User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7) Gecko/20040626 Firefox/0.9.1 Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7) Gecko/20040626 Firefox/0.9.1 When I load firefox the memory usage is around 15-20mb. After a few minutes of browsing the memory will rise up and up to over 100mb. The only way to stop it is to minimize the browser and then restore it. Once minimized the browser uses around 2 or 3mb of memory then once restored goes back to 15-20mb. Only it then rises again and repeats. Reproducible: Always Steps to Reproduce: 1.Open firefox 2.Browse pages 3.Minimize! Actual Results: Memory increases, then falls once minimized and then increases again once restored Expected Results: Memory should stay at around 15-20mb during browsing and drop once the browser is minimized

(In reply to comment #0) > When I load firefox the memory usage is around 15-20mb. After a few minutes of browsing the memory will rise up and up to over 100mb. Please give steps to reproduce, web sites and actions you take. Without steps it's impossible to fix this bug. (it works for me windows Firefox 0.9.1)

Hi, I do not know how to describe it further than i simply use the browser to browse sites. It happens on my daily usage, checking the forums i read, admin for my site and the odd news sites. it does not happen only with a specific site, and there is nothing wrong witht he sites that it does happen with. All I do is open up firefox, browse and then check the usage a while later and its up near the 100mb mark. if you tell me what to do to run tests to fnd the leak of the source or something then i will be more than happy to oblige Tingle

Windows that free memory upon minimizing is by OS design. I can view any process in the Windows XP's Task Manager, and it will reduce memory usage upon minimizing (and then spring back up when restored). Suggesting ->Invalid.

That comment is indeed true, however the bug is more really due to the fact that the memory usage is reaching 100mb before i minimize. Surely it should start at around 20mb and then stay there? 9assuming that i have memory cache disabled which i do) Tingle

Firefox has various memory leak bugs, hence why it's one of the most memory applications I'm running right now.

Firefox will not release memory of closed tabs until minimized under win 2000 according to taskmanager. Open two tabs one with: anything two with: http://bugzilla.mozilla.org/buglist.cgi?product=Firefox this will bring memory above 100mb close latter tab memory consumption doesn't drop until minimized on linux gentoo (2.6.7) memory won't even be released after minimize according to top. Bringing mem cosumption to about 120mb

(In reply to comment #6) forgot i check the memory consumption *after* i maximize firefox again in win 2000

Sorry for the above "bugs". After finding konqueror has the same memory behaviour under linux i tested this in firefox by doing the following. -open the bugs site (mem goes to 150 mem) -open a new tab with slashdot -open something else in the bugs tab -open the bugs site in the slashdot tab Memory consumption stays about the same while doing this. Instead my hypothesis it would double. So I have to conclude its the way linux manages its memory which results in this memory behaviour.

(In reply to comment #0) > User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7) Gecko/20040626 Firefox/0.9.1 > Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7) Gecko/20040626 Firefox/0.9.1 > > When I load firefox the memory usage is around 15-20mb. After a few minutes of > browsing the memory will rise up and up to over 100mb. The only way to stop it > is to minimize the browser and then restore it. Once minimized the browser uses > around 2 or 3mb of memory then once restored goes back to 15-20mb. Only it then > rises again and repeats. > > Reproducible: Always > Steps to Reproduce: > 1.Open firefox > 2.Browse pages > 3.Minimize! > > Actual Results: > Memory increases, then falls once minimized and then increases again once restored > > Expected Results: > Memory should stay at around 15-20mb during browsing and drop once the browser > is minimized I am having this same issue with Firefox 0.9.3 consuming 100MB+ of system memory. This happens when I attempt to play the Shockwave based game, Slingo, at Slingo.com. I have the latest version of Shoickwave player. Rich Specs: 750 P3 512MB RAM Win Xp Pro, Fully patched

(In reply to comment #1) > (In reply to comment #0) > > When I load firefox the memory usage is around 15-20mb. After a few minutes > of browsing the memory will rise up and up to over 100mb. > > Please give steps to reproduce, web sites and actions you take. > Without steps it's impossible to fix this bug. (it works for me windows Firefox > 0.9.1) > I have been about to reproduce this with Firefox 1.0PR repeatedly. System specs: Intel Pentium 4 2GHz 256MB RAM Windows XP With Windows Task Manager open, I started Firefox 1.0PR in Safe Mode. Firefox appears in the task list using about 15megs RAM. As I loaded different web pages, the memory usage slowly increased. After about 10 different web sites (no pages being reloaded from cache, all 10 different sites), the memory usage had increased to about 45megs. After about 15 more different sites, memory usage was up to 50megs. NOTE: this was in Safe Mode with only one tab. Minimizing brought memory down as expected. Continued same testing as before with the same results. Memory usage did not increase as quickly in safe mode as it did in normal mode, but I have a lot of extensions installed. Not sure if this is expected behavior but am willing to test it differently if required.

I had the same behaviour. I've been to jftp (http://www.jmethods.com/products/jftp/) and tried to launch the java application. I closed it before the java launched the software by clicking a google "ad link", then get back using mouse gesture to the left. The application crashed (the java console in my JRE1.4.2 told me about an exception), so I refreshed the page, and the application pops up. Then I close it and get back searching on google. The memory usage of firefox get as high as 72Mb (in WinXP task manager). When I minimized the browser, the memory get cleaned and the memory falled to 10Mb. However, the virtual memory is still in use (67Mb).

Seeing the problem here as well. Browser: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.7.5) Gecko/20041107 Firefox/1.0 Steps: 1. Open http://bugzilla.mozilla.org/buglist.cgi?product=Firefox in a new tab 2. (can be skipped) stop the page after a sizable memory usage shows up. 3. Close the tab from step 1 4. Check Task Manager. In my case Mem Usage: ~75MB, VM Size ~73MB 5. Minimize Firefox 6. Check Task Manager. In my case Mem Usage: ~1.5MB, VM Size ~73MB 7. Restore/maximize Firefox 8. Check Task Manager. In my case Mem Usage: ~20MB, VM Size ~73MB In the Memory Usage History graph, the memory usage goes up while the page is loaded, drops slightly after Firefox is minimized and restored, and then stabilizies at a high plateau (will post a screenshot)

see comment #12

I am seeing my memory usage for firefox go up to stupid amounts now. I disabled config.trim.on.minimize because i was getting major problems after having the browser minizmied as I have a very slow hard drive. So now the memory does not free itself on minimization, as expected. However, the most I have seen firefox using before I closed it was 400MB of ram. I am currently sat here using firefox whilst it is taking up 200MB of ram. I have upgraded to version 1.0. It would really be nice if someone could take a look at this for me! I will run whatever test you guys want me to do, but surely 100MB is high usage, let alone 400MB. It happens for me and my configuration, not with thunderbird of sunbird or any other app from the mozilla foundation which leads me to believe its a bug with firefox! Help please!

This "cleanup" of memory does not really occure. Win2000 and XP swap the memory out to the page file. In the task manager there are more than one columns for the memory usage, the one with the "virtual memory" not shown by default. If you enable this column, you will see, that minimizing firefox will decrease the normal memory usage, but will not decrease its virtual memory usage, which means that memory will be swapped in again when used from the swap file (one can see that both memory usages increase when continuing surfing after mini and maximizing). I'm using firefox 1.0 under Win2000 and WinXP and reach very often my memory capacity of my computer (384 MB RAM) only surfing for a few hours. Limiting caching in memory with the browser.cache.memory.capacity config did not help, the browser is comsuming memory for every Tab which seems not to be freed when the Tab is closed ... and this is MUCH memory ... even on my other machine where i have 1 GB RAM, it takes only a few hours until most of the memory is consumed by firefox and other apps are running slow because of swapping.

(In reply to comment #3) > Windows that free memory upon minimizing is by OS design. I can view any process > in the Windows XP's Task Manager, and it will reduce memory usage upon > minimizing (and then spring back up when restored). Suggesting ->Invalid. I think that this is correct. I think that it cannot discover the memory leak as long as source is not basically investigated.

This comment intends to confirm this bug. I started with a fresh rebooted W2K PC. The task manager was opened and showing the top memory eaters. Mozilla -last version- has QuickStart on, so initially the program uses VM=8596K and memory=15500K. I opened a local newspaper main page at http://lanacion.com.ar and left the browser window untouched, without focus for a couple hours. During that time the virtual memory consumed by Mozilla increased from 36388K to 101328K. Below you can see a recording of the RAM and VM numbers reported by the task manager: time RAM VM comment 15:40 27500 20368 start Mozilla 15:41 42336 36388 open http://lanacion.com.ar/ 15:46 46060 39276 15:49 47484 40468 15:55 46024 43260 15:57 47740 44988 16:03 49644 45500 16:11 52160 49944 auto-refreshed 16:20 48192 53976 16:21 54248 54132 auto-refreshed 16:27 49100 57236 16:31 59808 59480 auto-refreshed 16:33 60776 60436 16:38 63364 63308 16:41 64108 64844 auto-refreshed 16:57 57912 70732 some RAM reclaimed after using MS Word 17:21 74252 80001 17:45 75724 91128 approaching 100MB steadily 18.09 71116 101328 100MB reached in 2:30hs As this computer has only 128MB RAM I need to close the browser every hour or so, because else everything works slowly specially task swaps with alt+tab, that take up to 1 minute. In my other PC I have 512MB and the sale W2K OS. The Mozilla RAM issue is less noticeable and sometimes memory usage is 200MB and more. I have always opened three tabs: gmail, Netscape mail and the news LaNacion.com. Memory usage grows all the time. Today when I was watching the memory numbers grow the only variations I witnesed were those moving GIFs. My suspect (without any evidence at all) is that this might be the cause of the problem.

More on this: Yesterday I left the PC displaying the newspaper page overnight and this morning the task manager reported 190MB VM usage. The memory grow didn't happen today because the newspaper page was not open. Netscape mail and gmail didn't bloat my Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.7.8) Gecko/20050511

*** Bug 271760 has been marked as a duplicate of this bug. ***

*** Bug 258978 has been marked as a duplicate of this bug. ***

*** Bug 263470 has been marked as a duplicate of this bug. ***

*** Bug 263930 has been marked as a duplicate of this bug. ***

*** Bug 256424 has been marked as a duplicate of this bug. ***

*** Bug 275203 has been marked as a duplicate of this bug. ***

Some seem to believe config.trim_on_minimize=false pref is a 'fix' for this age-old FF memory consumption issue. It isn't. Disabling trimming to swapfile simply prevents FF/OS from swapping the leaky/unfreed pages to disk (which is obviously a very expensive operation). Why the hell FF/OS swaps so much data back in when maximising just one page is beyond me, but that's another story (bug). Setting "trim" just hides the leak for a while. It allows FF to continue the broken behaviour - but the b0rken-ness is just less visible because the painful disk swapping is avoided. BUT, eventually, you will run out of RAM. Even if you close all pages and tabs except one - "about:" - FF will still consume huge amounts of RAM and swap. I have a PC (WinXP) with 1024MB ram, and if I use FF for longer than two weeks without closing, I can chew the whole lot. And it's not because of the number of tabs/pages open, because when I SessionSaver Snapback after closing FF, voila, >400MB RAM freed! <apologies for the comment length>

beryan asked: > Why the hell FF/OS swaps so much data back in when maximising just one page? Without internal knowledge of FF ... For maximising a page FF maybe needs only a couple MBs. But if those MBs are scattered a few bytes in each virtual memory page then lots of memory pages have to be brought into RAM just to get the few bytes each contains. If this were true then the solution would be to reorganise FF's virtual memory arrangement, or reuse previous chunks. In the example I brought >> http://lanacion.com.ar there is a permanent traffic of animated gifs. If the display of each new gif is asking for a new memory chunk then this might be the problem. Please keep in mind that this is just a guess.

*** Bug 273615 has been marked as a duplicate of this bug. ***

I experience the following behavior when browsing multiple sites loaded with images at the same time: -FF seems to become slower until memory usage is up to about 200MB -FF freezes and finally disappears; no error message, no feedback agent, just a silent crash I run FF 1.0.5 on WinXP SP2 machine, Intel 2.8GHz with 1GB RAM. So FF has quite a playground here - but still this memory leak causes crashes since version 0.9+ Since Firefox AND Thunderbird are the two programs that use up the major amount of memory, this issue should be looked into. A while ago, I read somewhere on the web, that FF and TB are supposed to share the same engine in future versions to optimize memory consumption. Any news on this???

*** Bug 271904 has been marked as a duplicate of this bug. ***

I thought this was due to the memory cache respecting expiry dates, that's often set to insanely late IMHO, considering this is RAM being a very constrained resource compared to the disk cache. See also about:cache?device=memory and e.g. large JPG images with expire dates several DAYS after now in many cases, at least for me for a regular browsing session. No wonder browsing sessions in that case start running low on memory. Or are the expiry dates shown for the disk cache, and Firefox regularly cleans its disk cache? However, even with this consideration, it still seems a bit weird since that page now tells me it's using 64K, but Windows is reporting 114MB usage / 107MB VM.

Arrgh, I meant to write 64M above. Sorry!

*** Bug 299917 has been marked as a duplicate of this bug. ***

I definately see this as well. if you view pages such as fark.com photoshop contests where there are lots of pictures on one page you see the virtual memory usage climb quite rapidly. This virtual memory doesn't seem to be released regardless of whether you close the offending tab or not. The only way I've got to release it to end firefox completely and start up a new process.

quick snapshot of firefox memory usage after visiting a graphics intensive page *AND* close down all the tabs that held those pages. Using 1.0.4

I still see this bug in a big way. I had about 20-25 tabs open, about 10-15 of those were photos. I noticed my computer slowing down (after all the tabs were closed), and opened the task manager to find Fx using 575,000K+ of memory. I tried minimizing, but that number will not go down. Right now, all I have is this one window/one tab open, and the number has again risen to 584,000+. Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.8b4) Gecko/20050810 Firefox/1.0+ ID:2005081019

Using firefox 1.0.6, Windows XP. Short description: firefox.exe process memory usage goes to ~430MB growing by 0.5-1.5MB/s, then drops to ~70MB, then grows again to ~430MB, and that's no end. Pagefile usage remains the same about 970-990MB. I found this after weekend, when I left two firefox windows, one with 10 tabs (these tabs are open for 1-2 weeks already), and on friday I have opened another firefox window with 3 tabs, and left it for the weekend. This monday found, that process memory usage is above 400MB, and stays so. After trying to activate firefox, cycle started - 430 immediately drops to 70, and then grows again to 430MB. Closed second firefox with 3 tabs - window disappeared, memory usage didn't changed. Closed first firefox with 10 tabs - didn't helped too, firefox window disappeared, but firefox.exe process stayed growing to 430MB, and finally disappeared when reached this limit. Page file usage this moment dropped to 330MB from 970MB. I have copied the sites which were opened, if needed I can give the adresses.

just to confirm that i get this too, using 1.5 i open FF about half 8 in the morning and close it again about 5 at night after working as a web developer all day (now that's a LOT of page loads and reloads). during a CTRL ALT DELETE i noticed that FF was using > 80megs of memory at one point and the only site i had open was slashdot.

I had firefox open for several days (up to 15 windows at one point), system RAM usage was at 614MB. I closed all windows to make the popupsdie plugin install complete, and RAM usage dropped within seconds to 485MB used. IMO this bug isn't that bad to deal with if you have a lot of RAM, but it can be crippling on low RAM systems. It seems to only be releasing allocated memory for pages that aren't even loaded anymore when the browser closes fully, rather then when the tab or window containing the page closes. I'm currently using FireFox version 1.0.6 but will be updating to 1.0.7 shortly. The only plugins I have loaded are popupsdie and the imageshack rightclick menu adder.

grab as promised, i'd actually closed firefox at this point but it had gotten lodged in the xp task manager, not good!

re: https://bugzilla.mozilla.org/show_bug.cgi?id=249469#c33 Agreed. Fark is the best test case for this that I've seen. http://www.geekissues.org/c/76602 grows my VM to over a gig in a few minutes. Of course, IE behaves no better (its progressive display is much delayed over FF's, but it lets go of the VM quicker upon closing the window). I presume that this is a category of page that makes it tough to be a browser. Anybody know of any browsers that do it better? Opera? Safari?

(In reply to comment #40) > re: https://bugzilla.mozilla.org/show_bug.cgi?id=249469#c33 > > Agreed. Fark is the best test case for this that I've seen. > > http://www.geekissues.org/c/76602 > > grows my VM to over a gig in a few minutes. Of course, IE behaves no better > (its progressive display is much delayed over FF's, but it lets go of the VM > quicker upon closing the window). I presume that this is a category of page > that makes it tough to be a browser. Anybody know of any browsers that do it > better? Opera? Safari? Has it occured to you that all the pictures you see need to be stored somewhere? A single 1024x768 image (assuming 24 bit color / 3 bytes per pixel) is going to cost about 2.3MB. On a page with hundreds of images, well, do the math.

Almost completely corrected in FF 1.5 !!!

(In reply to comment #41) > > Has it occured to you that all the pictures you see need to be stored > somewhere? A single 1024x768 image (assuming 24 bit color / 3 bytes per > pixel) is going to cost about 2.3MB. On a page with hundreds of images, > well, do the math. Ayuh. It's occurred to me that they get stored on disk. In the fark case on my system (1G RAM), (VM Size)/(Mem Usage) >= 10. It usually starts spilling to pagefile.sys when VM and Usage get to around half a gig (after which Mem Usage drops off quickly as VM Size continues to grow). I wouldn't want a pathological case to dictate design. On the other hand, it'd be nice if FF had a clean, uncomplicated way to switch over to disk when it's clear there ain't gonna be enough RAM - and we're gonna be reading those images from disk either way. Whether it's from pagefile.sys or someplace else is the issue. (I tried growing my FF disk cache to 1500MB, but it didn't help.) I don't think what FF is currently doing is unreasonable. But I do think that handling it better than the competition would be another plus for FF. (N.B. - fixing this well includes taking into account MS Windows' file caching behavior. Otherwise you're just offloading runaway RAM consumption from FF to the OS. (I'm referring to Windoze 'cause that's where I'm running it at the moment. I also run it on Linux, where similar concerns apply))

Proposal to reduce the memory consumption of images in Mozilla http://primates.ximian.com/~federico/news-2005-11.html#moz-images Found this at Mozillazine.

(In reply to comment #42) > Almost completely corrected in FF 1.5 !!! > It is not true

Bugs filed without useful details are unfixable. There are multiple places where we could leak, and without specifics, its impossible to say that X leak bug is really fixed. Please file leak bugs according to the guidelines here: http://dbaron.org/log/2006-01#e20060110a

(In reply to comment #46) > Bugs filed without useful details are unfixable. There are multiple places > where we could leak, and without specifics, its impossible to say that X leak > bug is really fixed. Please file leak bugs according to the guidelines here: > http://dbaron.org/log/2006-01#e20060110a > Believe me, if I knew any more details I'd post them. I guess I'll just continue to close all of my Firefox browsers windows every few days. Huzzah for "customer service"? Oh, and my browser's memory is up to over 100mb, despite not having hit a single graphics heavy webpage in the few days I've been running this browser session.

(In reply to comment #47) > (In reply to comment #46) > > Bugs filed without useful details are unfixable. There are multiple places > > where we could leak, and without specifics, its impossible to say that X leak > > bug is really fixed. Please file leak bugs according to the guidelines here: > > http://dbaron.org/log/2006-01#e20060110a > > > Believe me, if I knew any more details I'd post them. I guess I'll just > continue to close all of my Firefox browsers windows every few days. Huzzah > for "customer service"? > > Oh, and my browser's memory is up to over 100mb, despite not having hit a > single graphics heavy webpage in the few days I've been running this browser > session. > Alas, the best "customer service" cannot fix a bug which is merely described as "my browser keeps eating more and more and more memory", especially when the "customer service people" are mostly volunteers and not salaried employees. I regret this tendency of Firefox's, but I lack the necessary expertise to sift through the whole Mozilla code hunting for potential memory leaks, so I suppose that me too, I'll have to restart Firefox maybe two or three times a day. Thanks to "Restart Firefox" and "Tab Mix Plus" extensions for a fairly streamlined single-click close-and-restart operation which keeps the same pages in the same tabs, and even (in most cases) at the same scroll point.

Reproducible: Always Steps to Reproduce: 1. Go to http://cars.ign.com/articles/681/681897p1.html 2. Scroll down to the bottom of the page and click on the Images button. 3. Click on the top left image. 4. After the page loads, click on the Next Page button. 5. Repeat step 4 about one hundred times. If you get to a page that has an annoying ad, just wait until the next image appears, and continue. Expected Results: Mem Usage shown in the Windows Task Manager is about 100 MB. Actual Results: Mem Usage shown in the Windows Task Manager is nearly 300 MB. I can reproduce in Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a1) Gecko/20060202 Firefox/1.6a1 with a new profile. I also reproduced it in a recent SeaMonkey trunk build.

The memory cache reports nearly 200 MB of RAM in use after following the set of steps above.

I can't reproduce in Firefox 1.5.0.1. After viewing just one image, the leak gauge reports (yes, I did quit Firefox before running the leak gauge!): Summary: Leaked 10 out of 10 DOM Windows Leaked 26 out of 28 documents Leaked 2 out of 3 docshells After viewing many more images, the leak gauge reports nearly 200 leaked DOM Windows and documents.

(In reply to comment #51) > I can't reproduce in Firefox 1.5.0.1. [...] Well, in "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.0.1) Gecko/20060202 Firefox/1.5.0.1", build ID:2006020204, I do see a steady rise in virtual memory size by the method set forth in comment #49 (by about 200M for 100 images, in addition to what was already there just after loading the indicated URL). AFAICT the only way to reclaim that memory is to close Firefox completely. (Reopening leaves it at its "starting" size -- or below, if there were other tabs from which memory could also be reclaimed. Note: With the extensions I have, it's a two-button click to restart Firefox with the same pages in the same tabs.)

Can any of you reproduce on recent trunk/1.8 branch builds and if so please attach leak-gauge logs to this bug?

Although if we were only leaking images leak-gauge might not show anything.

WFM with Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1a3) Gecko/20060626 BonEcho/2.0a3. The total memory usage stays below 60 MB. The memory cache stays within its default limit of 18 MB. leak-gauge.pl reports 0 leaks.

This hasn't regressed lately, has it?

I can't reproduce using the steps in comment #49 with Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a8pre) Gecko/2007082305 Minefield/3.0a8pre. Can you reproduce a problem with another set of steps?