Hi, I use kde sc 4.11.1 and kmail with an offline-imap setup.
Every mail-check I experience that the mysql-process started by akonadi is using the cpu 100% (all 8 cores). This holds until the mail-check is finished + about additional 5 seconds.
When actively working this can be very annoying since the whole desktop becomes less responsive during this period. And since mail-checks are done every 10 minutes or so this is a real pain.

Is this the "normal" behaviour or is something wrong with my system?
Any suggestions to improve the situation are appreciated. :)

I removed all akonadi-related error logs and did an akonadi-restart (through akonadi-kcm-module).
New error logs showed up after restart:
akonadiserver.error
akonadi_control.error.old
mysql.err

I'm not seeing this on two of my systems, so there's likely some issue somewhere.

1. Are you using mariaDB or MySQL?
2. Do you have mail indexing enabled (System Settings -> Desktop search)?

Thanks for the response.

einar wrote:1. Are you using mariaDB or MySQL?

MySQL - v5.5.32

einar wrote:2. Do you have mail indexing enabled (System Settings -> Desktop search)?

Yes, I have. I also tried deactivating it. But there is no difference.


In the meantime I also wiped my akonadi folders (~/.config/akonadi/ and ~/.local/share/akonadi/) and did a new imap-setup. After finishing importing the emails the behaviour was the same again (most errors in the log-files disappeared though).

I'lll be asking around on what is the best way to debug this (or if there is a need to report a bug).

A while ago the problem disappeared magically.
Now with kde 4.12.2 and 4.12.3 the issue shows up again. :(

You may want to try a database check: issue "akonadictl fsck" from the command line. Notice that the command will return immediately (the operation is done in the background).

einar wrote:You may want to try a database check: issue "akonadictl fsck" from the command line. Notice that the command will return immediately (the operation is done in the background).

Thanks for the hint. I tried this; sadly it didn't change anything.

I also did again a wipe of all akonadi stuff and mysql-db which also didn't change anything in the end.
I wonder if this has something to do with my mysql-setup on my gentoo-box (although I didn't do anything special, just emerged the mysql-package).

The only time I saw this is when Akonadi was built against MariaDB and I was using instead MySQL. That said, I'll see if I can poke some gurus that may help you with finding the root cause of the problem.

The high database load is most probably cause by the IMAP resource checking whether all emails in given folder are cached in Akonadi completely. This happens for each folder, so if you have many folders and some of them are bit larger, it makes the database busy for a while.

The "missing body check" is there as a sort of recovery from bugs we had in Akonadi in the past. Trust me, you are not the only one annoyed by this, but we haven't figured out yet a good way to run the check selectively. I'll bring it up on the KDE PIM sprint in couple weeks and maybe we can do something about it in 4.13.1 or .2.

Okay. At least I know now that it's most likely not my fault. :P

Thank you for taking the time making a statement and good luck in tackling these issues in your sprints!