Hi,

Since I updated to kdepim 15.08 on Chakra, akonadi + mysql (system instance) is the source of constant non-negligeable disk usage. This slows down the system considerably. Also kmail seems to resync completely all mailboxes on every start.
I'd like to avoid having to set everything up from scratch, if anyone has any ideas as to how to debug it, please suggest...

Thanks

In order to see what's wrong with it, there are a few things to check:

- Email indexing is complete? (check "akonadiconsole", "Indexing Agent" and see if there are status messages)
- If indeed it's continuously resyncing the mail (use a network monitor applet / program to see if that's the case)
- If you can tell via a system monitor which process is using the most disk I/O
- If you are using POP3 or IMAP, and an approximate (order of magnitude) number of messages you have.

Hi,

The status line for indexing agent is "online/idle", all the time.
I/O throughput is reported to be 90% of mysqld, which ATM has nothing else than akonadi connected to it.
Network is mostly idle. With spikes that could be attributed to almost anything. I'll try to get a more detailed network monitor running.
They're all IMAP accounts, one with about 15k emails, two < 1k emails.

Akonadiconsole debugger is full of :

Code: Select all

akonadi_imap_resource_0 (0x7fb2500bb120) 10644 { Command: "CreateItem" Merge mode: "(Remote ID, Silent)" Collection: "UID 54" MimeType: "message/rfc822" GID: "" Remote ID: "7104" Remote Revision: "" Size: "0" Time: "QDateTime( Qt::TimeSpec(LocalTime))" Tags: "Invalid scope" Added Tags: "Invalid scope" Removed Tags: "Invalid scope" Flags: "QSet(\SEEN)" Added Flags: "QSet()" Removed Flags: "QSet()" Attributes: { } Parts: "QSet()" }
akonadi_imap_resource_0 (0x7fb2500bb120) 10644 { Response: "FetchItems" Error Code: "0" Error Msg: "" ID: "10519" Revision: "0" Collection ID: "-1" Remote ID: "" Remote Revision: "" GID: "" Size: "0" Mimetype: "" Time: "QDateTime(2015-10-13 09:37:21.635 UTC Qt::TimeSpec(UTC))" Flags: "QVector()" Tags: { } Relations: { } Virtual References: "QVector()" Ancestors: { } Cached Parts: "QVector()" Parts: { } }
akonadi_imap_resource_0 (0x7fb2500bb120) 10644 { Response: "CreateItem" Error Code: "0" Error Msg: "" }
akonadi_imap_resource_0 (0x7fb2500bb120) 10645 { Command: "CreateItem" Merge mode: "(Remote ID, Silent)" Collection: "UID 54" MimeType: "message/rfc822" GID: "" Remote ID: "7105" Remote Revision: "" Size: "0" Time: "QDateTime( Qt::TimeSpec(LocalTime))" Tags: "Invalid scope" Added Tags: "Invalid scope" Removed Tags: "Invalid scope" Flags: "QSet(\SEEN)" Added Flags: "QSet()" Removed Flags: "QSet()" Attributes: { } Parts: "QSet()" }
akonadi_imap_resource_0 (0x7fb2500bb120) 10645 { Response: "FetchItems" Error Code: "0" Error Msg: "" ID: "10520" Revision: "0" Collection ID: "-1" Remote ID: "" Remote Revision: "" GID: "" Size: "0" Mimetype: "" Time: "QDateTime(2015-10-13 09:37:21.640 UTC Qt::TimeSpec(UTC))" Flags: "QVector()" Tags: { } Relations: { } Virtual References: "QVector()" Ancestors: { } Cached Parts: "QVector()" Parts: { } }
akonadi_imap_resource_0 (0x7fb2500bb120) 10645 { Response: "CreateItem" Error Code: "0" Error Msg: "" }

The gmail (15k emails) IMAP account status is "Online, running (40%)
Syncyng Folder '<folder name>' "

Thanks.

Edit : making an "akonotes" resource offline accelerated the gmail sync and it finished within 5 minutes. Will try logging out/in later today to check whether it still happens...

Update : Just received a mail in gmail, and mysqld has gone crazy.
Akonadiconsole shows the imap resource status as "online, syncing(3000000%)" percents still climbing...
And the debugger is full of

Code: Select all

akonadi_imap_resource_0 (0x7fb2500bb120) 82476 { Command: "CreateItem" Merge mode: "(Remote ID, Silent)" Collection: "UID 54" MimeType: "message/rfc822" GID: "" Remote ID: "47648" Remote Revision: "" Size: "0" Time: "QDateTime( Qt::TimeSpec(LocalTime))" Tags: "Invalid scope" Added Tags: "Invalid scope" Removed Tags: "Invalid scope" Flags: "QSet(\SEEN)" Added Flags: "QSet()" Removed Flags: "QSet()" Attributes: { } Parts: "QSet()" }
akonadi_imap_resource_0 (0x7fb2500bb120) 82476 { Response: "FetchItems" Error Code: "0" Error Msg: "" ID: "113206" Revision: "0" Collection ID: "-1" Remote ID: "" Remote Revision: "" GID: "" Size: "0" Mimetype: "" Time: "QDateTime(2015-10-13 10:09:54.228 UTC Qt::TimeSpec(UTC))" Flags: "QVector()" Tags: { } Relations: { } Virtual References: "QVector()" Ancestors: { } Cached Parts: "QVector()" Parts: { } }
akonadi_imap_resource_0 (0x7fb2500bb120) 82476 { Response: "CreateItem" Error Code: "0" Error Msg: "" }akonadi_imap_resource_0 (0x7fb2500bb120) 82476 { Command: "CreateItem" Merge mode: "(Remote ID, Silent)" Collection: "UID 54" MimeType: "message/rfc822" GID: "" Remote ID: "47648" Remote Revision: "" Size: "0" Time: "QDateTime( Qt::TimeSpec(LocalTime))" Tags: "Invalid scope" Added Tags: "Invalid scope" Removed Tags: "Invalid scope" Flags: "QSet(\SEEN)" Added Flags: "QSet()" Removed Flags: "QSet()" Attributes: { } Parts: "QSet()" }
akonadi_imap_resource_0 (0x7fb2500bb120) 82476 { Response: "FetchItems" Error Code: "0" Error Msg: "" ID: "113206" Revision: "0" Collection ID: "-1" Remote ID: "" Remote Revision: "" GID: "" Size: "0" Mimetype: "" Time: "QDateTime(2015-10-13 10:09:54.228 UTC Qt::TimeSpec(UTC))" Flags: "QVector()" Tags: { } Relations: { } Virtual References: "QVector()" Ancestors: { } Cached Parts: "QVector()" Parts: { } }
akonadi_imap_resource_0 (0x7fb2500bb120) 82476 { Response: "CreateItem" Error Code: "0" Error Msg: "" }

After about 5 minutes it stopped.

I'll ask someone knowledgeable on internals on what to do to diagnose further the issue.

It looks like you're having a *single* mail that keeps being synced over and over, a case I've never seen before (talk about lucky . At first, we should try to get more information out of your Akonadi.

To do so, run this in a terminal:

Code: Select all

# Yes, you need a newline after "true" here
QT_LOGGING_RULES="* = true
qt.* = false"


followed by

Code: Select all

akonadictl restart

Try to capture this debug once the sympthom occurs, and pastebin it somewhere (with long or no expiration).

Just happened, but completely locked the PC. So I couldn't get any output. Redirected to a file now.
But : At akonadi start I get :

Code: Select all

akonadi_baloo_indexer(2988): "DatabaseLockError" "DatabaseLockError: Unable to get write lock on /home/martin/.local/share/baloo/email/: already locked"
akonadi_baloo_indexer(2988): "DatabaseLockError" "DatabaseLockError: Unable to get write lock on /home/martin/.local/share/baloo/emailContacts/: already locked"
akonadi_baloo_indexer(2988): Could not obtain lock for Xapian Database. This is bad
akonadi_baloo_indexer(2988): "DatabaseLockError" "DatabaseLockError: Unable to get write lock on /home/martin/.local/share/baloo/notes/: already locked"
akonadi_baloo_indexer(2988): "DatabaseLockError" "DatabaseLockError: Unable to get write lock on /home/martin/.local/share/baloo/calendars/: already locked"

I suppose either removing the flintlock file from each of these dirs or removing the dirs themselves could fix this?

Thanks,

Just received a mail, imap resource status went up to a few million percent. Console output shows nothing. Akonadiconsole output (partial, all lines before this snippet are CreateItem/FetchItems combo) : https://paste.kde.org/pee5ufjsa