2.53 memory leaks

User Help for Seamonkey and Mozilla Suite
Post Reply
jwq
Posts: 77
Joined: May 13th, 2004, 3:53 am

Re: 2.53 memory leaks

Post by jwq »

I experience the same problems in 2.53 that others have reported and also note that they were not present in 2.49. In my case, memory usage starts at around 2.0 GB real and 0 GB VM compressed. The VM compressed rises with usage and when it exceeds aroung 3.0 GB (i.e. total memory usage exceeds 5.0 GB) then I start to see lag & beachballs. I'm not using the same OS as others in this thread, I am using uBlock Origin and NoScript.

MacFixIt posted a "rule of thumb" some time ago: if one person reported a problem then it was regarded as isolated to the individual. If two people reported the same problem then it was regarded as a widespread problem affecting 10's of thousands of users.

There are sufficient reports of high memory consumption affecting performance on 2.53 to consider this a widespread problem affecting a large number of users. I believe that there is a problem with 2.53 which needs to be fixed.

What is required for this bug to be reproduced or identified?
User avatar
Frank Lion
Posts: 21173
Joined: April 23rd, 2004, 6:59 pm
Location: ... The Exorcist....United Kingdom
Contact:

Re: 2.53 memory leaks

Post by Frank Lion »

jwq wrote:I experience the same problems in 2.53 that others have reported
I note that you are also providing exactly the same total lack of information as the others who reported this.

For support or a bug fix the deal's the same - Provide real information, as writing a 'Me Too!!!!' just doesn't cut it.

http://forums.mozillazine.org/viewtopic ... #p14867877
Last edited by Frank Lion on August 5th, 2020, 7:15 am, edited 1 time in total.
"The only thing necessary for the triumph of evil, is for good men to do nothing." - Edmund Burke (attrib.)
.
User avatar
ndebord
Posts: 1122
Joined: December 7th, 2002, 9:53 am

Re: 2.53 memory leaks

Post by ndebord »

Frank Lion wrote:
jwq wrote:I experience the same problems in 2.53 that others have reported
I note that you are also providing exactly the same total lack of information as the others who reported this.

For support or a bug fix the deal's the same - Provide real information, as writing a 'Me Too!!!!' just doesn't cut it.

http://forums.mozillazine.org/viewtopic ... #p14867877
Frank,

I continue to wonder why there are such disparate reports of memory usage. I have a pretty plain vanilla little laptop (W 10 Pro) and I see around 500 megs most of the time. Using that kludge to reduce memory usage, lowers that down to around 400 megs, but no less and I never see 2 gigs of usage. Just saying

Nick
-N- Si vis pacem, para bellum
FrameWork, SeaMonkey(64-bit),Windows 10 Pro (X64- 21H2), WinPatrol, Malwarebytes & Panda Dome
User avatar
Peter Creasey
Posts: 1340
Joined: October 26th, 2007, 2:32 pm
Location: Texas

Re: 2.53 memory leaks

Post by Peter Creasey »

Frank, the point you make here is often right on; however, in this case, the info provided in this thread has been thorough.
. . . . . . . . . . Pete
User avatar
mightyglydd
Posts: 9813
Joined: November 4th, 2006, 7:07 pm
Location: Hollywood Ca.

Re: 2.53 memory leaks

Post by mightyglydd »

Hmm.. :-k Has anyone with this problem actually tried a New Test Profile..with NOTHING added ?
#KeepFightingMichael and Alex.
User avatar
Frank Lion
Posts: 21173
Joined: April 23rd, 2004, 6:59 pm
Location: ... The Exorcist....United Kingdom
Contact:

Re: 2.53 memory leaks

Post by Frank Lion »

ndebord wrote: Frank,

I continue to wonder why there are such disparate reports of memory usage. I have a pretty plain vanilla little laptop (W 10 Pro) and I see around 500 megs most of the time. Using that kludge to reduce memory usage, lowers that down to around 400 megs, but no less and I never see 2 gigs of usage. Just saying

Nick
I wondered that and, like you, the about:memory thing only ever has a minimal effect on my 200 - 800mb, whereas people with this problem seem to be saying it has a huge impact.

Maybe one day with far more information all will be revealed, or at least allow a bug to be filed. As it is, everything is just guesswork.

Peter Creasey wrote:Frank, the point you make here is often right on; however, in this case, the info provided in this thread has been thorough.
Yeah? You mean in a 'Just one guy on this whole thread tried SafeMode and no one complaining tried an additional new testing profile' type way thorough?

We even had LJ (an experienced extension author) come here complaining that memory leaks were causing their crashes, which turned out to be an extension issue. SafeMode before even posting would have saved some blushes there.


jwq wrote: MacFixIt posted a "rule of thumb"
Guess where they can put their thumb.
"The only thing necessary for the triumph of evil, is for good men to do nothing." - Edmund Burke (attrib.)
.
jwq
Posts: 77
Joined: May 13th, 2004, 3:53 am

Re: 2.53 memory leaks

Post by jwq »

ndebord wrote:I continue to wonder why there are such disparate reports of memory usage. ... I never see 2 gigs of usage.
I'm not sure there's anything relevant there. What's relevant is the fact that with version 2.53 the memory footprint increases with use until becomes unusable, when it was fine with version 2.49. BTW, I only see about a 0.1 GB reduction in usage when flushing memory from the about:memory page.
mightyglydd wrote:Hmm.. :-k Has anyone with this problem actually tried a New Test Profile..with NOTHING added ?
Yes. The memory footprint increases with usage.

Today I ran in "safe mode". Memory usage started at 2.0/0.0 GB (2.0 GB total) and rose to 3.0/1.3 GB (4.3 GB total) within a few hours of usage.
User avatar
Frank Lion
Posts: 21173
Joined: April 23rd, 2004, 6:59 pm
Location: ... The Exorcist....United Kingdom
Contact:

Re: 2.53 memory leaks

Post by Frank Lion »

jwq wrote:
mightyglydd wrote:Hmm.. :-k Has anyone with this problem actually tried a New Test Profile..with NOTHING added ?
Yes. The memory footprint increases with usage.

Today I ran in "safe mode". Memory usage started at 2.0/0.0 GB (2.0 GB total) and rose to 3.0/1.3 GB (4.3 GB total) within a few hours of usage.
Good, that's the first two diagnostic steps done.

As those results didn't improve anything (plus those steps are required anyway for a bug to be filed), you are now in a position to file a bug (see Bugzilla section here - https://www.seamonkey-project.org/dev/get-involved)

You will need to provide 'Steps to Reproduce' for the bug filing as, pretty obviously, the devs need to be able to see what it is they are fixing. Those steps are, in this case, anything you can think of. For example, my memory usage starts (just now) at 241mb and yours at 2.0gb, so anything like, how many tabs are you starting with, are you loading them all at startup, details of any particular websites you feel might be a problem, that sort of thing.
"The only thing necessary for the triumph of evil, is for good men to do nothing." - Edmund Burke (attrib.)
.
User avatar
ndebord
Posts: 1122
Joined: December 7th, 2002, 9:53 am

Re: 2.53 memory leaks

Post by ndebord »

Frank Lion wrote:
ndebord wrote: Frank,

I continue to wonder why there are such disparate reports of memory usage. I have a pretty plain vanilla little laptop (W 10 Pro) and I see around 500 megs most of the time. Using that kludge to reduce memory usage, lowers that down to around 400 megs, but no less and I never see 2 gigs of usage. Just saying

Nick
I wondered that and, like you, the about:memory thing only ever has a minimal effect on my 200 - 800mb, whereas people with this problem seem to be saying it has a huge impact.

Maybe one day with far more information all will be revealed, or at least allow a bug to be filed. As it is, everything is just guesswork. .
Frank,

Today my memory usage is down from before: One tab, 234M, 4 tabs 285M to 301M
and the sites are Memeorandum, Politico, CNN and MozillaZine Forums.

Nick
-N- Si vis pacem, para bellum
FrameWork, SeaMonkey(64-bit),Windows 10 Pro (X64- 21H2), WinPatrol, Malwarebytes & Panda Dome
User avatar
Frank Lion
Posts: 21173
Joined: April 23rd, 2004, 6:59 pm
Location: ... The Exorcist....United Kingdom
Contact:

Re: 2.53 memory leaks

Post by Frank Lion »

ndebord wrote:Today my memory usage is down from before: One tab, 234M, 4 tabs 285M to 301M
and the sites are Memeorandum, Politico, CNN and MozillaZine Forums.
Thanks, useful to know and I have a few ideas about this one, but long Firefox experience tells me that once people have decided there is a bug in a program, then nothing will change their minds, so it's far easier * just to steer them to how to file a bug.

And yeah, I ask questions for a reason and here's what happens when a user files a 'Because I say so' bug - http://forums.mozillazine.org/viewtopic ... #p14794702 Frg's reply there is pretty funny. Not as funny as this though -
bug Reporter wrote: There is a Mozillazine thread started on 2017-07-25 July 2017 which is clearly the same problem and it explicitly refers to versions 2.40 & 2.46 and affects users of Windows OSes.

http://forums.mozillazine.org/viewtopic ... &t=3032024
*cough* the bug 'Reporter' totally missing the fact that the 2017 problem there had been solved and was an extension/plugin problem. :)

Anyway, here's the filed bug - https://bugzilla.mozilla.org/show_bug.cgi?id=1444584



* easier in a 'not being accused of being a 'denier' for trying to get clarity, not having to read tiresome 'Why should a user have to....?' posts when I do offer suggestions, but mainly not having to read abuse aimed at me or generally to not read the constant self-righteous tetchyness of the afflicted on memory leak threads' type way.
"The only thing necessary for the triumph of evil, is for good men to do nothing." - Edmund Burke (attrib.)
.
frg
Posts: 1361
Joined: December 15th, 2015, 1:20 pm

Re: 2.53 memory leaks

Post by frg »

Well I have yet to see one of my many installations using 2GB when starting and I really have a few of them. I am on Windows and can't reproduce it. IanN is on Linux and can't reproduce it. I occasionally try various Linux versions plus macOS 10.15 and can't reproduce it. Have not heard a beep about this from another dev. Does not mean there is no bug just that no dev is able to reproduce it.

Other than the things Frank stated in his list I would like to know:

What OS including architecture, version and distribution.
How much ram?

If you are on linux: do you run a distribution or self compiled build: If yes just stop and download the official release and try it.

Are you using pop3 imap news rss?
Are you using oauth2?
Are you using caldav?
Are you using Flash (death to it)?

If you are starting with 2GB just for browsing please try a new test profile first.

Also check the error log for startup errors. There are usually 4 to 5 but not more.
I also suggest backing up the profile when SeaMonkey is not running and deleting the storage directory, storage.* and webappsstore.* files
If you are on Windows delete your whole AppData\Local\Mozilla\SeaMonkey in your profile. Not Appdata/Roaming\Mozilla\SeaMonkey because this is your actual profile. The local dir just contains the cache files and safebrowsing lists.
BillT52
Posts: 119
Joined: March 30th, 2008, 10:31 am
Location: New York

Re: 2.53 memory leaks

Post by BillT52 »

It seems that the web sites you visit have an effect. I was running SM with a clean profile and only 1 tab SM started with about 200MB. I visited www.yahoo.com and usage quickly grew to about 8-900MB. I came here to forums.mozillazine.org and the usage dropped slightly to 7-800 MB. I let it just sit for an hour (had dinner), and the usage actually dropped back to around 300M. I went back to yahoo.com and it climbed again to the 1GB range. Followed a link to a page on stylecaster.com and in a few minutes (reading 1 article) it climbed to 1.9 GB. Memory data from Process Explorer private bytes and working set columns.

Unscientific monitoring makes me think that the amount and/or type of advertising is affecting the memory usage.

And it just occurred to me it may not be a "leak" (usage does drop eventually), it just may be that some ad servers result in pages that require a lot of memory that SM just can't clean up faster than it's being consumed by new ads.

My 2 cents of uninformed (and maybe misleading) conjecture.
RDaneel
Posts: 603
Joined: January 19th, 2004, 2:43 pm
Location: Puget Sound, WA
Contact:

Re: 2.53 memory leaks

Post by RDaneel »

This is absolutely the case! I am glad it got mentioned... :)

What is on the pages you visit has a tremendous effect on memory usage - and it isn't always obvious which page elements are the ones contributing "more than their share" to your overall memory usage.

I will say that - as someone that previously only used NoScript, and has recently added UBlock Origin into the mix - that there are "protections" against runaway memory usage. ;)

UB-O seemed to have a very salutary effect on controlling memory usage - and even though I mostly see the "recovery" effect mentioned, there is usually a long-term growth that gets you in the end (referring to WS getting up to 3 GB and more) - but heck, why are people so wrapped up about this? All you have to do is kill SM, saying "Yes" to the offer to save your state/tabs, and then restart... everything comes back, but now your memory usage is back down to the minimal levels, and only grows as you actually visit the tabs.

(EDIT: For the record, I am using SM 64-bit on Windows 10 "2004" or "20H2", and regularly run with multiple windows, some with >200 tabs! And even with this "load", response is still smooth and doesn't become annoying / "laggy" until my WS is showing towards the higher end of between 2 and 3 GB.)
lem3
Posts: 74
Joined: May 15th, 2015, 10:29 am

Re: 2.53 memory leaks

Post by lem3 »

In my case I have seen SeaMoney behave normally and become deathly slow in both cases using about 2GB memory. Response time seems more closely related to CPU usage (as shown in Task Manager). Under conditions I haven't yet been to specify, after some time SeaMonkey CPU usage slowly cycles between (about) 0.5% to over 15%. When it's over about 11% things get very slow. At about 14% I start getting the "Not Responding" message from the operating system. Once the load decreases things are "fine" for a bit. The "about:memory" trick makes things better most of the time (temporarily) but not always.

Restarting SeaMonkey reduces the CPU usage cycle down to between 0.5% and 2.5%.

It's possible the memory size is what's driving the CPU load but they don't seem tightly coupled here.

I have more things to try but I can answer a few questions:
Mail is IMAP only with "check for new messages" turned off
Five or so lightly used RSS newsgroup feeds
I'm not using oauth, oauth2, caldav or Flash
OS is Win 10 Pro, version 1903, 8GB RAM, x64
UBlock Origin installed and turned on for all pages
User avatar
-Px-
Posts: 480
Joined: April 20th, 2011, 1:56 am

Re: 2.53 memory leaks

Post by -Px- »

@frg
Seems web workers may be the source of the leak, I've checked about:memory and found a bunch of workers present there, despite all windows/tabs with Paperspace were closed already. That is a cloud used by my project customer, not sure if they have a free version to test/reproduce the issue. Pressing GC removed those, but I wonder why they were still present

Code: Select all

5,598.98 MB (100.0%) -- explicit
├──3,169.06 MB (56.60%) -- workers
│  ├──3,058.67 MB (54.63%) -- workers(paperspace.com)
│  │  ├────980.75 MB (17.52%) -- worker(https://console.paperspace.com/DecoderWorker.js, 0x1fcd08b5800)
│  │  │    ├──979.47 MB (17.49%) -- zone(0x1fcf4ae3000)
│  │  │    │  ├──979.23 MB (17.49%) -- compartment(web-worker)
│  │  │    │  │  ├──979.19 MB (17.49%) -- classes
│  │  │    │  │  │  ├──978.44 MB (17.48%) -- class(ArrayBuffer)/objects
│  │  │    │  │  │  │  ├──512.00 MB (09.14%) ── non-heap/elements/wasm
│  │  │    │  │  │  │  ├──466.44 MB (08.33%) ── malloc-heap/elements/normal
│  │  │    │  │  │  │  └────0.00 MB (00.00%) ── gc-heap
│  │  │    │  │  │  └────0.75 MB (00.01%) ++ (4 tiny)
│  │  │    │  │  └────0.04 MB (00.00%) ++ (3 tiny)
│  │  │    │  └────0.24 MB (00.00%) ++ (7 tiny)
│  │  │    └────1.28 MB (00.02%) ++ (3 tiny)
│  │  ├────706.56 MB (12.62%) -- worker(https://console.paperspace.com/DecoderWorker.js, 0x1fce44ba000)
│  │  │    ├──705.16 MB (12.59%) -- zone(0x1fd24740000)
│  │  │    │  ├──704.92 MB (12.59%) -- compartment(web-worker)
│  │  │    │  │  ├──704.88 MB (12.59%) -- classes
│  │  │    │  │  │  ├──704.12 MB (12.58%) -- class(ArrayBuffer)/objects
│  │  │    │  │  │  │  ├──512.00 MB (09.14%) ── non-heap/elements/wasm
│  │  │    │  │  │  │  ├──192.12 MB (03.43%) ── malloc-heap/elements/normal
│  │  │    │  │  │  │  └────0.00 MB (00.00%) ── gc-heap
│  │  │    │  │  │  └────0.76 MB (00.01%) ++ (4 tiny)
│  │  │    │  │  └────0.04 MB (00.00%) ++ (3 tiny)
│  │  │    │  └────0.24 MB (00.00%) ++ (7 tiny)
│  │  │    └────1.41 MB (00.03%) ++ (3 tiny)
│  │  ├────558.56 MB (09.98%) -- worker(https://console.paperspace.com/DecoderWorker.js, 0x1fcb5446800)
│  │  │    ├──557.28 MB (09.95%) -- zone(0x1fc9d229000)
│  │  │    │  ├──557.05 MB (09.95%) -- compartment(web-worker)
│  │  │    │  │  ├──557.01 MB (09.95%) -- classes
│  │  │    │  │  │  ├──556.26 MB (09.94%) -- class(ArrayBuffer)/objects
│  │  │    │  │  │  │  ├──512.00 MB (09.14%) ── non-heap/elements/wasm
│  │  │    │  │  │  │  └───44.26 MB (00.79%) ++ (2 tiny)
│  │  │    │  │  │  └────0.75 MB (00.01%) ++ (4 tiny)
│  │  │    │  │  └────0.03 MB (00.00%) ++ (3 tiny)
│  │  │    │  └────0.24 MB (00.00%) ++ (7 tiny)
│  │  │    └────1.28 MB (00.02%) ++ (3 tiny)
│  │  ├────558.56 MB (09.98%) -- worker(https://console.paperspace.com/DecoderWorker.js, 0x1fcb0691000)
│  │  │    ├──557.28 MB (09.95%) -- zone(0x1fd0425e000)
│  │  │    │  ├──557.04 MB (09.95%) -- compartment(web-worker)
│  │  │    │  │  ├──557.01 MB (09.95%) -- classes
│  │  │    │  │  │  ├──556.26 MB (09.94%) -- class(ArrayBuffer)/objects
│  │  │    │  │  │  │  ├──512.00 MB (09.14%) ── non-heap/elements/wasm
│  │  │    │  │  │  │  └───44.26 MB (00.79%) ++ (2 tiny)
│  │  │    │  │  │  └────0.75 MB (00.01%) ++ (4 tiny)
│  │  │    │  │  └────0.03 MB (00.00%) ++ (3 tiny)
│  │  │    │  └────0.24 MB (00.00%) ++ (7 tiny)
│  │  │    └────1.28 MB (00.02%) ++ (3 tiny)
│  │  └────254.23 MB (04.54%) ++ (14 tiny)
│  ├─────60.03 MB (01.07%) ++ workers(atlassian.net)
│  └─────50.35 MB (00.90%) ++ (2 tiny)
Post Reply