1219334 - Memory usage >1gb after about a week in safe mode (no addons) since TB38. Normal message activity. OK when using 37.0a1 2014-12-05

Status: RESOLVED WORKSFORME

(Thunderbird :: General, defect)

Description

[Silex]

Attachment: memory-report.json.gz

User Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:41.0) Gecko/20100101 Firefox/41.0
Build ID: 20151014143721

Steps to reproduce:

Keep thunderbird open for a few days (around 2 or 3)


Actual results:

Use of more of 1 Gb of RAM, freeze and/or slow the software


Expected results:

Keep RAM around 400/600 Mb

Comment 1

[Silex]

Please note i got this problem on 2 thunderbird on 2 computers (windows) with different mail account (but lightning on both)

Here are the debug data of thunderbird

  Paramètres de base de l'application

    Nom: Thunderbird
    Version: 38.3.0
    Agent utilisateur: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:38.0) Gecko/20100101 Thunderbird/38.3.0
    Dossier de profil: Ouvrir le dossier correspondant

              (Lecteur local)
    Identifiant de compilation de l'application: 20150928051427
    Plugins activés: about:plugins
    Configuration de compilation: about:buildconfig
    Utilisation mémoire: about:memory

  Comptes courrier et groupes
    account1:
      INCOMING: account1, , (imap) imap.univ-lille1.fr:993, SSL, passwordCleartext
      OUTGOING: smtps.univ-lille1.fr:587, alwaysSTARTTLS, passwordCleartext, true

    account2:
      INCOMING: account2, , (none) Local Folders, plain, passwordCleartext

    account3:
      INCOMING: account3, , (imap) imap.googlemail.com:993, SSL, passwordCleartext
      OUTGOING: smtp.googlemail.com:465, SSL, passwordCleartext, true

    account5:
      INCOMING: account5, , (imap) imap.googlemail.com:993, SSL, passwordCleartext
      OUTGOING: smtp.googlemail.com:465, SSL, passwordCleartext, true

  Rapports de plantage

  Extensions
    Adblock Plus, 2.6.11, true, {d10d0bf8-f5b5-c8b4-a8b2-2b9879e08c5d}
    Display Mail User Agent, 1.7.0, true, {F8147CF4-B9E3-445B-AA87-081ED66548F8}
    Enigmail, 1.8.2, true, {847b3a00-7ab1-11d4-8f02-006008948af5}
    hubiC for Filelink, 0.3.1, true, [e-mail]
    Lightning, 4.0.3.1, true, {e2fda1a4-762b-4020-b5ad-a41df1933103}
    Mark All Read Button, 0.7.1, true, {DA506923-55C4-4fbe-9E4C-D24B80D7DE9F}
    Module d'Antidote, 8.38.56, true, [e-mail]
    Provider for Google Calendar, 1.0.4, true, {a62ef8ec-5fdc-40c2-873c-223b8a6925cc}
    Remove Duplicate Messages (Alternate), 0.3.12, true, {a300a000-5e21-4ee0-a115-9ec8f4eaa92b}
    Google Contacts, 0.7.12, false, {BDD92442-0534-4D6F-A966-BAB7D561D781}

  Préférences modifiées importantes

    Nom: Valeur

      accessibility.typeaheadfind.flashBar: 0
      browser.cache.disk.capacity: 358400
      browser.cache.disk.smart_size_cached_value: 358400
      browser.cache.disk.smart_size.first_run: false
      browser.cache.disk.smart_size.use_old_max: false
      dom.max_chrome_script_run_time: 0
      extensions.lastAppVersion: 38.3.0
      font.internaluseonly.changed: false
      font.name.monospace.el: Consolas
      font.name.monospace.tr: Consolas
      font.name.monospace.x-baltic: Consolas
      font.name.monospace.x-central-euro: Consolas
      font.name.monospace.x-cyrillic: Consolas
      font.name.monospace.x-unicode: Consolas
      font.name.monospace.x-western: Consolas
      font.name.sans-serif.el: Calibri
      font.name.sans-serif.tr: Calibri
      font.name.sans-serif.x-baltic: Calibri
      font.name.sans-serif.x-central-euro: Calibri
      font.name.sans-serif.x-cyrillic: Calibri
      font.name.sans-serif.x-unicode: Calibri
      font.name.sans-serif.x-western: Calibri
      font.name.serif.el: Cambria
      font.name.serif.tr: Cambria
      font.name.serif.x-baltic: Cambria
      font.name.serif.x-central-euro: Cambria
      font.name.serif.x-cyrillic: Cambria
      font.name.serif.x-unicode: Cambria
      font.name.serif.x-western: Cambria
      font.size.fixed.el: 14
      font.size.fixed.tr: 14
      font.size.fixed.x-baltic: 14
      font.size.fixed.x-central-euro: 14
      font.size.fixed.x-cyrillic: 14
      font.size.fixed.x-unicode: 14
      font.size.fixed.x-western: 14
      font.size.variable.el: 17
      font.size.variable.tr: 17
      font.size.variable.x-baltic: 17
      font.size.variable.x-central-euro: 17
      font.size.variable.x-cyrillic: 17
      font.size.variable.x-unicode: 17
      font.size.variable.x-western: 17
      gfx.direct3d.last_used_feature_level_idx: 0
      mail.openMessageBehavior.version: 1
      mail.winsearch.enable: true
      mail.winsearch.firstRunDone: true
      mail.winsearch.global_reindex_time: 1355215280
      mailnews.database.global.datastore.id: 3e5c822c-b5c3-4de4-a6a9-ef96e4bbf45
      mailnews.database.global.views.conversation.columns: {"threadCol":{"visible":true,"ordinal":"1"},"flaggedCol":{"visible":true,"ordinal":"3"},"attachmentCol":{"visible":false…
      network.cookie.prefsMigrated: true
      network.predictor.cleaned-up: true
      places.database.lastMaintenance: 1445968916
      places.history.expiration.transient_current_max_pages: 52497
      plugin.importedState: true
      plugin.state.nppdf: 0
      print.printer_CP2025dn.print_bgcolor: false
      print.printer_CP2025dn.print_bgimages: false
      print.printer_CP2025dn.print_colorspace:
      print.printer_CP2025dn.print_command:
      print.printer_CP2025dn.print_downloadfonts: false
      print.printer_CP2025dn.print_duplex: 1515870810
      print.printer_CP2025dn.print_edge_bottom: 0
      print.printer_CP2025dn.print_edge_left: 0
      print.printer_CP2025dn.print_edge_right: 0
      print.printer_CP2025dn.print_edge_top: 0
      print.printer_CP2025dn.print_evenpages: true
      print.printer_CP2025dn.print_footercenter:
      print.printer_CP2025dn.print_footerleft: &PT
      print.printer_CP2025dn.print_footerright: &D
      print.printer_CP2025dn.print_headercenter:
      print.printer_CP2025dn.print_headerleft: &T
      print.printer_CP2025dn.print_headerright: &U
      print.printer_CP2025dn.print_in_color: true
      print.printer_CP2025dn.print_margin_bottom: 0.5
      print.printer_CP2025dn.print_margin_left: 0.5
      print.printer_CP2025dn.print_margin_right: 0.5
      print.printer_CP2025dn.print_margin_top: 0.5
      print.printer_CP2025dn.print_oddpages: true
      print.printer_CP2025dn.print_orientation: 0
      print.printer_CP2025dn.print_page_delay: 50
      print.printer_CP2025dn.print_paper_data: 9
      print.printer_CP2025dn.print_paper_height: 11,00
      print.printer_CP2025dn.print_paper_name:
      print.printer_CP2025dn.print_paper_size_type: 0
      print.printer_CP2025dn.print_paper_size_unit: 1
      print.printer_CP2025dn.print_paper_width: 8,50
      print.printer_CP2025dn.print_plex_name:
      print.printer_CP2025dn.print_resolution: 1515870810
      print.printer_CP2025dn.print_resolution_name:
      print.printer_CP2025dn.print_reversed: false
      print.printer_CP2025dn.print_scaling: 1,00
      print.printer_CP2025dn.print_shrink_to_fit: true
      print.printer_CP2025dn.print_to_file: false
      print.printer_CP2025dn.print_unwriteable_margin_bottom: 0
      print.printer_CP2025dn.print_unwriteable_margin_left: 0
      print.printer_CP2025dn.print_unwriteable_margin_right: 0
      print.printer_CP2025dn.print_unwriteable_margin_top: 0
      print.printer_HP_LaserJet_P2015_PCL6.print_bgcolor: false
      print.printer_HP_LaserJet_P2015_PCL6.print_bgimages: false
      print.printer_HP_LaserJet_P2015_PCL6.print_colorspace:
      print.printer_HP_LaserJet_P2015_PCL6.print_command:
      print.printer_HP_LaserJet_P2015_PCL6.print_downloadfonts: false
      print.printer_HP_LaserJet_P2015_PCL6.print_duplex: 81799280
      print.printer_HP_LaserJet_P2015_PCL6.print_edge_bottom: 0
      print.printer_HP_LaserJet_P2015_PCL6.print_edge_left: 0
      print.printer_HP_LaserJet_P2015_PCL6.print_edge_right: 0
      print.printer_HP_LaserJet_P2015_PCL6.print_edge_top: 0
      print.printer_HP_LaserJet_P2015_PCL6.print_evenpages: true
      print.printer_HP_LaserJet_P2015_PCL6.print_footercenter:
      print.printer_HP_LaserJet_P2015_PCL6.print_footerleft: &PT
      print.printer_HP_LaserJet_P2015_PCL6.print_footerright: &D
      print.printer_HP_LaserJet_P2015_PCL6.print_headercenter:
      print.printer_HP_LaserJet_P2015_PCL6.print_headerleft: &T
      print.printer_HP_LaserJet_P2015_PCL6.print_headerright: &U
      print.printer_HP_LaserJet_P2015_PCL6.print_in_color: true
      print.printer_HP_LaserJet_P2015_PCL6.print_margin_bottom: 0.5
      print.printer_HP_LaserJet_P2015_PCL6.print_margin_left: 0.5
      print.printer_HP_LaserJet_P2015_PCL6.print_margin_right: 0.5
      print.printer_HP_LaserJet_P2015_PCL6.print_margin_top: 0.5
      print.printer_HP_LaserJet_P2015_PCL6.print_oddpages: true
      print.printer_HP_LaserJet_P2015_PCL6.print_orientation: 0
      print.printer_HP_LaserJet_P2015_PCL6.print_page_delay: 50
      print.printer_HP_LaserJet_P2015_PCL6.print_paper_data: 1
      print.printer_HP_LaserJet_P2015_PCL6.print_paper_height: 11,00
      print.printer_HP_LaserJet_P2015_PCL6.print_paper_name:
      print.printer_HP_LaserJet_P2015_PCL6.print_paper_size_type: 0
      print.printer_HP_LaserJet_P2015_PCL6.print_paper_size_unit: 0
      print.printer_HP_LaserJet_P2015_PCL6.print_paper_width: 8,50
      print.printer_HP_LaserJet_P2015_PCL6.print_plex_name:
      print.printer_HP_LaserJet_P2015_PCL6.print_resolution: 81799264
      print.printer_HP_LaserJet_P2015_PCL6.print_resolution_name:
      print.printer_HP_LaserJet_P2015_PCL6.print_reversed: false
      print.printer_HP_LaserJet_P2015_PCL6.print_scaling: 0,60
      print.printer_HP_LaserJet_P2015_PCL6.print_shrink_to_fit: false
      print.printer_HP_LaserJet_P2015_PCL6.print_to_file: false
      print.printer_HP_LaserJet_P2015_PCL6.print_unwriteable_margin_bottom: 0
      print.printer_HP_LaserJet_P2015_PCL6.print_unwriteable_margin_left: 0
      print.printer_HP_LaserJet_P2015_PCL6.print_unwriteable_margin_right: 0
      print.printer_HP_LaserJet_P2015_PCL6.print_unwriteable_margin_top: 0
      privacy.donottrackheader.enabled: true

  Accélération graphique

      Description de la carte: NVIDIA Quadro 2000
      ID du vendeur: 0x10de
      ID du périphérique: 0x0dd8
      RAM de la carte: 1024
      Pilotes de la carte: nvd3dumx,nvwgf2umx,nvwgf2umx nvd3dum,nvwgf2um,nvwgf2um
      Version du pilote: 10.18.13.5585
      Date du pilote: 8-28-2015
      Direct2D activé: false
      DirectWrite activé: false (6.2.9200.17461)
      Paramètres ClearType: Paramètres ClearType introuvables
      Rendu WebGL: false
      Fenêtres avec accélération graphique: 0

      AzureCanvasBackend: skia
      AzureSkiaAccelerated: 0
      AzureFallbackCanvasBackend: cairo
      AzureContentBackend: cairo

  JavaScript

  Ramasse-miettes incrémentiel: 1

  Accessibilité

    Activée: 0
    Empêcher l'accessibilité: 0

  Versions des bibliothèques

      Version minimale attendue
      Version utilisée

      NSPR
      4.10.8
      4.10.8

      NSS
      3.19.2 Basic ECC
      3.19.2 Basic ECC

      NSS Util
      3.19.2
      3.19.2

      NSS SSL
      3.19.2 Basic ECC
      3.19.2 Basic ECC

      NSS S/MIME
      3.19.2 Basic ECC
      3.19.2 Basic ECC

Comment 2

[Wayne Mery (:wsmwk)]

Likely one of these. Any possibility you might be able to disable any for a breif period to rule them out?
 Lightning, 4.0.3.1, true, {e2fda1a4-762b-4020-b5ad-a41df1933103}
 Provider for Google Calendar, 1.0.4, true, {a62ef8ec-5fdc-40c2-873c-223b8a6925cc}
 Google Contacts, 0.7.12, false, {BDD92442-0534-4D6F-A966-BAB7D561D781}

Comment 3

[Silex]

i disable all of them on one computer, it might take a few days to test/reproduce it

Comment 4

[Silex]

(you can rule out google contact who is incompatible with thunderbird current version)

Comment 5

[Silex]

After a few days of uptime (around 6), thunderbird memory increase until more than 1Gb, even without lightning, provider and g.contacts. The leaks seems to be smaller (if my impression are correct, it should have take 2 or 3 days to be at this state).
I had a second memory report at this state.

Comment 6

[Silex]

Attachment: memory-report2.json.gz

Comment 7

[Wayne Mery (:wsmwk)]

Thanks for testing. Was this also with Enigmail and Adblock Plus disabled

Comment 8

[Silex]

Yes, with all extension/plugin disabled.
I saw i had the same problem (not the exact amount) on mac (10.8.5)

Comment 9

[Wayne Mery (:wsmwk)]

So this did not happen with the immediate prior versions 38.2.0 or 38.1.0?
If you reinstall them from http://releases.mozilla.org/pub/thunderbird/releases/ does the memory usage reproduce?   (you'll need to drill down though directories of version, platform win32, language, etc)

Comment 10

[Silex]

After reinstalling 38.1.0,
it makes the same with it, it reach 1Gb+ after a few days open (maybe more, but it might change with use of thunderbird)

Comment 11

[Wayne Mery (:wsmwk)]

Please create a diagnostic setting at preferences | advanced | general | config editor, find mailnews.database.dbcache.logging.console and set the string to 'debug'.  

After some period of time, does "open db count" in the error console stay sustained above 30?  You can open it at appmenu | tools | error console

Comment 12

[Silex]

done

it took longer than usual  (around 10 days) to reach 1.2 Gb of Memory, atm in the log, open db count is around 6 or 7, periodic check of cached dbs is around 12 or 13

Comment 13

[Silex]

After around 17 days, Thunderbird reach 1.9 Gb of Memory,
open db count have gone once between 16 and 18, then go back between 6 and 10
periodic check is around 12 (once 25)

i have to relaunch it now, and update it to 38.4 since it's impossible to use in this state

Comment 14

[Silex]

Attachment: memory-report-15-12-14.json.gz

here the last memory report (hope it'll help)

Comment 15

[Wayne Mery (:wsmwk)]

Was it OK when you were using version 31?
And if it was OK, can you install version 31 please, and tell us your results?
Install from http://releases.mozilla.org/pub/thunderbird/releases/31.7.0/win32/fr/

Comment 16

[Silex]

i will try to do this test at the end of the week, until next year,
but after the 5 jan. 2016 i'll have to stop test (for personal reason) until february.
So i'll keep you update on result around 4 or 5 jan.

Comment 17

[Carl]

I'm suffering similar problem. After a few days memory use can be 800MB and Thunderbird gets too painfully slow to use anymore. Closing Thunderbird takes so long to close and/or doesn't close that I sometimes have to kill the process.

Upgrading from 38.5.0 to 38.5.1 didn't help and I think I was suffering this same problem with earlier version, but cannot be sure. Definitely was not suffering this problem a few versions back.

I am using Thunderbird with the same IMAP account I have had for years. No extensions added or removed in years (see below for installed/enabled list).

I see in bug #1225491 a comment from Wayne Mery that "Other possibility - high memory is sometimes caused by bad .msf files". Deleting all MSF files is safe, correct? They will all just get cleanly recreated?

Mozilla/5.0 (Windows NT 5.1; rv:38.0) Gecko/20100101 Thunderbird/38.5.1

Extensions:
  Addressbooks Synchronizer 1.1.4
  ImportExportTools 3.2.4.1
  Lightning 4.0.5.1
  MoreFunctionsForAddressBook 0.7.3
  Quicktext 0.9.11.6

Comment 18

[Wayne Mery (:wsmwk)]

Silex...

(In reply to Silex from comment #16)
> i will try to do this test at the end of the week, until next year,
> but after the 5 jan. 2016 i'll have to stop test (for personal reason) until
> february.
> So i'll keep you update on result around 4 or 5 jan.


Carl, please test with one or more addons disabled

Comment 19

[Silex]

Sorry, i wasn't able to make the test in time between my absence (too many things to solve)

Like i said before, i was away for more than a month, i let thunderbird run on version 38.5, it seems the problem has disappeared : it was around 700 Mb memory when windows was reduce (but was still working, for filters at last), get up until 1.2 Gb once until fully opened. But i still had to close, and relaunch Thunderbird to get back control of it.

Is the test on version 31.x still usefull, or does it changes something ?

regards

Comment 20

[Wayne Mery (:wsmwk)]

 memory-report.json.gz 2015-10-28 
├────609.07 MB (49.95%) ── maildb/database(<anonymized>) [12]

memory-report2.json.gz 2015-11-09 
├──1,008.02 MB (101.15%) ── maildb/database(<anonymized>) [?!] [10]

memory-report-15-12-14.json.gz 2015-12-14 "After around 17 days, Thunderbird reach 1.9 Gb of Memory"
├────934.28 MB (51.27%) ── heap-unclassified

I'm not sure what to make of any of this, especially the last one. 

My first thought is to run calendar addon and its friends out of a separte profile. https://support.mozilla.org/en-US/kb/using-multiple-profiles describes using multiple profiles. And you would need to start one of your profiles with 
 -P -no-remote

Comment 21

[Wayne Mery (:wsmwk)]

p.s. the purpose of running calendar addon out of a separate profile is to be able to clearly identify which environment has the issue - the one with calendar or the one without.

Comment 22

[Carl]

(In reply to Wayne Mery (:wsmwk, use Needinfo for questions) from comment #18)
> 
> Carl, please test with one or more addons disabled

I've just completed at least 3 days running Thunderbird with all extensions disabled. There was NO runaway memory usage, staying relatively stable in the 220MB vicinity. Refer to comment 17 for my usual list of extensions. This does not, of course, necessarily imply that the problem lies with one of these extensions, as it could still be a bug in Thunderbird that is exposed by an extension, but it does raise a question in my mind. I've used Lightning for many years, but I believe it recently became a bundled component of the standard Thunderbird distribution. Was my existing Thunderbird/Lightning installation silently modified in some way such that I now get Lightning upgrades as part of Thunderbird upgrades or am I still getting those from the original extension repository? Is there a difference between the bundled Lightning versus the separately distributed version?

I have now re-enabled Lightning and Addressbooks Synchronizer as those are pretty important to day-to-day activity.

Comment 23

[Carl]

@Wayne Mery: I just realized you're asking us to use the Needinfo feature. I'd appreciate answers to my questions in comment 17 and comment 22.

Comment 24

[Matt]

To answer your last question,  the versions of Lightning bundled with Thunderbird and those available on AMO are identical.  To prevent duplicated updates the version bundled with Thunderbird has a 0.0.1 point version difference,  but the binary is identical.

To answer your first question,  deleting them is usually safe.  What you may see is a loss of tags if the accounts are IMAP and the server does not support tags as in those circumstances the tags are stored in the MSF file.  Otherwise they will be cleanly regenerated.  However if disabling add-ons ends the issue it is unlikely to be a solution.

Comment 25

[Silex]

Well, my collegue had to relaunch when i wasn't here, so nothing is solved (after a week, thunderbird use 1.5 Gb of Memory atm) My bad, please don't mind what i said in comment #19 (In reply to Silex from comment #19)
I will start test with different profile today.

Comment 26

[Silex]

After 9 days, thunderbird (fresh profile, one extension, no ligthning) use 1 Gb of Memory ... again :)
Do you need more test or info ?

Comment 27

[Wayne Mery (:wsmwk)]

@Carl,  800MB is not excessive memory depending what you are doing in calendar. So please do not mess with .msf files until you eliminate lightning as the probable cause.  Please check - memory: http://mzl.la/1TL3iVh   perf: http://mzl.la/1TL3ovW  and report your results

@Silex, If you understand your comments, you did not experience this problem when using version 31. So I would ask that you a) disable application updates at tools | options | advanced | update  b) install an older version [1] and report your results (you may need to adjust your addon versions to match)

[1] http://releases.mozilla.org/pub/thunderbird/releases/31.8.0/win32/en-US/

Comment 28

[Carl]

Attachment: memory-report,2016-02-28,Carl.json.gz

Comment 29

[Carl]

(In reply to Wayne Mery (:wsmwk, use Needinfo for questions) from comment #27)
> @Carl,  800MB is not excessive memory depending what you are doing in
> calendar. [ ... ] Please check - memory:
> http://mzl.la/1TL3iVh   perf: http://mzl.la/1TL3ovW  and report your results

The sort of memory usage I've been seeing is not normal for me, or at least it wasn't in all the many years I've been using Lightning. In any case, I'm not sure what you want me to do with the 38 different bug reports you've pointed me to. I did spot someone mentioning the about:memory thing, which is new to me, so I've attached a current anonymized memory report. When I use the "Show memory reports" feature of about:memory, the following text in red precedes the numerical data:

  WARNING: the following values are negative or unreasonably large.

      * explicit/maildb/database(<anonymized>)
      * explicit/(19 tiny)
      * explicit/(19 tiny)/heap-unclassified 

  This indicates a defect in one or more memory reporters. The invalid values are highlighted.

In the numeric data, explicit/maildb/database is red highlighted "605.95 MB", explicit/(19 tiny) is "-532.72 MB", and explicit/(19 tiny)/heap-unclassified is "-552.05 MB". At one point there was a fourth bullet in that list, but it disappeared when I clicked the Measure button again and I hadn't made a note what it was.

Comment 30

[Silex]

(In reply to Wayne Mery (:wsmwk, use Needinfo for questions) from comment #27)
> @Silex, If you understand your comments, you did not experience this problem
> when using version 31. So I would ask that you a) disable application
> updates at tools | options | advanced | update  b) install an older version
> [1] and report your results (you may need to adjust your addon versions to
> match)
> 
> [1] http://releases.mozilla.org/pub/thunderbird/releases/31.8.0/win32/en-US/

Thanks
I begin test now with thunderbird 31.8, i'll keep you update

Comment 31

[Carl]

Attachment: memory-report,2016-03-01,Carl.json.gz

Comment 32

[Carl]

Attachment: memory-report,2016-03-03,Carl.json.gz

Still running the same session as my last memory report. This latest attachment corresponds to Sysinternals Process Explorer telling me that Thunderbird is using 609.5K Private Bytes and 618.7K Working Set. The numbers keep growing with this session, as before. Why is the status still UNCONFIRMED? Am I wasting my time providing these memory reports?

Comment 33

[Carl]

Attachment: memory-report,2016-03-05,Carl.json.gz

Still the same Thunderbird session, but its responsiveness is now insufferable, so I'll be restarting it after this about:memory attachment.

522.79 MB (100.0%) -- explicit
├──1,671.90 MB (319.80%) ── maildb/database(<anonymized>) [?!] [9]
├──275.04 MB (52.61%) -- window-objects
...
└──-1,570.30 MB (-300.37%) -- (20 tiny) [?!]
   ├───────3.59 MB (00.69%) ++ workers/workers(chrome)
   ├───────3.33 MB (00.64%) ++ images
   ...
   ├───────0.00 MB (00.00%) ++ media
   └──-1,593.36 MB (-304.78%) ── heap-unclassified [?!]

Comment 34

[Wayne Mery (:wsmwk)]

(In reply to Carl from comment #32)
> ...
> 
> Still running the same session as my last memory report. This latest
> attachment corresponds to Sysinternals Process Explorer telling me that
> Thunderbird is using 609.5K Private Bytes and 618.7K Working Set. The
> numbers keep growing with this session, as before. Why is the status still
> UNCONFIRMED? Am I wasting my time providing these memory reports?

For several reasons, including because we don't confirm until we have a strong sense of cause and steps that someone else can reproduce; because your issue might not be the same as Silex's; and in your most recent comments you haven't stated addons enabled. I suggest if you file your own bug report with full details, and if only see this problem with calendar installed then file it in the calendar product https://bugzilla.mozilla.org/enter_bug.cgi?product=Calendar

Comment 35

[Wayne Mery (:wsmwk)]

(In reply to Silex from comment #30)
> > ...
> > [1] http://releases.mozilla.org/pub/thunderbird/releases/31.8.0/win32/en-US/
> 
> Thanks
> I begin test now with thunderbird 31.8, i'll keep you update

Thanks

Comment 36

[Carl]

(In reply to Wayne Mery (:wsmwk, use Needinfo for questions) from comment #34)
> and in your most recent
> comments you haven't stated addons enabled.

I told you in comment #22 that I had "re-enabled Lightning and Addressbooks Synchronizer" add-ons, having disabled all add-ons prior to that. I would have told you if I'd enabled any additional add-ons.

> I suggest if you file your own
> bug report with full details, and if only see this problem with calendar
> installed then file it in the calendar product
> https://bugzilla.mozilla.org/enter_bug.cgi?product=Calendar

At this point it is not totally obvious whether this is a calendar problem. It is not obvious that the cause of the problem is different than Silex. And since you do not "have a strong sense of cause" either, it seems a bit hasty to me to be disregarding what I'm seeing.

Comment 37

[Silex]

(In reply to Wayne Mery (:wsmwk, use Needinfo for questions) from comment #35)
> (In reply to Silex from comment #30)
> > > ...
> > > [1] http://releases.mozilla.org/pub/thunderbird/releases/31.8.0/win32/en-US/
> > 
> > Thanks
> > I begin test now with thunderbird 31.8, i'll keep you update
> 
> Thanks

One week after, Thunderbird use 418 Mo (that seems to be normal) on 31.8
Any other test ? Do you need a memory report ?

Comment 38

[Wayne Mery (:wsmwk)]

Thanks for the update. No memory reports needed. 

It may be a tad tedious but the next thing to do is determine which beta version between 31 and 38 *first* shows this behavior. (we actually need to get more fine grained than that but the betas are a safe start)  At http://releases.mozilla.org/pub/thunderbird/releases/ run 36.0b1 a few days and if it also shows the problem then jump back to 34.0b1, etc

Comment 39

[Silex]

After 6 day and 23 hours, Thunderbird 36.0b1 takes 1.2 Gb of memory
I'll try with 34.0b1 in a few hours

Comment 40

[Silex]

After 6 day and 17 horus, Thunderbird 34.0b1 use 328 Mb of memory
i wait for further instructions :)

Comment 41

[Wayne Mery (:wsmwk)]

(In reply to Silex from comment #40)
> After 6 day and 17 horus, Thunderbird 34.0b1 use 328 Mb of memory
> i wait for further instructions :)

Next step is to determine which nightly build between the first 35.0a1 (~2014-09-02) and last 36.0a1 (2014-11-28) first causes problems.  You might start with https://archive.mozilla.org/pub/thunderbird/nightly/2014/10/2014-10-29-03-02-04-comm-central/  See how it is for 3 days.  Then if it worked OK pick a build half way between 2014-10-29 and 2014-11-28 (which might be challenging - we have many broken periods in 2014 resulting in missing builds)

Comment 42

[Wayne Mery (:wsmwk)]

(In reply to Silex from comment #40)
> After 6 day and 17 horus, Thunderbird 34.0b1 use 328 Mb of memory
> i wait for further instructions :)

Antropik,
We still very much need more of your details. Please see comment 41

Comment 43

[Silex]

Hello,

for now, i'm on version 36.0a1 / 2014-11-25 (so pretty close to 2014-11-28 :/)
Still no memory leak, i'll try next one today or tomorrow

Comment 44

[Silex]

hi, and sorry for the wait, the version of 2014-11-25 seems have not leak, 
and i don't find any windows version for 26, 27 and 28 in central, what should i do ?

Comment 45

[Wayne Mery (:wsmwk)]

(In reply to Silex from comment #44)
> hi, and sorry for the wait, the version of 2014-11-25 seems have not leak, 
> and i don't find any windows version for 26, 27 and 28 in central, what
> should i do ?

the next windows build occurs on 2014-12-05. Can you confirm https://archive.mozilla.org/pub/thunderbird/nightly/2014/12/2014-12-05-03-02-01-comm-central/ does not leak. If it doesn't leak then our regression range doesn't have many patches in http://hg.mozilla.org/comm-central/pushloghtml?startdate=2014-11-25+03%3A05%3A00&enddate=2014-12-04+03%3A05%3A00

Comment 46

[Silex]

ok, installing now,
thank you

Comment 47

[:aceman]

Great work narrowing this down guys!
Silex, Carl, what is the number of folders you have? Are there any folders with many messages (like above 50000) ?

Comment 48

[Silex]

for what i see, around 4k max for biggest folders (if you have a method to have all the counts, i can give you more exact stats)

Comment 49

[:aceman]

ON TB 38+ you can try View->layout->folder pane columns, then in the appeared header in the folder pane, click the column picker and enable "Total". You'll see if you have any folder with many messages.

I also asked about the number of folders.

Also, do you need to scroll in the folder pane (folder list) frequently?

Comment 50

[Carl]

(In reply to :aceman from comment #47)
> Silex, Carl, what is the number of folders you have? Are there any folders
> with many messages (like above 50000) ?

As a reminder, I am using an IMAP server, but Synchronization is enabled in Thunderbird such that all folders are available offline. My current stats as reported by Thunderbird (rightly or wrongly):

- IMAP "INBOX" contains 172 normal folders and subfolders, max (sub)folder size is 7917, total is 26027, total storage is 1272MB
- IMAP "INBOX/user" (subscriptions to shared IMAP folders) contains 48 folders and subfolders, max (sub)folder size is 4976, total is 10424, total storage is 736MB
- "Local Folders" contains 305 folders and subfolders, max (sub)folder size is 3384, total is 27715, total storage is 1485MB
- "News Feeds" contains 2 folders, max (sub)folder size is 658, total is 658, total storage is 1MB

Current memory usage is a little under 300MB, so not a problem at the moment. Currently running v45.1.0. As Thunderbird is inclined to do every couple of years or so, yesterday it decided to rebuild a new copy of my offline IMAP files (ie. INBOX.sbd abandoned and INBOX-1.sbd created), so local storage requirements just doubled and all MSF files have been freshly generated. I really hate when that happens :-(

I do scroll in the folder pane occasionally, but not frequently.

Comment 51

[Carl]

(In reply to Carl from comment #50)
> Currently running v45.1.0.

Correction: Currently v45.0, imminently v45.1.0.

Comment 52

[Silex]

> Currently running v37.0a1 since 2016-05-19 (around 2 weeks)

The memory used by thunderbird is now 796 Mb (instead of ~390 Mb on previous version), but it seems there 's no real leak (it's small, around 2 or 3Mb for 24h)


any new instructions ?

Comment 53

[Wayne Mery (:wsmwk)]

(In reply to Silex from comment #52)
> > Currently running v37.0a1 since 2016-05-19 (around 2 weeks)
> 
> The memory used by thunderbird is now 796 Mb (instead of ~390 Mb on previous
> version), but it seems there 's no real leak (it's small, around 2 or 3Mb
> for 24h)

Is that with any addons installed, like lightning?


> any new instructions ?

Is your  memory usage still below 800MB?
Is the comment 41 regression range definitely correct?

Comment 54

[Silex]

There were no addons installed (not even Lightning which makes things not easy for work :)

For now, i'm using normal release (the version 38.5 goes quickly higher than 1.2 Gb), i'll go away for holydays in a few days, and makes test again around 20 July if you want.

Comment 55

[Wayne Mery (:wsmwk)]

(In reply to Silex from comment #54)
> There were no addons installed (not even Lightning which makes things not
> easy for work :)
> 
> For now, i'm using normal release (the version 38.5 goes quickly higher than
> 1.2 Gb), i'll go away for holydays in a few days, and makes test again
> around 20 July if you want.

yes please

Comment 56

[:aceman]

Do you also use a news account in TB?

Comment 57

[Silex]

(In reply to :aceman from comment #56)
> Do you also use a news account in TB?

No, i don't use newsgroup on thunderbird :)

Comment 58

[Wayne Mery (:wsmwk)]

Silex,
What happened with the windows build 2014-12-05 mentioned in comment 45?

Comment 59

[Silex]

(In reply to Wayne Mery (:wsmwk, NI for questions) from comment #58)
> Silex,
> What happened with the windows build 2014-12-05 mentioned in comment 45?

If i'm not wrong, it's in comment 52

For now, i'm still testing again a few versions to be sure :
36.0a1 (2014-10-29), (2014-11-05), (2014-11-15), (2014-11-22) and (2014-11-25) seems to have no real leak.

i was going to change soon to another version, do you have any precise build in mind to test ?

Comment 60

[Wayne Mery (:wsmwk)]

Thanks for asking.  

If it is not happening with 2014-12-05, please determine which build starts showing the behavior between 2014-12-06 and 2015-01-12-03 (the last v37 build).

Comment 61

[Silex]

Sorry for taking this long, in fact, what ever the version of thunderbird is used, i don't have major leak anymore ... 
I guess the time i spend between tests on up-to-date version corrected something in the database of thunderbird (but what and how ?)

Comment 62

[Wayne Mery (:wsmwk)]

Odd. Well, it's been a long slog and we can be grateful it's gone (for now). Unfortunate we didn't find the cause. And I'm not finding other bug reports about memory in that time frame of 37.0a1

We'll close it for now and if the problem returns please comment in the bug with new info.  Thanks.