Closed Bug 268138 Opened 20 years ago Closed 18 years ago

Extensive memory usage over time

Categories

(Firefox :: General, defect)

Product:
Firefox
Component:
General
Version:
1.0 Branch
Platform:
x86
All
Type:
defect
Priority:
Not set
Severity:
major

Tracking

()

Status:
RESOLVED DUPLICATE of bug 320915

People

(Reporter: kaeptn00, Assigned: bugzilla)

Details

Attachments

(1 file)

Collage of four examples of FF trapping real and VM
200.89 KB, image/png
Details
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.3) Gecko/20041026 Firefox/1.0RC1 Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.3) Gecko/20041026 Firefox/1.0RC1 Firefox eats an exzessive amount of memory over the time. Current browser session is about 12 hours old and now uses 730MB of memory. This is making my system and Firefox slow. Physical memory is 1GB so it is not a real problem to me. It seems to be related to not freeing the memory plugis like Quicktime, Flash and such use. Don't know if this is a duplicate bug. Sorry for not being able to give more specific results. If there's a test case let me know. Reproducible: Couldn't Reproduce Steps to Reproduce: 1. 2. 3.
Well, I'm having some problems like this. I opened www.deviantart.com and starte do browse the site and open the pictures at different tabs. +- 5 tabs that I opened and closed always and 5 more that stayed opened. I did this for some time and when I saw, firefox was using more than 300MB. I minimized it and the memory intantly dropped below 40MB, but then it was going up slowly and then fast while the window was still minimized. The memory freed while minimized is small compared to what it was. Firefox and the system start to be very slow. The OS I'm using is Windows 2000 sp4 and FireFox v1.0 Preview Release. I hope I have helped.
Hi there, running Firefox and Thunderbird on a W2K SP4, PIII 450MHz, 512MB RAM, perhaps you can soon change the OS marker even to "All", as I see similar things happening in Linux. > Don't know if this is a duplicate bug. Sorry for not being able to give more > specific results. If there's a test case let me know. It might be, there are similar ones: at least 265139 and 222953 etc. etc., just check https://bugzilla.mozilla.org/buglist.cgi?bug_status=UNCONFIRMED&bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&field0-0-0=product&type0-0-0=substring&value0-0-0=memory&field0-0-1=component&type0-0-1=substring&value0-0-1=memory&field0-0-2=short_desc&type0-0-2=substring&value0-0-2=memory&field0-0-3=status_whiteboard&type0-0-3=substring&value0-0-3=memory&field1-0-0=product&type1-0-0=substring&value1-0-0=usage&field1-0-1=component&type1-0-1=substring&value1-0-1=usage&field1-0-2=short_desc&type1-0-2=substring&value1-0-2=usage&field1-0-3=status_whiteboard&type1-0-3=substring&value1-0-3=usage > Reproducible: Couldn't Reproduce I can, and have been able to in earlier Mozilla versions too. > Steps to Reproduce: 1. Open an app where you can see memory and CPU usage 2. Note memory usage by FF when starting to use it 3. Keep check on the amount of memory it's munching over time It's in the trend, not in the single web page cases. Rather alarmingly, memory and resource munching seems to be a general open source problem, that really deserves to be adressed seriously. Even new computers often come with just 256 MB, so even such a new computer would choke quickly from running, say, OOo, Firefox and Thunderbird. I do just that right now, and they eat 225 MB (!) between themselves. Starting e. g. The GIMP adds another 19 MB, _before_ loading any graphics. This is on Win2K, haven't yet checked the impact on e. g. SUSE 9.1 + KDE or the like, but I know memory to be an issue there too, as I experience it on an LTSP system. Thunderbird is a true memory muncher too, munching up a fat 30 MB just for starting up, stalling the whole system with 100% CPU usage on startup and closing. Don't know if that should go into another bug report. Indeed, it seems all current Mozilla products are more or less affected. The other software isn't "your problem", but the compound effect of your software in a common environment of other (open source) software is, IMHO. And then there's the CPU munching too, where it seems Firefox is a real culprit, munching 80% and more of my CPU resources (not just relative to other, absolute usage), sometimes making it peak at a potentially system-stalling 100% total. I beleive the priority for this should be set to a fairly high level. Isn't there something to be done about this, each contributing their piece of the puzzle?
Ok ... I see that I am not the only one having this problem. Could please someone of the staff make a comment on this? So please take my excuse for reassigning ;) BTW, Why can't I reassign this? The system does not let me. It should, shouldn't it? I understand that XUL applications eat more memory than other apps due to the nature (of the beast ;)). Any ideas beyond that which could explain why memory usage grows and grows and grows over time? I see the duplicate bugs and changed the OS field to "All". I am not going to resolve this one here as a duplicate of really old and unassigned bugs ;)
OS: Windows XP → All
Version: unspecified → 1.0 Branch
hello, i have the same problem under windows 2000. after using firefox some hours and open many new tabs and close them again, the browser need more then 300 mb ram also when i close all windows/tabs since one. i must restart it after 24 to 48 hours. goodby
I've found that memory usage increases greatly after downloading files unless the Download box is 'cleaned up' after they finish downloading. Memory usage does not necessarily correspond to how large the files are, as downloading 20 or 30 images will use 50+ MB of ram, despite each image only being 100-200kb in size. The browser will stop responding and sometimes crash after reaching a certain amount of memory usage. CPU usage also escalates to 99% when this happens. The only way to prevent it is to 'Clean Up' your downloads as they finish if you are downloading more than a few files.
same problem here. after about two days, firefox eats up 38% of memory (showing in top). normally at startup (with the same tabs), it eats up ~20%. i don't think that it depends on downloads, cause i don't do a lot of downloads. i'm just browsing and doing a lot of tab open/close operations. my firefox also crashes after a few days, maybe caused by too much memory usage!? i'm using linux debian sid 2.6.10-mm1 here.
I think this should be a duplicate of bug 130157.
Might be. Might not be. In fact there are other bugs related to this topic which are also not makred dupes of other bugs ... this does not mean I am not willing to close this one here. Some threads mention XUL as the nasty, some tabs, some this some that. The truth must lie in there somewhere ;) Or in malloc() or what else ... My fears are since this one here has not gotten an "we are working on it" answer and the one you are refering to is ages old (opened in 2002) there's no real focus on this issue. Or is there? Text-Poll ;) Should I dupe it or not?
OS: W2K v5.00.2195 SP4 Chip:Intel P3 750MHz RAM: 512Mb Viewing Gallery sites (web sites that display images) causes VM to grow, which is expected, but when the page or tab is closed, the VM is not free'd. Even closing the browser does not free the VM, until every FF browser is closed. This is a problem, since i have to close all browsers once every 24hours. This is my normal session of browsers and you can try these to reproduce the problem: Note the VM level. 1. Open DWMX (v7.0.1), Fireworks (v7.0.2), a few Windows Explorers Note the VM level. 2. start 1 ff browser and open a few tabs to normal sites, such as yahoo mail, google and other stuff i need for access - usually about 4 or 5 tabs. Note the VM level. 3. start another ff browser and open a few tabs to forum sites and other reference sites - usually about 4 to 5 sites. Note the VM level. 4. start another ff browser, which i use for searching when i am working, like php.net, dev.mysql.com, google, etc, so this can grow from 1 to about 10 to 15 tabs. Note the VM level. 5. start another ff browser and using Flash Got (see note below to install Flash Got) i may build a gallery to search for images to use in forums, for example. Try this in Build Gallery http://www.itv-f1.com/ImageLibrary/3[1000-2000;1]_2.jpg Flash Got - installed from http://www.flashgot.net/getit You will notice the VM grow, but when you close the tab, the VM is not released, nor when the 4th ff browser is close. Only when you close all of them does the VM drop.
REPRODUCIBLE: Simply open & close a new tab repeatedly - worked for me to get memory usage up to 100MB, just by repeatedly opening and closing tabs (Ctrl + T, then Ctrl + W) in rapid succession. Anyway, I notice that almost every morning when I come in to work, VM Size reported on WinXP for Firefox is ~100MB, with exactly 3 tabs open: 1) Gmail 2) a small .java file I am browsing 3) FishEye web page When I completely close & reopen Firefox, with the same 3 pages in 3 tabs, VM size is ~15MB or so. -dave-
This is a general problem of Mozilla, not just of Firefox. I'm running Windows XP SP2 with NO virtual memory, in a system with 1GB of RAM. I'm using Mozilla 1.8b, and I'm measuring the memory usage through the "private bytes" column of Process Explorer (www.sysinternals.com). After having used Mozilla on and off for a few days for browsing and reading mail, without closing it, I closed all tabs of the browser window except one, in which I entered the URL "about:blank". I also closed the mail window and cleared the cache in the preferences. The only Mozilla window that was left was the one showing "about:blank". After this, Process Manager indicated a use of 174472 kbytes by Mozilla. Minimizing Mozilla made only a very slight change in this figure. I then closed Mozilla and started it again, and again opened "about:blank" in a single tab of a browser window. The memory usage indication given by Process Explorer was then 17588 kbytes. I think this is a clear indication of some memory management problem in Mozilla. For some special reasons, I don't use virtual memory in my system. This behavior of Mozilla means that I have to close it and restart it every few days, to prevent it from eating up too much RAM.
Intel Celery 1.6, 256Mb RAM, W2K SP4, Firefox 1.0.3. I am getting the same memory leak condition with a specific website. I have a javascript running a webcam image that auto-refreshes every interval. After running Firefox for around 24 hours with that window open, VM usage will exceed system thresholds (easily the largest memory consumption on my system. Note: IE does NOT exhibit this behavior on the same page. Page is: http://members.cox.net/griffon/thefield/webcamremote.html Have gone into this bug before with Mozilla, only to be told it has something to do with GDI handles. That fix never apparently solved the problem. I have a screen print of the VM usage if requested.
Seems to happen when a java applet is in use in a minimized screen. Environment info: Windows 2000 SP4, Firefox 1.0.3 Steps to reproduce: 1. Goto http://everyvideogame.com/nes_Mike_Tyson's_Punch-Out!!.htm 2. Start the game 3. Minimize the window After approx 30-45min of minimized memory usage is over 100K
1.0.4, Windows 2000 SP4, 512MB RAM. I experience the same thing. Closing all windows fails to release any significant amount of memory. This again is related to using Flash, and other embeddings. This is a very serious bug. Why is it not Critical?
(Sorry, that should have read "closing all windows except one fails to release any significant amount of memory".)
(In reply to comment #15) > (Sorry, that should have read "closing all windows except one fails to release > any significant amount of memory".) It keeps all its memory until the program itself is terminated.
I have this problem too! The usage climbs to around 150MB. I am on the latest Deer Park Alpha 2 from 7/20/2005. At first, this Deer Park Alpha seems to say pretty good around 26mb. But over time, with never closing the browser, it gets really high!
This bug is terrible. I also expirence this. My friends complain to me how slow it gets, and I open up task manager there and the memory usage is in the hundreds of MBs. This is not normal and should be resolved.
This behaviour has been evident for at least one year. Every nightly I try shows the same problem after a few days. The only plugin I use is SessionSaver (because FF gets bloated so I have to close down). Flash is not installed. Clean profiles are used often. This behaviour is reproducible every single time. This is the single largest problem with FF. Take a look at the bugs over time for memory issues.
I have been reading through the bug lists and variations of this are RAMPANT. I wanted to add a quick test case to reproduce the issue: 1. Activate Task Manager and select the performance tab. 2. Note the amount of 'Available' memory in the 'Physical Memory (K) pane. Alternatively and more specifically, note the used memory of 'firefox.exe' on the 'Processes' tab. 3. Open an instance of Firefox. Note the available memory. Leave this one alone. 4. Open another instance of Firefox. Note the available memory. 5. Open another instance of Firefox. Note the available memory. 6. Close the last opened instance. NOTE that the memory used by the process and by the operating system DOES NOT decrease. 7. Open some more; watch the memory use climb. Close some more; watch the memory usage stay the same. 8. Memory is completely release when ALL instances of Firefox are closed. It SEEMS that saving/downloading a .jpg here and there, or possibly hitting pages with javascript, rapidly accellerates the process of memory consumption.
(In reply to comment #20) > I have been reading through the bug lists and variations of this are RAMPANT. > I wanted to add a quick test case to reproduce the issue: > > 1. Activate Task Manager and select the performance tab. > 2. Note the amount of 'Available' memory in the 'Physical Memory (K) pane. > Alternatively and more specifically, note the used memory of 'firefox.exe' on > the 'Processes' tab. > 3. Open an instance of Firefox. Note the available memory. Leave this one > alone. > 4. Open another instance of Firefox. Note the available memory. > 5. Open another instance of Firefox. Note the available memory. > 6. Close the last opened instance. NOTE that the memory used by the process > and by the operating system DOES NOT decrease. > 7. Open some more; watch the memory use climb. Close some more; watch the > memory usage stay the same. > 8. Memory is completely release when ALL instances of Firefox are closed. > > It SEEMS that saving/downloading a .jpg here and there, or possibly hitting > pages with javascript, rapidly accellerates the process of memory consumption. > Unless maybe by means of undocumented or poorly-documented command-line arguments (which, in particular, do not appear in the output of "firefox -h |more") it is not possible to open several instances of Firefox. What you get is one or more windows operated by a single instance of the binary executable (as shown by Task Manager, where the "Processes" tab shows at most one process named "firefox.exe". (OTOH, when using Internet Explorer, each window corresponds to a different instance of the binary) This said, I also get slow but irreversible memory increase over time, which can be reclaimed only by closing Firefox completely; usually, closing and reopening it with the _same_ HTML pages in the _same_ tabs reclaims a lot of memory, let's say 300KB of a 500KB process (while the biggest non-Firefox process, typically Thunderbird or Windows Explorer, uses less than 50KB).
I just wanted to bring attention to this bug again. This is probably the #1 issue in firefox that has not been addressed yet. Ram usage grows over time to unreasonable levels. This bug has been around forever. I really hope there is a concerted effort to fix this before Firefox 2.0 is released.
(In reply to comment #22) > I just wanted to bring attention to this bug again. This is probably the #1 > issue in firefox that has not been addressed yet. Ram usage grows over time to > unreasonable levels. > I'll second that. Increasing ram consumption while effectively doing nothing is not acceptable.
(In reply to comment #22) > This bug has been around forever. I really hope there is a concerted effort to > fix this before Firefox 2.0 is released. The problem I see in this bug report is that there's no one clearly identified bug. It seems to be a mishmash of various memory problems. If you want memory problems fixed, please file good memory leak bug reports <http://dbaron.org/log/2006-01#e20060110a>, with one clearly defined issue per bug. Once you've filed them, you can mention them here. Then this bug should be closed because it's not getting anywhere.
Status: NEW → RESOLVED
Closed: 18 years ago
Resolution: --- → INVALID
(In reply to comment #24) Hmm ... don't get me wrong but setting the bug to RESOLVED/INVALID means that in the bugs lifecycle it now is "A resolution has been taken" and "The problem described is not a bug." This might be right cause it is more than one bug. Somehow I feel very sad as the blog gives me the feeling that bug reports from "uneducated" users don't really trigger the developers "hunting instinct". And they "prey" here is a rather big one in my opinion. Good to see though that after almost 2 years someone at least got annoyed by memory leakage even though it is not quiet resolved yet. Don't dare to say ever again that open source projects have shorter reaction times on severe bugs. I'll see what I can do on reporting a better bug on this but don't count on it too much ... =================== Letting my personal disappointment alone, I ask anyone reading this bug report to try to get involved in this again and follow the blogs guidelines. ===================
First, this is not a (In reply to comment #25) > Somehow I feel very sad as the blog gives me the feeling that bug reports from > "uneducated" users don't really trigger the developers "hunting instinct". And > they "prey" here is a rather big one in my opinion. The problem is that the original bug report contains no steps to reproduce a problem. How can we tell what the problem is or whether the problem is fixed? > Good to see though that after almost 2 years someone at least got annoyed by > memory leakage even though it is not quiet resolved yet. Don't dare to say ever > again that open source projects have shorter reaction times on severe bugs. If you're trying to imply there's some sort of serious bug with plug-ins, please understand that the vast majority of users don't see any problem. That's why you need to give the steps to reproduce, so we can see what the problem is. It can be as simple as: 1. Go to http://www.gamezworld.com/ 2. Click on the Foobnitz link If the memory use goes way up when we perform those steps, and the problem does not occur with other browsers, we can see what the problem is and it can be fixed. If no one explains how to reproduce the bug, don't be surprised if it's not fixed for several more years.
(In reply to comment #26) > If the memory use goes way up when we perform those steps, and the problem does > not occur with other browsers, we can see what the problem is and it can be > fixed. If no one explains how to reproduce the bug, don't be surprised if it's > not fixed for several more years. The bug is really hard to nail down in a reproducable way. The bug resides deep in the core, IMHO. This makes it much much harder for non programming savy people to actualy think of a way to make it reproducable. Oh, and some of the reports above are quiet specific, IMHO. Like the webcam guy on #12. A simple task eats memory. How more specific could it be from a users point of view? A hard to find instruction on how to enter debug mode and putting a perl source online is only half of a step. The blog may be helpful here but having people install perl is not a very convenient way. Script it to php or to cgi and write a web interface to it. It is "unconvenient" anyways, why make it even more complicated? You state that "the vast majority of people" does not see a problem here. What kind of people do you relate to here? The normal "John Doe" does not even care about memory usage. I would guess that about 75% of users using Firefox never opened a task manager and if, that they would not realize the memory usage of FF is high. If you would say that the vast majority of developers don't have memory leakage this would mean that some have. Those should be the "ancor points" of effort. I never saw a statement that this is taken care of. If there is, point me to it, please. Make it public, make it a number one target, have people unite. Don't mark bugs as resolved which actualy really aren't just because the bug report is too unspecific. This is damn near to denying this specific bug by pulling it of from a "open bugs" list. It makes me ask "Do you guys really want to fix this?". If needed, expand the bugs status options. How about a "find that leak" initiative on a larger scale? I'd be glad to help if it is easy for me. If I could afford spending hours on it I'd propably be in the dev team ;)
(In reply to comment #27) > Oh, and some of the reports above are quiet specific, IMHO. Like the webcam guy > on #12. A simple task eats memory. How more specific could it be from a users > point of view? That's exactly the problem... "some of the reports". There are *many* bug reports in this "bug report". They need to be put in their own bug reports, as I described above. One problem per bug report. > You state that "the vast majority of people" does not see a problem here. What > kind of people do you relate to here? The normal "John Doe" does not even care > about memory usage. I would guess that about 75% of users using Firefox never > opened a task manager and if, that they would not realize the memory usage of > FF is high. Please take this discussion over to MozillaZine. The "memory problems" have been discussed there many times. In nearly all cases, there is no actually memory problem, or it has something to do with a user's configuration, such as an extension that has a bad memory leak.
There are reports over so many possibly different problems in this bug, it looks somewhat like a meta bug. kaeptn00, would you be more at peace if instead of being declared INVALID, this bug were marked as DUPLICATE of meta bug 320915 ? I cannot make that change, but I think you can; and I have a hunch that the Mozilla gurus wouldn't bitch much about the difference. Steve, if you disagree, please speak up.
(In reply to comment #29) > There are reports over so many possibly different problems in this bug, it > looks somewhat like a meta bug. kaeptn00, would you be more at peace if instead > of being declared INVALID, this bug were marked as DUPLICATE of meta bug 320915 > ? I cannot make that change, but I think you can; and I have a hunch that the > Mozilla gurus wouldn't bitch much about the difference. Steve, if you disagree, > please speak up. > Thank you for pointing me there. Trying to reopen and dupe ... BTW, I just did a "test drive" in safe mode loosing 1MB of mem in 20 minutes on a really simple, self-reloading file displaying a local picture. Closing the file left a higher mem footprint than I had right after loading the file. This might be a misinterpretation. Was a small test anyways ... there's now a log file which I can't process due to lack of perl. Anyone?
Status: RESOLVED → REOPENED
Resolution: INVALID → ---
*** This bug has been marked as a duplicate of 320915 ***
Status: REOPENED → RESOLVED
Closed: 18 years ago → 18 years ago
Resolution: --- → DUPLICATE
(In reply to comment #30) [...] > This might be a misinterpretation. Was a small test anyways ... there's now a > log file which I can't process due to lack of perl. Anyone? If you are on Linux, you can probably find a perl package on your distribution CD, or wherever you got your Linux distribution. You gan also get perl for free (for Windows or Unix/Linux/MacOsX IIUC) at http://aspn.activestate.com/ASPN/Downloads/ActivePerl/Source .
(In reply to comment #30) > Thank you for pointing me there. Trying to reopen and dupe ... This isn't a duplicate of that bug. That one is about new memory problems on the Gecko 1.8 branch (Firefox 1.5). This bug report was opened on Firefox 1.0. But whatever. > BTW, I just did a "test drive" in safe mode loosing 1MB of mem in 20 minutes on > a really simple, self-reloading file displaying a local picture. Closing the > file left a higher mem footprint than I had right after loading the file. It's normal for memory use to be higher after completing a sequence of actions. It does not indicate a memory leak. > This might be a misinterpretation. Was a small test anyways ... there's now a > log file which I can't process due to lack of perl. Anyone? Why not use the JavaScript version of the leak gauge? Also, I should remind everyone that this is not a discussion forum. The discussions here should be taking place in the Firefox Bugs forum on MozillaZine. When you have collected enough information to write a legitimate bug report, including steps to reproduce, you can file a bug report and discuss the well-defined bug there.
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: