I have a total of 4 imap huge accounts (take that literal), which I try to manage with the kde pim suite (now at K:R:49 = 4.9.5 from OBS, running on openSUSE 12.2/x86_64). Thanks god, my primary system is well equipped with a fast CPU and 8 GB RAM, as well as are the mail server. Unfortunately, I have to restart akonadi every two days at least, because: hp 9112 5.8 0.5 2470692 41104 ? Sl Jan04 260:52 akonadiserver hp 9126 3.5 2.8 766564 230144 ? Sl Jan04 155:56 /usr/bin/akonadi_imap_resource hp 9127 0.3 0.6 625376 49176 ? Sl Jan04 14:10 /usr/bin/akonadi_imap_resource hp 9128 0.0 0.1 554872 11324 ? Sl Jan04 0:35 /usr/bin/akonadi_imap_resource hp 9129 0.0 0.1 555712 13284 ? Sl Jan04 0:46 /usr/bin/akonadi_imap_resource hp 9134 0.0 0.1 507080 11804 ? S Jan04 0:20 /usr/bin/akonadi_mailfilter_agent hp 9135 1.5 19.5 1895240 1596364 ? S Jan04 68:35 /usr/bin/akonadi_nepomuk_feeder From this, you can see, that I really use one primary and one secondary imap account, while the other two are kept silent (no accesses, no manual nor auto updates). I've stopped kmail already, since that wasn't responding anymore. The semantic desktop is on, file indexer and email indexer are switched off. Looking at the numbers, the 1.5 GB from akonadi_nepomuk_feeder disturbs me most. I've no idea what the code is doing or trying to archive, but feeding nepomuk must be possible with much less resource consumption or something is conceptually wrong. Reproducible: Always Steps to Reproduce: 1. run kmail with huge imap mailboxes for a while 2. check memory consumption 3. I can provide more detailed information, if necessary as I control the cyrus imap server as well.
Interestingly, after akonadictl start, it takes 3 cpu seconds for akonadi_nepomuk_feeder to reserve 1.6 GB ram now: hp 7931 12.9 19.7 1897084 1616672 ? S 20:51 0:03 /usr/bin/akonadi_nepomuk_feeder Hmm.
Here are the biggest suckers after a day: hp 7908 6.5 0.7 2271880 61072 ? Sl Jan08 64:20 akonadiserver hp 7922 3.7 3.1 767232 257012 ? Sl Jan08 36:54 /usr/bin/akonadi_imap_resource hp 7923 0.5 0.5 583472 48560 ? Sl Jan08 5:34 /usr/bin/akonadi_imap_resource hp 7924 0.0 0.2 554788 20572 ? Sl Jan08 0:08 /usr/bin/akonadi_imap_resource hp 7925 0.0 0.2 555856 21504 ? Sl Jan08 0:12 /usr/bin/akonadi_imap_resource hp 7931 4.2 19.6 1897648 1606864 ? S Jan08 42:09 /usr/bin/akonadi_nepomuk_feeder hp 8077 0.6 16.0 3238776 1312468 ? Sl Jan08 6:04 /usr/bin/kmail -caption KMail 3.3 GB reserved memory for E-Mail functionality. Now I'm about to restart akonadi since it doesn't send any massages. akonadiconsole shows: Error during executing query "SELECT CollectionAttributeTable.id, CollectionAttributeTable.collectionId, CollectionAttributeTable.type, CollectionAttributeTable.value FROM CollectionAttributeTable WHERE ( collectionId = :0 )" : "MySQL server has gone away QMYSQL: Unable to execute query" Error during executing query "SELECT PimItemTable.id, PimItemTable.remoteId, MimeTypeTable.name, ResourceTable.name, PimItemTable.rev, PimItemTable.remoteRevision, PimItemTable.size, PimItemTable.datetime, PimItemTable.collectionId FROM PimItemTable INNER JOIN MimeTypeTable ON ( PimItemTable.mimeTypeId = MimeTypeTable.id ) INNER JOIN CollectionTable ON ( PimItemTable.collectionId = CollectionTable.id ) INNER JOIN ResourceTable ON ( CollectionTable.resourceId = ResourceTable.id ) WHERE ( ( PimItemTable.id = :0 ) ) ORDER BY PimItemTable.id DESC" : "MySQL server has gone away QMYSQL: Unable to execute query" ItemRetrieverException : Unable to retrieve item from resource: <html>Invalid item retrieved</html> akonadiconsole(30100)/libakonadi Akonadi::EntityTreeModelPrivate::fetchJobDone: Job error: "Unable to fetch item from backend (collection 8, resource -1)" for collection: 8 Obviously, akonadi is unable to reestablish the autoclosed mysql connection, which I had to relocate to my server here, because HOME is NFS mounted.
*** This bug has been marked as a duplicate of bug 305352 ***