Why journal full occurs and how to deal with it?
(1) Journal files size are too small
Journal size default value is 4M, if it's caused by journal size too small, you can increase the journal size as the following steps:
A. Stop DB.
B. Increase the size of DB_JNLSZ or add the number of DB_JnFil, and set DB_SMODE=2 in dmconfig.ini.
C. Restart DB.
(2) Long transactions exist
- Split long transaction to some short transactions.
- Don’t execute "commit" for a long time when setting is “autocommit off”
(3) Backup mode is on
When backup server being turned on, lot of journal size will be hold until backup operation finished, so it’s better to execute backup in time.
(4) In table replication environment
A. In our testing, journal full in master side caused by creating triggers while loading DB, the workaround is to use "column number" options (default 1000) as following:
dmSQL> load db from xxxxx 100;
B. If journal full happen in slave side, it should be caused by creating replication with "CLEAR and FLUSH", you can you turn off journal while loading DB to avoid journal full as following:
dmSQL> set loaddb fast;
dmSQL> load db from xxxxx 100;
dmSQL> set loaddb safe;
Version: All version
Product: Normal/Bundle
Platform: Windows/Linux