Email import errors

[Rafael03]

If your MariaDB tables are still on the older utf8 charset, definitely upgrade them to utf8mb4. It's almost always a weird character in the email subject causing the crash. As for the 'Record has changed' error, it’s definitely a locking issue - parallel cron is the right move, but also double-check that your server's system time is perfectly synced (NTP), as tiny offsets can sometimes mess with record timestamps in high-volume imports.

[rabii]

i guess that the first error has nothing to do with import failing - as first error is just that your instance tried to save an email with a subject that contains an invalid or non-UTF-8 character (\xD0). MySQL rejected it because the database/table/column is not using utf8mb4.

However from the second error is different and it show that this could an optimistic concurrency failure - meaning while the import is running there could be another process running for the same resource and the resource keeps changing.

Try to activate parallel cron job.

[a.slyzhko]

Hi,
we occasionally lose incoming emails because some messages fail to import. In logs we see two repeating errors:

ERROR: EmailAccount 67e563c63b1cdc4e9, import message; 22007 SQLSTATE[22007]: Invalid datetime format: 1366 Incorrect string value: '\xD0' for column `espo`.`email`.`name` at row 1

and

ERROR: EmailAccount 61a3511f0a13a5e8f, import message; HY000 SQLSTATE[HY000]: General error: 1020 Record has changed since last read in table 'email_user'

Also similar to the second one, but with 'entity_team' table, occurs from time to time, but less frequently than with 'email_user'.

Environment: Ubuntu 22.04, MariaDB 11.8, EspoCRM 9.2.5.

Does anyone know what typically causes these, and how to prevent the import from failing? Any suggestions appreciated.