[Akonadi] [Bug 334280] IMAP Synchronisation with GMAIL takes 13 minutes to complete !!

Christian Mollekopf mollekopf at kolabsys.com
Sun May 4 21:24:02 BST 2014


https://bugs.kde.org/show_bug.cgi?id=334280

--- Comment #9 from Christian Mollekopf <mollekopf at kolabsys.com> ---
(In reply to comment #8)
> (In reply to comment #7)
> > (In reply to comment #6)
> > > (In reply to comment #5)
> > > > This should generate network traffic indeed.
> > > 
> > > Strangely enough I didn't notice any network traffic during this period. As
> > > indicated network traffic is only visible around the end. 
> > 
> > Please let me know if the situation is improved or still the same with the
> > fix for the other bug.
> 
> With the fix for the other bug, the synchronization went back to about 5
> minutes and now I see network traffic during the flag retrieval. 
> 
> What exactly does the number mean in this line :
> akonadi_imap_resource_16(30157)/kdepimlibs (kimap)
> BatchFetcher::fetchNextBatch: Fetching  257301  to  257400
> 

It means UID's 257301 to 257400 are fetched which only correlates to the number
of messages that ever have been in the folder (not to the amount you currently
have).

> Does that mean that I have 257400 emails in that folder ? If so, then
> something else is wrong. I know that I have a number of high volume email
> lists (e.g. kde-commit at kde.org), but I am deleting the emails from there
> every 3 days to reduce the total number of emails. I also cleaned up in
> gmail.com itself, so I shouldn't have those many emails.
> 
> > Flag fetching definitely is slower because I'm currently fetching the flags
> > just as messages in batches of 100 to avoid overloading akonadi. This
> > obviously results in many roundtrips, and I suppose the batchsize could be
> > greatly increased for flags (because a lot less data than full messages).
> 
> Can I increase the number somewhere ?  Just to test if I can find a better
> number for my system and to retrieve things a little faster ?

In imapresource.cpp (setItemSyncBatchSize) we define the batch size. This
directly correlates to the number of messages that are loaded into memory, so I
wouldn't make this one too big.

In retrieveitemstask.cpp:668 you could increase the passed-in batch-size for
flag fetching only.
I think it should work if you simply set a multiplier (e.g. 10*batchSize()).

It's not configurable in a configfile.

-- 
You are receiving this mail because:
You are on the CC list for the bug.



More information about the Kdepim-bugs mailing list