GNOME Bugzilla – Bug 575039
Evolution startup takes too long
Last modified: 2013-09-13 01:03:57 UTC
In my system (Intel Dual Core 1,83GH) evolution takes around 15-20 seconds to show the main window, but it still needs around 20 seconds more to become responsive. I think this is too long, I don't know whether it has to do with the port to sqlite but I think it didn't take so long before. In addition, evolution is so agressive with the disk usage during startup that makes the whole system quite slow too.
What do you mean by responsive? What is the default folder it opens on (e.g. INBOX)? How many emails are in that folder? What sort of folder is it (e.g. IMAP, Local, ...)?
Ok, I've just built evolution from git and tried again. It still takes around 20 seconds to show the main window (too much time IMO), but now it's faster to show the contents of the folders once the main window is shown. It's still also very agressive with the disk usage during startup. (In reply to comment #1) > What do you mean by responsive? I meant that if I clicked on send/receive button it didn't do anything until another 20 seconds. > What is the default folder it opens on (e.g. INBOX)? it depends, but I think it's doesn't really matter. > How many emails are in that folder? > What sort of folder is it (e.g. IMAP, Local, ...)? > They are all local folders, I use only pop accounts. Thanks!
Try running http://www.gnome.org/~sragavan/evolution-rebuild-summarydb
(In reply to comment #3) > Try running http://www.gnome.org/~sragavan/evolution-rebuild-summarydb > It worked, it's much faster now. evo should probably run 'vacuum' periodically. Thanks.
Yeah, we're trying to figure out how to run that in the background without locking up the mailer for extended periods of time. Anyway, closing this.
(In reply to comment #3) > Try running http://www.gnome.org/~sragavan/evolution-rebuild-summarydb Is it safe to run this while evolution is running or should I shut it and/or EDS down? As for this bug's disposition, I don't really see how a) it's solved or b) it's not a bug given that people have to perform an "out of band" operation in order to avoid it. In the least, this should be a duplicate of some other bug, if closed at all.
Hello? Anyone? Bueller? Anyone at all care to answer the question in comment #6?