Closed Bug 445825 Opened 17 years ago Closed 16 years ago

Computer freezes repetitively when using Firefox 3

Categories

(Firefox :: General, defect)

Product:
Firefox
Component:
General
Platform:
x86
Windows XP
Type:
defect
Priority:
Not set
Severity:
critical

Tracking

()

Status:
VERIFIED INVALID

People

(Reporter: g.churchill, Unassigned)

Details

User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1.16) Gecko/20080702 Firefox/2.0.0.16 Build Identifier: Firefox 3 Since upgrading to Firefox 3 earlier this week, my computer freezes repetitively. The freezing is initiated by a mouse click which renders both mouse and keyboard inoperative. This happens on any website including Firefox home site. When I tried IE7, I had no problems with freezing. Right now I've switched back to Firefox 2 and all seems to be ok. I notice the same problem being reported by other users on a number of online forums. Reproducible: Always Steps to Reproduce: 1. Use Firefox to access the internet 2. Usually freezes within 2-3 minutes. 3. Sorry about that, I am not sure what theme was being used........I think it was the default but didn't pay close attention to that.
Component: Software Update → General
QA Contact: software.update → general
Summary: Computer repetitively when using Firefox 3 → Computer freezes repetitively when using Firefox 3
Did you whole Operating system froze ? You can not get the task manager with ctrl+alt+del ?
I can confirm this behavior. I have not been able to replicate it in any meaningful way. What I can tell you so far is that it seems to happen immediately after a disk access (I hear HDD activity). First, I will notice Firefox has locked up. Trying to click on some tabs or scroll the page will make the Windows (Not Responding) message appear in the title bar. Shortly after that the entire system freezes. The system is completely unresponsive - even the clock stops.
Status: UNCONFIRMED → NEW
Ever confirmed: true
That is correct, the whole operating system freezes so that you can not access the task manager with the keyboard.
That looks like a OS bug for me or one of your drivers kills the system. Could graphic card, Firewall or AV drivers. Firefox doesn't install system drivers and using 100% still allows to open the taskmanager. Either using all available memory or using all GDI resources could trash the OS a little bit but it should not kill the whole os. I don't think this is fixable by us (and it will not in Firefox:General and more informations).
I'm curious to know what forms the basis of your "OS bug" diagnosis. I would be even more curious to know what it is that is peculiar to FF3 that uncovered this bug that previous versions of FF did not uncover. Prior to FF3 I could run about 28 - 40 days between reboots, but the lockups became daily with the installation of FF3. I've experienced the same issue on the office machine, but much less frequently. There is definitely some relationship to FF3, but it's too early to know whether FF is the cause or if it has exposed another product's bug.
It's not to early. A userspace application can not easy crash a modern OS, that's for example the difference between windows95/98/ME, MAc OS classic and NT/2k/2003/XP/Vista, OS X, Linux. Yes, your lookups could be releated to Firefox3. Firefox3 could trigger a bug in the OS (kernel) or in one of your used drivers running in the kernel mode (graphic card, firewall driver).
You are assuming that a system level driver is causing the issue. It may or may not be the case. I don't think there's enough information to warrant that assumption yet.
No,i assume that a userspace application can not lockup a proteced-mode OS. They are enough information to warrant that.
use kd/ntsd/windbg and a second computer and chase down the problem. it isn't our fault. please don't waste our qa's time. you could try booting *windows* into safemode and using firefox (also in safe mode). if the problem goes away, then it's a driver or other program that loads a driver which doesn't get triggered when you use safe mode. figure out which and then complain to your vendor. if it does, then it's probably a file system, or antivirus driver (possibly a crypto layer if you have such a thing).
Status: NEW → RESOLVED
Closed: 16 years ago
Resolution: --- → INVALID
Final comment as I've found the problem but first of all I have to say I'm disappointed in 'timeless' response. I have no doubt that this individual possesses greater computer expertise than I do but I'm not a novice either. It was obvious from the start that the problem wasn't solely in FF3's design as numerous people were using FF3 without any problem. But FF2 had been working fine and the problem only arose when FF3 was downloaded. Rather than flippantly dismissing the issue, I'd expect customer service to recognize there must be a connection with that. The CMA approach is far from professional. For any who have the problem or may encounter it in the future, my problem was that the two COM+ services in Windows were set to "disabled" ( I have no idea how that came to be) and when reset to 'manual' my issue was resolved. This is done through the menu: start......control panel.....performance and maintenance.....administrative tools.....services. Once the services screen opens up, click on COM+ Event System and set 'startup type' to 'manual'. Repeat the resetting to manual with COM+ System Application. Yes the problem was within the Windows settings on my computer but there is something different between how FF2 and FF3 interact with my Windows settings.
Resolution: INVALID → FIXED
None of the COM+ services are disabled on my machine. Please don't change the status of bugs to "FIXED". Only the module owners should do that.
Status: RESOLVED → REOPENED
Resolution: FIXED → ---
invalid is the right solution. Disabling this service seems to hang the whole OS and that is not our fault, it could be a buig in windows itself that it doesn't handle such a disable service.
Status: REOPENED → RESOLVED
Closed: 16 years ago → 16 years ago
Resolution: --- → INVALID
Is it normal procedure to mark a bug as invalid when you do not know the cause? Or, is this being attributed to the other poster's COM+ issue? If that's the case, I will open another FF3 freeze bug.
It is the normal procedure to mark a bug as invalid if it's sure that this the issue is not our (Firefox) fault like in this case. A whole system freeze is ALWAYS not our fault by definition. That would be only different on windows9x/ME but this OS is not supported anymore. You can debug your system with microsoft tools and a second PC on the serial console (AFAIK) and then send this info to Microsoft or in the case of a driver fault send it to the driver author. That the reporter of this bug found the issue on his system makes no difference for us but I'm sure he is happy to found problem. There is only one exception for this, if Firefox allocates all memory. This can look like a system freeze but it's not. You get problems opening the taskmanager because the taskmanager needs some RAM.
Perhaps I'm off base, but I'm thinking the goal shouldn't be to make sure you're not at fault, and instead it should be to see if there's anything that can be done so that Firefox doesn't cause the user's computer to freeze. No?
As an aside, I've been able to cure the freezes so far (about a week) by making sure I close Firefox every time I step away from the computer.
It's impossible to add workarounds for broken systems for every user with such a problem and as you can even read in this bug, they are all different. It's also possible that a someone with a broken RAM get freezes only with Firefox because Firefox may use the same RAM area on his system. How should we debug such a problem ? A OS freeze debug means that you have to use a second system attached to the first one.... Feel free to analyze your issue and come back if you know what's going wrong. But that would mean that you know why your OS is failing and can fix your system instead. There is one exception: If a known OS problem hits all people. We did that for example for a OS X 10.4 bug. But there is no general OS problem because we would have thousands of reports in that case. Bugs that are not our fault are in general invalid, examples are: Plugin crashes, addon causes a firefox hang, system freeze/crash/BSOD with Firefox. for comment#17: screensaver activated, Energy saving (hhd spindown) activated ?
It's true that you can't fix each person's problems that are unique to their system. Dismissing this problem as such has the appearance of a cop-out. *ALL* bug reports will only be filed by one person and could be equally dismissed. Even the very worst of bugs will only be filed by a small subset of users. That doesn't mean the problem is limited to that subset. In this case, it seems that there is a significant, if small, number of users affected. I don't understand your eagerness to dismiss the issue without knowing how many are affected, under what conditions, and what might be triggering it. At the very least I would think you would want to find out so that others could benefit. To answer your question, yes this machine suspends to RAM after an hour of inactivity. However, the freezes do not appear related as they don't happen right after the machine wakes up. I could try leaving FF on while the machine stays running (no APM or ACP), but the machine eats about 4 kWh a day. I'm reluctant to burn through 30 kWh to test it at this point.
I'm near every day in bugzilla as my time permits it. I saw 2-4 OS freeze bugs and a few more crash bugs in the FF3 timeframe. There are enough users who are filing bugs, there are enough examples about that in bugzilla. All OS freezes are different issues, the reporter of this bug got it with a disabled service, you get it when firefox is idle, another one is getting it when opening pdf files. I don't understand why you want something getting fixed on the firefox side that is unique to your system and that is not a firefox bug. I can understand that you want your issue fixed but bugzilla is not a problem solver. And don't forget that it's not possible to debug your OS freeze without sitting in front of your system, read for example http://www.wd-3.com/archive/RemoteDbg.htm If there would be an easy way to find out what is causing this, I or others would help but it's difficult to debug such a OS freeze and impossible if you don't sit in front of that freezing system.
(In reply to comment #20) > I saw 2-4 OS freeze bugs and a few more crash bugs in the FF3 timeframe. What do you believe that translates into actual numbers of users experiencing the issue? > There are enough users who are filing bugs, there are enough examples about > that in bugzilla. Exactly. If each bug was dismissed out of hand, not much would be fixed. I agree that this is pretty much what has been happening since 1997 when I started here, but that doesn't kill my hope that maybe it won't always be that way. > All OS freezes are different issues, the reporter of this bug got it with a > disabled service, you get it when firefox is idle, another one is getting it > when opening pdf files. I get that one too, but that's an Acrobat issue and it doesn't freeze the system, just FF. > I don't understand why you want something getting fixed on the firefox side > that is unique to your system and that is not a firefox bug. I don't want it fixed. I want to find out what it is first. Then, if it's something that CAN be worked around by FF, it would be nice if that could be done. Of course, there's no obligation for Mozilla Corporation to do that just like there's no obligation to help someone up who has just fallen. Sometimes people do things because they want to help instead of being part of the problem. > I can understand that you want your issue fixed but bugzilla is not a problem > solver. > And don't forget that it's not possible to debug your OS freeze without sitting > in front of your system, read for example > http://www.wd-3.com/archive/RemoteDbg.htm > If there would be an easy way to find out what is causing this, I or others > would help but it's difficult to debug such a OS freeze and impossible if you > don't sit in front of that freezing system. It's not impossible. There are other ways to find the cause if issues other than attaching a debugger. Good, old-fashioned investigation and deductive reasoning can go a long way.
(In reply to comment #21) > (In reply to comment #20) > > I saw 2-4 OS freeze bugs and a few more crash bugs in the FF3 timeframe. > > What do you believe that translates into actual numbers of users experiencing > the issue? There is no "this" issue. Users who getting the exact same os freeze are very low. > > There are enough users who are filing bugs, there are enough examples about > > that in bugzilla. > > Exactly. If each bug was dismissed out of hand, not much would be fixed. I > agree that this is pretty much what has been happening since 1997 when I > started here, but that doesn't kill my hope that maybe it won't always be that > way. No, i mean if many users hits the same problem, then you get many bug reports about such an issue and we have examples for such bugs which affects many people and I don't mean os crash/freeze bugs. We had only a os crash with somes images on windows9x with nvidia (?) drivers that affected many people. Many people had this problem and no workaround was added in Gecko. > > > All OS freezes are different issues, the reporter of this bug got it with a > > disabled service, you get it when firefox is idle, another one is getting it > > when opening pdf files. > > I get that one too, but that's an Acrobat issue and it doesn't freeze the > system, just FF. No, there is a user who gets an OS freeze if he opens a PDF file. That is not our fault, it's a system/driver/hardware fault. That is the same thing that you call it an Adobe bug if the pdf plugin only hangs Firefox. > > I don't understand why you want something getting fixed on the firefox side > > that is unique to your system and that is not a firefox bug. > > I don't want it fixed. I want to find out what it is first. Then, if it's > something that CAN be worked around by FF, it would be nice if that could be > done. Of course, there's no obligation for Mozilla Corporation to do that just > like there's no obligation to help someone up who has just fallen. Sometimes > people do things because they want to help instead of being part of the > problem. To summarize: You want a workaround in Firefox because there is an unknown problem with your system that nobody else gets and which is impossible to debug ? > It's not impossible. There are other ways to find the cause if issues other > than attaching a debugger. Good, old-fashioned investigation and deductive > reasoning can go a long way. Great, then start with that and if you found the reason of your system freeze then post it here.
(In reply to comment #22) > There is no "this" issue. Users who getting the exact same os freeze are very > low. I would like access to the data you are using to make this assessment. I have no data to indicate that there is no freezing issue as it has been reported by more than a couple of individuals. > We had only a os crash with somes images on windows9x with nvidia (?) drivers > that affected many people. Many people had this problem and no workaround was > added in Gecko. Failure to perform in the past isn't an excuse for failure to perform in the present. Why stop there? I would prefer that the bar for Mozilla Corporations customer service be set higher than, "hey, we told you to get f'ed before, so we can do it now too." > To summarize: > You want a workaround in Firefox because there is an unknown problem with your > system that nobody else gets and which is impossible to debug ? Is it your position that the other parties in this bug are just my alter egos? Or that every person posting about freezing in FF on MozillaZine are my alter egos also? If not, then your position that it's a problem "nobody else gets" is patently absurd. > Great, then start with that and if you found the reason of your system freeze > then post it here. That's my intention. It's just hard to keep the signal strong in the noise of FF cheerleaders trying to dismiss an issue they have no knowledge of by making unsubstantiated allegations. Having a discussion with you is pointless because you are emotionally invested in sticking to your conclusions despite any evidence to the contrary.
For those of us out here who are having problems utilizing FF3 the perception that comes across from Mozilla support is that if the problem isn't totally within FF3 then thre is no problem (can you imagine what message a newcomer reading this ongoing dialogue stream would get?). I think the odds are that the freezing problem doesn't lie totally with FF3 as many are using it without an issue. And it appears that there are more than one specific 'alignment' that triggers the freezing. (for example, mine just happened to be the COM+ services). Personally I think the focus should not on pointing fingers to fix 'blame' but should be on helping those who have a problem. I don't see that as meaning Mozilla should be trouble shooting individual computers but at the minimum Mozilla could host a repository for us who have or had a problem with one of their products. If that was done, then people would soon want to come to the website to either learn or to contribute. As well, if a pattern emerges, it may highlight weaknesses within FF3 that Mozilla could attack. Keep this in mind, we are only taking the time to write and report because we WANT to use FF3. But for every one of us who do make the effort how many more users out there just shrug their shoulders and move on to another browser? I doubt that every person that has had an issue with FF3 has reported it......that just isn't human nature. Mozilla might consider that they may be missing an opportunity to assemble knowledge from many of us who care about their products.
George: The point is the you get no support here in bugzilla. Bugzilla is something that you usually not see in the public for commercial products. We have http://support.mozilla.com for support related questions, including a Forum.. Support questions are just invalid in bugzilla because bugzilla si a tool for the developer only to track their work. I help here the whole time but it's not possible to help with a system freeze, there is no way with debugging tools for a normal user to find the problem. That means that there is no way for support. Updating the whole system including all drivers is the only way I would try if i would have the same problem. I could ignore such bug reports with the result that nobody comments in the bug and the bug stays unconfirmed with no comment after x months/years. I prefer to tell the user the truth, that there is no help/support for such a problem is available and that this problem will and can not be fixed in firefox. Ignoring such bugs would be a much easier way for me because the user usually tend to protest with the result that you can see in this bug report. Don't forget that many of people in bugzilla are just users like you and we don't write one bug report, we comment in many bug reports. I commented on over 9500 bug reports in 8 years to give you an impression.
Status: RESOLVED → VERIFIED
I'm curious about what procedure or testing was performed in order to verify that this issue does not involve Firefox code.
You need to log in before you can comment on or make changes to this bug.