feeds IT © 1998–2017