it amazes me how much misinformation is still going around. Peoples conceptions about how or why they get virii infecting their systems. Since one of the primary focuses of my job is remove malware from client computers, I've become exceedingly good at it. The stuff that amazes me is when people wonder why they have to pay for a service to remove the virus that's now stopping them from getting online.
There are a few major schools of thought on this, I'll go over three, starting with a common one, but not one I'd like to focus on today.
"You're supposed to be providing a service. I can't access that service. This should be free!" ... Since our positioning is as PART of the ISP's infrastructure... (factually not the case)... they believe that if for any reason, they cannot get online, it falls into our (the ISP's) job to fix whatever the issue may be. No exceptions. Unfortunately, we only guarantee internet coming out of the back of OUR modem. If YOUR equipment can't receive the signal and utilize it, well, that falls into our department as much as your broken TV falls into the task of your Cable Co.'s jurisdiction to fix it. AKA, not.
This next one is super common...
"But I have XYZ Security software from you to prevent this type of thing!" ... Okay. Just because you got a virus while using a service we re-sell on behalf of another company, does not entitle you to use a paid service for free because something got past it. If you would've read the service agreement, I'm sure you would have seen a 'best effort' clause... stating that the protection is not 100%, that they can only defend against KNOWN virii. Hense, if the virus is NEW, or not yet known, you can still get it.
I love that one.
... not.
This last one is less common but just as irritating:
"I get internet through you and therefore I got the virus from you, so it's your obligation to fix it" ... okay. You didn't get your virus from us, you got it from the INTERNET. There are a lot of people and places on the internet, and scanning all traffic for virii is an ASTRONOMICAL TASK!! We're not going to do that, nevermind the fact that even if we did, it wouldn't be perfect (see previous point). YOU were the one dumb enough to click the link and run the software, on YOUR equipment... the internet works fine, you're just a dumbass. Just because we provide you with a connection between you and the web doesn't make it our responsibility what you do with that.
Enjoy your virus.
Anyways, I had to rant there briefly, I hope noone minds. Play safe, and keep away from those damn dirty virii.
Monday, November 1, 2010
Saturday, October 2, 2010
Internet Explorer Woes
As time rolls on, I sometimes get to wondering why people still use IE. It's simple enough I guess: buy a computer, use the internet. This happens to be Internet Explorer, and that's fine for most people.
Anyone I know, who seems to think they know anything about computers at all, is using Firefox. A good choice. I use Chrome as my primary browser of choice, and fall back on FF for anything that Chrome seems to choke on (not much). if that doesn't work, I'll fall back to IE8, just to see if it will work, and if that doesn't fix the problem the website is having, I usually chalk it up to webpage design and move on.
HOWEVER, IE has so many quirks and quarks.
Working in IT, it's plainly obvious that IE relies on jscript.dll for the javascript function calls. It's also plainly obvious that the Java runtime included with Windows/IE is terrible, at best. It's almost always replaced with Sun Microsystem's JRE. It's also plainly obvious that without Javascript, almost all flash will not function, since the calls made to make the Flash work, is largely JScript fueled.
There lies your three most problematic components of IE. Now, I'm sure FF isn't terribly better, but seems so, because not as many inexperienced users (who are getting viruses and messing up settings) are using it.
So, if JScript works, and Java works and Flash works, you usually have a good-enough browser for anyone.
Here's the problem. it seems, that 64-bit dll's and 32-bit applications don't mix well.
In Microsoft's "infinite wisdom", Internet Explorer 8, 32-bit and 64-bit, both use the same registry keys for normal operation, which streamlines settings. If IE8 32-bit gets setup for a proxy, then the 64-bit version will also use that same proxy.
The problem lies in the fact that once the 64-bit jscript.dll file takes over the registry key, you're basically FUBAR for getting javascript to run in 32-bit IE. While Javascript will run in the 64-bit IE, nothing else really works well, or at all.
Sure, there's a 64-bit workaround for Flash, and a 64-bit JRE from Sun, but installation is painful at best and nothing, besides those three things would ever work with the browser, probably for many years (possibly ever).
This has all become painfully obvious to me since I had a customer at work who has a copy of Windows 7, Home Premium, x64, with both 32 and 64 bit Internet Explorers, installed.
Personally, I havn't found a good solution for the customer. I was excited to give it a good try today when I saw the chat session (at the time, upwards of 30 hours connected), drop into the queue. I'll also be looking for them tomorrow and the next day, hoping to pick up their chat. (I may even give them a call back to try to resolve it).
I suspect reassigning the jscript.dll to the 32-bit version would fix the issues they're having... I would just need to make sure they won't accidentally drop into the 64 bit version.
Two questions remain for that: first, where the heck is the jscript.dll reference in the registry for IE? second, would the system even have the 32-bit jscript.dll installed?
I'm not sure of either, but I'm excited to try.
Anyone I know, who seems to think they know anything about computers at all, is using Firefox. A good choice. I use Chrome as my primary browser of choice, and fall back on FF for anything that Chrome seems to choke on (not much). if that doesn't work, I'll fall back to IE8, just to see if it will work, and if that doesn't fix the problem the website is having, I usually chalk it up to webpage design and move on.
HOWEVER, IE has so many quirks and quarks.
Working in IT, it's plainly obvious that IE relies on jscript.dll for the javascript function calls. It's also plainly obvious that the Java runtime included with Windows/IE is terrible, at best. It's almost always replaced with Sun Microsystem's JRE. It's also plainly obvious that without Javascript, almost all flash will not function, since the calls made to make the Flash work, is largely JScript fueled.
There lies your three most problematic components of IE. Now, I'm sure FF isn't terribly better, but seems so, because not as many inexperienced users (who are getting viruses and messing up settings) are using it.
So, if JScript works, and Java works and Flash works, you usually have a good-enough browser for anyone.
Here's the problem. it seems, that 64-bit dll's and 32-bit applications don't mix well.
In Microsoft's "infinite wisdom", Internet Explorer 8, 32-bit and 64-bit, both use the same registry keys for normal operation, which streamlines settings. If IE8 32-bit gets setup for a proxy, then the 64-bit version will also use that same proxy.
The problem lies in the fact that once the 64-bit jscript.dll file takes over the registry key, you're basically FUBAR for getting javascript to run in 32-bit IE. While Javascript will run in the 64-bit IE, nothing else really works well, or at all.
Sure, there's a 64-bit workaround for Flash, and a 64-bit JRE from Sun, but installation is painful at best and nothing, besides those three things would ever work with the browser, probably for many years (possibly ever).
This has all become painfully obvious to me since I had a customer at work who has a copy of Windows 7, Home Premium, x64, with both 32 and 64 bit Internet Explorers, installed.
Personally, I havn't found a good solution for the customer. I was excited to give it a good try today when I saw the chat session (at the time, upwards of 30 hours connected), drop into the queue. I'll also be looking for them tomorrow and the next day, hoping to pick up their chat. (I may even give them a call back to try to resolve it).
I suspect reassigning the jscript.dll to the 32-bit version would fix the issues they're having... I would just need to make sure they won't accidentally drop into the 64 bit version.
Two questions remain for that: first, where the heck is the jscript.dll reference in the registry for IE? second, would the system even have the 32-bit jscript.dll installed?
I'm not sure of either, but I'm excited to try.
Sunday, September 19, 2010
usb.brando.com
The following is an open letter that I've sent to brando, my correspondence with them so far has been sub-par, primarily with people who don't seem to have a concept of the English language, never mind any idea how USB charging works, or the nuances with the differences in voltages.
Dear Brando,
I purchased your USB Wrist Band battery pack, upon receiving it, I was thoroughly impressed with the overall design, and comfort of the device, after charging I was eager to start using it.
Turns out, that the unit you sold and shipped me was defective, only outputting 4.0 of the rated 5.5V (according to your webpage).
I will not be returning this product for a refund, since shipping costs would far exceed the original cost of the product (35 USD), and I will thoroughly ensure that noone I know, or have any contact with, ever buys anything you sell, since, so far, the only responses I've gotten from your customer service department have been bullshit.
So you know, the 4.0V output doesn't even push any of my USB devices into recognising that a power supply is attached, nevermind charge anything.
Thanks for nothing.
Dear Brando,
I purchased your USB Wrist Band battery pack, upon receiving it, I was thoroughly impressed with the overall design, and comfort of the device, after charging I was eager to start using it.
Turns out, that the unit you sold and shipped me was defective, only outputting 4.0 of the rated 5.5V (according to your webpage).
I will not be returning this product for a refund, since shipping costs would far exceed the original cost of the product (35 USD), and I will thoroughly ensure that noone I know, or have any contact with, ever buys anything you sell, since, so far, the only responses I've gotten from your customer service department have been bullshit.
So you know, the 4.0V output doesn't even push any of my USB devices into recognising that a power supply is attached, nevermind charge anything.
Thanks for nothing.
Wednesday, September 8, 2010
Women in Technical Fields
This article was posted both on Twitter and on Facebook by the lovely and talented Felicia Day, who, a friend of mine had the recent opportunity to meet at FanExpo, which I wasn't even aware was happening... Next year... Next year.
Aside from all that jealousy about what my friend was able to see/do at FanExpo without me, Felicia, interestingly enough, is making quite the point. If you read the article (and admittedly, I was agreeing with the article so much I never quite made it to the end), it's describing a situation where we're almost forcing women into technical fields. I've always felt forcing an issue is wrong, especially of this nature.
This actually reminds me of an episode of the show "rescue me" that I recently re-watched where the lead character "Tommy" (Denis Leary) goes on a rant regarding racial equality in the FDNY; he makes the very important point that maybe it isn't the FDNY that's not hiring the blacks, jews or whatever into the department, but rather that it's the blacks, jews or what-have-you that simply don't want the damn job.
I would apply that logic here, while I agree with this article. While there's an opportunity here to promote women in the workplace, and certainly there are many women who have the physical capability of working in IT/Tech, as well as the brain power to do it as well (if given proper motivation, interest, etc.), however, just simply do not take enough interest in HOW the Internet works, to actually get into the inner-workings of the field. If these people, regardless of their gender, do not have the interest in the job, then are they really the right people for the job? Why should we, or anyone for that matter, be putting them into a position where they are, perhaps, doing something they may not enjoy?
To be totally blunt, I've always had the mindset that, I don't care if you're black, white, Jewish, Spanish, Mexican, Egyptian, Chinese, Cuban or Hungarian; if you can do the job, and do it well, then you should be doing the job. It's all about the right person for the right job, and getting the job done; It's never been WHO does the job, not to me. And that goes for every profession. Everything from Firefighters, to IT and beyond. If you want the job, and you can do it, then go for it.
I don't see why anyone would want anything different from that.
Aside from all that jealousy about what my friend was able to see/do at FanExpo without me, Felicia, interestingly enough, is making quite the point. If you read the article (and admittedly, I was agreeing with the article so much I never quite made it to the end), it's describing a situation where we're almost forcing women into technical fields. I've always felt forcing an issue is wrong, especially of this nature.
This actually reminds me of an episode of the show "rescue me" that I recently re-watched where the lead character "Tommy" (Denis Leary) goes on a rant regarding racial equality in the FDNY; he makes the very important point that maybe it isn't the FDNY that's not hiring the blacks, jews or whatever into the department, but rather that it's the blacks, jews or what-have-you that simply don't want the damn job.
I would apply that logic here, while I agree with this article. While there's an opportunity here to promote women in the workplace, and certainly there are many women who have the physical capability of working in IT/Tech, as well as the brain power to do it as well (if given proper motivation, interest, etc.), however, just simply do not take enough interest in HOW the Internet works, to actually get into the inner-workings of the field. If these people, regardless of their gender, do not have the interest in the job, then are they really the right people for the job? Why should we, or anyone for that matter, be putting them into a position where they are, perhaps, doing something they may not enjoy?
To be totally blunt, I've always had the mindset that, I don't care if you're black, white, Jewish, Spanish, Mexican, Egyptian, Chinese, Cuban or Hungarian; if you can do the job, and do it well, then you should be doing the job. It's all about the right person for the right job, and getting the job done; It's never been WHO does the job, not to me. And that goes for every profession. Everything from Firefighters, to IT and beyond. If you want the job, and you can do it, then go for it.
I don't see why anyone would want anything different from that.
Sunday, September 5, 2010
Telus Milestone rooting/flashing
I don't want to post much about personal stuff but this is far too interesting/helpful than to keep quiet about. So here goes.
As many will know, I purchased a Motorola Milestone from Telus a while back. When I got it, it was stock updated to Android 2.1 (originally released with 2.0.1), and since then, there was an update to 2.1, done OTA from Telus to correct call quality issues.
I started, very happy with the phone, but after dealing with the brief (but noticable) slow downs and delays, plus the possibility of never being updated to 2.2, plus the possibility of never having the Flash player or any of the 2.2 official updates, I started looking around.
To be clear, the Milestone here in Canada is supposed to be getting 2.2 (Froyo) in Q1 2011, still quite a ways away, I'm impatient.
When I started, I began by 'rooting' my phone... This process was easy, using a paid, but very easy to obtain rooting software "Universal Androot" from the market... After using this, I unsuccessfully attempted to use ROM Manager to swap ROMs, which resulted in paying for software that doesn't actually work on the Milestone. I don't mind, I'll support good developers, regardless.
The best way to root, and set up a Milestone, running the newest Telus ROM, in preparation to reflash it, is to use RSD Lite (available almost everywhere) and get the milestone "vulnerable recovery" SBF file. The installation of this SBF will allow your milestone to boot an update.zip that is not signed by a certified authority (AKA Motorola). After you have that basic file and app prepared, before flashing you need to setup an Update.zip and nandroid on your SD card on the device. I used Open Recovery on this.
In theory, you could use many of the various options for Recovery software... Open Recovery was my choice and I feel it's been a good one.
once open recovery is situated with the update.zip file on your SD card, you're free to apply the vulnerable recovery sbf to your phone using RSD Lite. this process is highly automatic, and only requires you to have the specific files, as well as your phone connected to the system via USB.
but hold on, don't let that phone start back up. Everytime the phone starts, with the Telus ROM, it checks that bootloader's checksum. if the checksum does not match the checksum on file, it reflashes the bootloader to spec. Before this happens, you'll need to boot into open recovery to make a change.
After the phone restarts after applying the vulnerable recovery patch (it will say update OK, then shutdown for a restart), you want to hold your phone's camera button as it's starting, you'll get to the phone with the exclamation point screen (you'll know it when you see it), at that point, hold volume UP, and press the camera button again, you'll get a menu.
Using your d-pad, select "apply update.zip" (a process you'll have to repeat to get into open recovery, all the time).
Open Recovery loads. browse over to console. At the console run the following command:
rm -f /system/etc/install-recovery.sh
this should delete the specific file that will reflash the bootloader/recovery and prevent you from getting back to the nandroid/open recovery.
after that, you're rooted with recovery, leave update.zip, nandroid, and open recovery on your SD card for future reference, and download a ROM of your choice. it may be a good idea to use nandroid to do a back up of your current ROM before deciding to wipe and flash a new one.
Whenever flashing a new ROM delete all user data, dalvik and cache. This should put the phone into a very RAW/Fresh state for the new ROM and prevent a lot of problems you may run into.
Enjoy, happy flashing.
As many will know, I purchased a Motorola Milestone from Telus a while back. When I got it, it was stock updated to Android 2.1 (originally released with 2.0.1), and since then, there was an update to 2.1, done OTA from Telus to correct call quality issues.
I started, very happy with the phone, but after dealing with the brief (but noticable) slow downs and delays, plus the possibility of never being updated to 2.2, plus the possibility of never having the Flash player or any of the 2.2 official updates, I started looking around.
To be clear, the Milestone here in Canada is supposed to be getting 2.2 (Froyo) in Q1 2011, still quite a ways away, I'm impatient.
When I started, I began by 'rooting' my phone... This process was easy, using a paid, but very easy to obtain rooting software "Universal Androot" from the market... After using this, I unsuccessfully attempted to use ROM Manager to swap ROMs, which resulted in paying for software that doesn't actually work on the Milestone. I don't mind, I'll support good developers, regardless.
The best way to root, and set up a Milestone, running the newest Telus ROM, in preparation to reflash it, is to use RSD Lite (available almost everywhere) and get the milestone "vulnerable recovery" SBF file. The installation of this SBF will allow your milestone to boot an update.zip that is not signed by a certified authority (AKA Motorola). After you have that basic file and app prepared, before flashing you need to setup an Update.zip and nandroid on your SD card on the device. I used Open Recovery on this.
In theory, you could use many of the various options for Recovery software... Open Recovery was my choice and I feel it's been a good one.
once open recovery is situated with the update.zip file on your SD card, you're free to apply the vulnerable recovery sbf to your phone using RSD Lite. this process is highly automatic, and only requires you to have the specific files, as well as your phone connected to the system via USB.
but hold on, don't let that phone start back up. Everytime the phone starts, with the Telus ROM, it checks that bootloader's checksum. if the checksum does not match the checksum on file, it reflashes the bootloader to spec. Before this happens, you'll need to boot into open recovery to make a change.
After the phone restarts after applying the vulnerable recovery patch (it will say update OK, then shutdown for a restart), you want to hold your phone's camera button as it's starting, you'll get to the phone with the exclamation point screen (you'll know it when you see it), at that point, hold volume UP, and press the camera button again, you'll get a menu.
Using your d-pad, select "apply update.zip" (a process you'll have to repeat to get into open recovery, all the time).
Open Recovery loads. browse over to console. At the console run the following command:
rm -f /system/etc/install-recovery.sh
this should delete the specific file that will reflash the bootloader/recovery and prevent you from getting back to the nandroid/open recovery.
after that, you're rooted with recovery, leave update.zip, nandroid, and open recovery on your SD card for future reference, and download a ROM of your choice. it may be a good idea to use nandroid to do a back up of your current ROM before deciding to wipe and flash a new one.
Whenever flashing a new ROM delete all user data, dalvik and cache. This should put the phone into a very RAW/Fresh state for the new ROM and prevent a lot of problems you may run into.
Enjoy, happy flashing.
Saturday, August 21, 2010
I'm at a loss
It's not a frequent thing when I'm at a loss on how to fix something, especially with anything electrical.
It's usually a cold day when I give up on something, and I'm usually the last to give up, and I usually give up for good reason. I only had one false-positive, in relation to a bad device, ever. I rid myself of a Linksys LNE100TX, which is a very good NIC, for a D-Link DFE-430TX, also a good NIC, and when I returned back and installed the D-Link, I found the source of my problem, which persisted through NIC changes, however I had garbage-d the Linksys already and was unable to retrieve it.
I've done crazy things with computers, from building something from scratch, to fixing the otherwise unfix-able, that others have given up on as "too frustrating", all the way to re-soldering resistors onto the power-riser on a laptop that wouldn't power-up (it worked afterward by the way).
My point here is, I'm usually the last to give up, and if I give up, I have a good reason. My expertise with all these things is fairly expansive, and only hits a limit when dealing with determining problems on a circuit-by-circuit basis. I have a multimeter, but I'm not always sure what information it's giving me, or how relevant it is.
I recently have been working on a friends Dell XPS m1330, which has a bad video processor (on-board on motherboard), requiring a motherboard replacement. I ordered one, it happened to be from China. The shipping time and method was reasonable, when receiving it, the packing was adequate and the board looks to be in good condition. Upon installing the board, and applying all the screws, cases, cards, thermal paste and coolers (including modifications to the original design, ergo: "copper mod"), I attempted to connect power before I powered up the unit, to test for function, and the charging light didn't come on... that's odd.
I checked, and found the power-light on the PSU wasn't lit. okay? it's plugged in, what gives?
after extensive testing, the new motherboard seems to be causing the power supply to randomly turn off. Probably a fault safety or something. I connected the system to another PSU (variable Voltage, adjustable, universal laptop power supply), and all that accomplished was to reduce the (albeit small) display on the cord, which normally shows the voltage, to a very dim jumble of non-sense data. Removing from the system, the display reset to it's lowest value, 15V, attempting a 15V connection, resulting in the same problem.
Attach either of these power supplies to the original mainboard, that works apart from the graphics being completely unusable, and the result is the same: it works.
I emailed the supplier who claims "it was tested before being sent out"... if that's the case, then sometime between you testing it, and me testing it, it stopped working.
I've handled the motherboard actually BETTER than the original one that came with the system (which I accidentally dropped on the floor from about a foot up), and the original works, the "new" board doesn't. explain that.
The supplier is willing to do an exchange, so the board will be going out on Monday.
I can't help but be a tad bit insulted by the insinuation from the supplier that I'm wrong, and the mobo works. I tested the thing, I even checked the resistance of the original mainboard's pos and neg terminals (incoming to the system), and the resistance between them is very similar (there will be variance between the two due to a tolerance in the resistance of every resistor on the board, eg. 100ohm resistor with +/- 5% tolerance could be anywhere from 95 to 105 ohms in resistance). So, in theory, they should both be very similar. I even tested it with the elbow for the universal power adapter plugged in to check for faults in the power adapter, which I would've been willing to desolder and exchange with the power connector from the original mobo, if I thought it would fix the problem, however, the resistance doesn't change, indicating no short in the connector.
I'm baffled. Apart from a major fault of a circuit that handles the current on the mainboard, that isn't obvious, I don't see why this would be happening.
Oh well, another 2-3 weeks before I can reassemble this with yet another motherboard.
It's usually a cold day when I give up on something, and I'm usually the last to give up, and I usually give up for good reason. I only had one false-positive, in relation to a bad device, ever. I rid myself of a Linksys LNE100TX, which is a very good NIC, for a D-Link DFE-430TX, also a good NIC, and when I returned back and installed the D-Link, I found the source of my problem, which persisted through NIC changes, however I had garbage-d the Linksys already and was unable to retrieve it.
I've done crazy things with computers, from building something from scratch, to fixing the otherwise unfix-able, that others have given up on as "too frustrating", all the way to re-soldering resistors onto the power-riser on a laptop that wouldn't power-up (it worked afterward by the way).
My point here is, I'm usually the last to give up, and if I give up, I have a good reason. My expertise with all these things is fairly expansive, and only hits a limit when dealing with determining problems on a circuit-by-circuit basis. I have a multimeter, but I'm not always sure what information it's giving me, or how relevant it is.
I recently have been working on a friends Dell XPS m1330, which has a bad video processor (on-board on motherboard), requiring a motherboard replacement. I ordered one, it happened to be from China. The shipping time and method was reasonable, when receiving it, the packing was adequate and the board looks to be in good condition. Upon installing the board, and applying all the screws, cases, cards, thermal paste and coolers (including modifications to the original design, ergo: "copper mod"), I attempted to connect power before I powered up the unit, to test for function, and the charging light didn't come on... that's odd.
I checked, and found the power-light on the PSU wasn't lit. okay? it's plugged in, what gives?
after extensive testing, the new motherboard seems to be causing the power supply to randomly turn off. Probably a fault safety or something. I connected the system to another PSU (variable Voltage, adjustable, universal laptop power supply), and all that accomplished was to reduce the (albeit small) display on the cord, which normally shows the voltage, to a very dim jumble of non-sense data. Removing from the system, the display reset to it's lowest value, 15V, attempting a 15V connection, resulting in the same problem.
Attach either of these power supplies to the original mainboard, that works apart from the graphics being completely unusable, and the result is the same: it works.
I emailed the supplier who claims "it was tested before being sent out"... if that's the case, then sometime between you testing it, and me testing it, it stopped working.
I've handled the motherboard actually BETTER than the original one that came with the system (which I accidentally dropped on the floor from about a foot up), and the original works, the "new" board doesn't. explain that.
The supplier is willing to do an exchange, so the board will be going out on Monday.
I can't help but be a tad bit insulted by the insinuation from the supplier that I'm wrong, and the mobo works. I tested the thing, I even checked the resistance of the original mainboard's pos and neg terminals (incoming to the system), and the resistance between them is very similar (there will be variance between the two due to a tolerance in the resistance of every resistor on the board, eg. 100ohm resistor with +/- 5% tolerance could be anywhere from 95 to 105 ohms in resistance). So, in theory, they should both be very similar. I even tested it with the elbow for the universal power adapter plugged in to check for faults in the power adapter, which I would've been willing to desolder and exchange with the power connector from the original mobo, if I thought it would fix the problem, however, the resistance doesn't change, indicating no short in the connector.
I'm baffled. Apart from a major fault of a circuit that handles the current on the mainboard, that isn't obvious, I don't see why this would be happening.
Oh well, another 2-3 weeks before I can reassemble this with yet another motherboard.
Friday, July 30, 2010
Routing tables
I've been dealing with this problem for a while, and I've only had a few instances where I *thought* it might be the problem with a specific person's internet, but I've never actually SEEN it on any system other than my own.
I work next to people that see this all the time, their solution is to Uninstall the NIC from device manager and get it to re-detect it. This is obviously not the best solution.
I'm not sure where the error comes from but it seems to primarily effect wired interfaces where you will shutdown and restart connected to the same, or a different network. I havn't been able to determine exactly WHY or WHEN it happens, so all of that is theory, but here's what happens.
The routing tables receive an incorrect entry stating that the local area adapter has a default gateway of 0.0.0.0
This entry, while not being assigned by any dhcp, is immune to release/renew attempts made by normal adapter "repair" operations.
Essentially, this entry tells the computer that every possible address on the internet should be sent out onto the local area network. That every possible IP address is local. So your computer, when connecting to Google, (for the purposes of this example, we'll use a common Linksys, NAT, private Class C network with local DNS and DHCP), you would need to go ask your DNS for Google's IP... in this case, your DNS will be an address local to you, such as 192.168.1.1 (most common router IPs), which will respond with something not-local, such as 173.194.32.104 . Your computer then checks it's routing tables and falsely determines that it should be able to connect to Google directly, which it obviously can't do... so it requests a connection and fails (on OSI layer 2) every time.
The easiest solution to the problem is to reboot, uninstall the adapter or similarly clear the routing tables using an indirect method.
The most efficient way is to do a "route delete 0.0.0.0" with added options to tell the route command to delete just the incorrect entry.
MY solution is just to perform a "route delete 0.0.0.0" which will clear ALL default routes, then do an IPCONFIG release and renew.
I don't know where this problem came from but it seems to only affect Windows Vista and Windows 7.
Hopefully that helps some of you folks out there. Happy routing.
I work next to people that see this all the time, their solution is to Uninstall the NIC from device manager and get it to re-detect it. This is obviously not the best solution.
I'm not sure where the error comes from but it seems to primarily effect wired interfaces where you will shutdown and restart connected to the same, or a different network. I havn't been able to determine exactly WHY or WHEN it happens, so all of that is theory, but here's what happens.
The routing tables receive an incorrect entry stating that the local area adapter has a default gateway of 0.0.0.0
This entry, while not being assigned by any dhcp, is immune to release/renew attempts made by normal adapter "repair" operations.
Essentially, this entry tells the computer that every possible address on the internet should be sent out onto the local area network. That every possible IP address is local. So your computer, when connecting to Google, (for the purposes of this example, we'll use a common Linksys, NAT, private Class C network with local DNS and DHCP), you would need to go ask your DNS for Google's IP... in this case, your DNS will be an address local to you, such as 192.168.1.1 (most common router IPs), which will respond with something not-local, such as 173.194.32.104 . Your computer then checks it's routing tables and falsely determines that it should be able to connect to Google directly, which it obviously can't do... so it requests a connection and fails (on OSI layer 2) every time.
The easiest solution to the problem is to reboot, uninstall the adapter or similarly clear the routing tables using an indirect method.
The most efficient way is to do a "route delete 0.0.0.0" with added options to tell the route command to delete just the incorrect entry.
MY solution is just to perform a "route delete 0.0.0.0" which will clear ALL default routes, then do an IPCONFIG release and renew.
I don't know where this problem came from but it seems to only affect Windows Vista and Windows 7.
Hopefully that helps some of you folks out there. Happy routing.
Subscribe to:
Posts (Atom)