Bug 401054

Note that this is the version included with Opensuse 15.0

> akonadictl status
Akonadi Control: running
Akonadi Server: running
Akonadi Server Search Support: available (Remote Search)
Available Agent Types: akonadi_akonotes_resource, akonadi_archivemail_agent, akonadi_birthdays_resource, akonadi_contacts_resource, akonadi_davgroupware_resource, akonadi_ews_resource, akonadi_ewsmta_resource, akonadi_facebook_resource, akonadi_followupreminder_agent, akonadi_googlecalendar_resource, akonadi_googlecontacts_resource, akonadi_ical_resource, akonadi_icaldir_resource, akonadi_imap_resource, akonadi_indexing_agent, akonadi_invitations_agent, akonadi_kalarm_dir_resource, akonadi_kalarm_resource, akonadi_knut_resource, akonadi_kolab_resource, akonadi_maildir_resource, akonadi_maildispatcher_agent, akonadi_mailfilter_agent, akonadi_mbox_resource, akonadi_migration_agent, akonadi_mixedmaildir_resource, akonadi_newmailnotifier_agent, akonadi_notes_agent, akonadi_notes_resource, akonadi_openxchange_resource, akonadi_pop3_resource, akonadi_sendlater_agent, akonadi_tomboynotes_resource, akonadi_vcard_resource, akonadi_vcarddir_resource

Here is the verbose output I am seeing:

org.kde.pim.akonadiserver: processing retrieval request for item (3483523, 3483582, 3483615, 3483619, 3483630, 3483633, 3483635, 3483657, 3483660, 3483676, 3483681, 3483685, 3483701, 3483703, 3483705, 3483707, 3483709, 3483715, 3483719, 3483730, 3483738, 3483739, 3483756, 3483759, 3483827, 3483844, 3483856, 3483857, 3483881, 3483928, 3483935, 3483946, 3483969, 3483994, 3483997, 3484027, 3484028, 3484032, 3484038, 3484050, 3484072, 3484080, 3484098, 3484104, 3484111, 3484209, 3484263, 3484307, 3484336, 3484354, 3484369, 3484393, 3484399, 3484438, 3484452, 3484466, 3484469, 3484493, 3484538, 3484616, 3484715, 3484758, 3484767, 3484846, 3484867, 3484917, 3484918, 3485055, 3485070, 3485078, 3485080, 3485084, 3485107, 3485124, 3485148, 3485248, 3485279, 3485432, 3485455, 3485461, 3485463, 3485476, 3485477, 3485480, 3485503, 3485509, 3485518, 3485524, 3485539, 3485541, 3485546, 3485550, 3485564, 3485570, 3485572, 3485574, 3485580, 3485607, 3485608, 3485617, 3485635)  parts: ("RFC822", "HEAD")  of resource: "akonadi_akonotes_resource_0"
org.kde.pim.kimap: The stream parser raised an exception: Unable to read more data
org.kde.pim.kimap: The stream parser raised an exception: Unable to read more data
org.kde.pim.akonadiserver: ItemRetrievalJob finished for request 0x7f1608221d90 , error: ""

In the Mail accounts page it keeps flicking between Ready (0%) and briefly flashing other stuff so fast I can't read it (something established.) This is with the Office365 account. My other account with Cyrus rarely, if ever, has any problems.

This error (which comes from the function below in kimap) means that no data came onto the IMAP socket for 30 seconds. I guess this could happen with very slow networks?


bool ImapStreamParser::waitForMoreData(bool wait)
{
    if (wait) {
        if (m_socket->bytesAvailable() > 0 ||
                m_socket->waitForReadyRead(30000)) {
            m_data.append(m_socket->readAll());
        } else {                                                                                                                                                                                                                           
            return false;
        }
    }
    return true;
}

Well, either that or the socket got closed meanwhile. I managed to reproduce this a few times so I added a QElapsedTimer and 0 ms had elapsed. So it wasn't a timeout, but a closed socket, very likely.

I believe this is due to https://bugs.kde.org/show_bug.cgi?id=403903

I am seeing this happen again. It now happens immediately with Microsoft office365.com. There is no delay.

This happens every time I receive a corporate-wide email with an extremely long CC list (the cc line is over 15,000 characters in length). Once this email is hit no more emails can be downloaded. Usually, this is an invitation to a corporate-wide meeting in my case. The only way to recover is to:
1. Stop akonadi
2. Delete the calendar entry in the Office365 web GUI
3. Delete the email in my inbox in the web GUI
4. Delete it from the deleted items folder in the web gui
5. Make sure it's really gone
6. Start akonadi
7. Sometimes I can use akonadiconsole to delete the offending email and allow akonadi to work again, but this isn't reliable.
8. If 7 doesn't work:
 9. stop akonadi
 10. delete .local/share/akonadi/file_db_data and .local/share/akonadictl/search_db
 11. start akonadi
 12. run akonadictl fsck

Once this is done things will work again, however it is a very slow process for Akonadi to rebuild everything since in my case I have several IMAP accounts with decades of emails. Also, I have to repeat this process every single time I receive a corporate-wide email or calendar invitation and it means I can't add it to my calendar.
I appologize for the two duplicate bugs I entered but they might add some additional info not listed here.

I consider this a critical bug, since this basically kills Akonadi every time an email with a huge CC list is received where recovery is difficult.

*** Bug 440345 has been marked as a duplicate of this bug. ***

You could try increasing the 30000 in 
bool ImapStreamParser::waitForMoreData (akonadi git repository), just to see if this 30s timeout is the reason for the problem? Maybe the server takes a long time to prepare sending this line...

I have seen values of 5 minutes or even longer being returned by the server.