Thank Ondrej.Zizka !
I agree that there is no parameter that change the behavior of the locks. IBM requires it to be executed at the moment the issue occurs:
-> db2mon.pl and db2fodc -hang full
I understand to detect something that could be tuned in the db2 engine.
We know that the application has worked well with Oracle and they did not have this Issue of locks.
The table ATMDEVICESTATE is the table that more OVERFLOWs and more UPDATE suffers, here statistics since the BD is reactivated at
Start Date Start Time
2018/06/05 02:25:15
TABLE_NAME PAGES OVERFLOWS
--------------- ----------- -----------
ATMDEVICESTATE 12 1,192,483
The Database page size = 4096 but the table was moved to tablespaces of 32K, nevertheless OVERFLOWs still appears at the top.
The ATMDEVICESTATE table only has 1063 rows with two indexes and has columns VARCHAR (10,32,50,64 and 256) we have considered rebuilding the table and change VARCHAR by CHAR, of course VARCHAR (256) changing it only by CHAR (255), we have monitored the column VARCHAR (256) and the maximum reached at the moment is 116 in length.
Do you think this could help ???, Any other idea ???
I'm going to review "FOR READ ONLY and the USE AND KEEP UPDATE LOCKS"