Re: How to reindex akonadi database without having to reconfigure?

On Thursday, 2 February 2017 09:45:59 GMT Martin Steigerwald wrote:
> Am Donnerstag, 2. Februar 2017, 07:56:16 CET schrieb ianseeks:
> > Hi
> > 
> > I'm having the "Retrieving folder..." error message on every folder that
> > contains an unread item. Akonadi fsck/vacuum/top/start no longer fixes the
> > problem.
> > I have in the past followed the how to fix akonadi when there was a failed
> > migration and that did all the work but it seemed like a hammer to crack a
> > nut as you have to reconfigure kmail etc.
> > 
> > Is there a way to just delete the data and any related indexes to force a
> > reindexing process?
> > 
> > I'm running out of ideas hence i hope this idea may help
> 
> Did you try "Clear Akonadi Cache" on one of the folders in question in
> Akonadiconsole. As far as I understand this is a partial clearing of the
> cache which forces Akonadi to redo the cache for the particular folder.
> 
> I bet I´d do it in the following order:
> 
> 1) make sure you have backup
> 2) stop Kmail
> 3) stop Akonadi
> 4) make sure database process is gone
> 5) start Akonadiconsole
> 6) clear cache of one folder
> 7) start kmail and test result
> 
> As far as I understand you maybe could loose data with doing that, just as
> you can loose data with removing the complete database, cause while Akonadi
> is mostly a cache, there can be circumstances where it contains more than
> only cached items, i.e. when an IMAP server is not reachable to add a newly
> written mail into the sent folder… (but on that I also found that the sent
> mail was stored into the local sent mail folder instead).
> 
> I do not know exactly whether the data loss risk is only when you wipe the
> whole database, or also when you clear only the cache of one folder, but I
> warned you :)
> 
> Also I believe there was a recent bug which caused Retrieving folder
> contents displayed forever… you could research that. I am still using
> KDEPIM 16.04 as Debian has nothing newer packaged due to challenges and
> difficulties packaging Qt Webengine with Debian quality standards.
> 
> Also what I believed helped timeouts a lot is raising InnoDB buffer pool
> size *a lot*. I raised it to 1 GiB. Mysqltuner.pl recommended even more.
> But keep the amount of free memory on your system in mind. Whether that
> helps you or not mostly depends on whether Akonadi is waiting on mysqld to
> complete the queries it floods it with.
> 
> Thanks,
Thanks for your response.  I spent about 12 hours yesterday trying your 
suggestions and loads of others.  I gave up in the end and it did not improve 
my situation at all.  i created a new user with a blank email database and its 
better, akonadictl fsck/vacuum/stop/start seem to clear the problem a lot 
better now.

-- 
opensuse:tumbleweed:20170201
Qt: 5.7.1
KDE Frameworks: 5.30.0
KDE Plasma: 5.9.0
kwin5-5.9.0-1.1.x86_64
kmail2 5.4.1
Kernel:  4.9.6-1-default
Nouveau:  1.0.13_2.2