After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 586882 - evolution takes a long time to start up
evolution takes a long time to start up
Status: RESOLVED INCOMPLETE
Product: evolution
Classification: Applications
Component: Mailer
2.26.x (obsolete)
Other Linux
: Normal major
: ---
Assigned To: evolution-mail-maintainers
Evolution QA team
Depends on:
Blocks:
 
 
Reported: 2009-06-24 19:51 UTC by Brian J. Murrell
Modified: 2017-11-20 21:12 UTC
See Also:
GNOME target: ---
GNOME version: 2.25/2.26


Attachments
stack traces while starting after crash (200.00 KB, application/x-tar)
2009-07-16 15:10 UTC, Brian J. Murrell
Details

Description Brian J. Murrell 2009-06-24 19:51:54 UTC
Usually only after having to force it down with evolution --force-shutdown.

It seems it takes many many minutes going through the .evolution/mail/vfolder/folders.db database.  Mine here is: 472,728,576.

But this is obviously something that is not scaling.
Comment 1 Srinivasa Ragavan 2009-06-26 13:56:29 UTC
One obvious problem here is that Evo doesn't do vacuum, which is bad and pending implementation :/. There should be a scheme to do auto vacuum once in a month or so. Can you see, if 'sqlite3 ~/.evolution/mail/vfolder/folders.db "vacuum;"' help improve the situation? If so then I can finalize its that. 
Comment 2 Brian J. Murrell 2009-07-08 14:09:47 UTC
So, I did a vacuum.  It took 10 minutes.  And even after that, starting evolution took minutes after the shell appeared before e-mail was usable.

This change to sqlite for scalability doesn't seem to be working too well.
Comment 3 Brian J. Murrell 2009-07-14 13:01:48 UTC
Any update or thoughts on this?  Having to wait over 10 minutes for evo to start after it has crashed is just unusable.

Where's the performance enhancement this sqlite migration was supposed to provide?
Comment 4 Srinivasa Ragavan 2009-07-14 14:48:11 UTC
Can you get some glances of 'thread apply all bt' 4-5 traces while starting up evo? in some good intervals if its 10minutues.
Comment 5 Brian J. Murrell 2009-07-16 15:10:12 UTC
Created attachment 138536 [details]
stack traces while starting after crash

Some stack traces while Evo takes many, many minutes starting up after a crash.
Comment 6 Brian J. Murrell 2009-07-16 15:12:07 UTC
(In reply to comment #4)
> Can you get some glances of 'thread apply all bt' 4-5 traces while starting up
> evo? in some good intervals if its 10minutues.

Attached.  FWIW, it seems to spend most time preparing vfolders for news groups.  I have one vfolder that aggregates all of the NNTP groups I'm subscribed to via gmane.  If you can trust the article count in the left pane these days, the largest vfolder has 406K unread NNTP articles.
Comment 7 Srinivasa Ragavan 2009-07-16 15:25:29 UTC
I think most of the time is spent in "newsgroup/all" vfolder. Is that the one you said?
Comment 8 Brian J. Murrell 2009-07-16 15:52:35 UTC
(In reply to comment #7)
> I think most of the time is spent in "newsgroup/all" vfolder. Is that the one
> you said?

'Tis indeed.
Comment 9 Brian J. Murrell 2009-10-01 10:54:57 UTC
Anything more on this bug?
Comment 10 Akhil Laddha 2011-04-12 05:57:24 UTC
Can you please check again whether this issue still happens in Evolution 2.32.2 or 3.0 and update this report by adding a comment and changing the "Version"
field? Thanks a lot.
Comment 11 Fabio Durán Verdugo 2011-05-30 03:36:08 UTC
Closing this bug report as no further information has been provided. Please feel free to reopen this bug if you can provide the information asked for.
Thanks!