We have performance problem when begin some batch process . when this process begin in DB another sessions have wait with event latch: undo global data, in some sites is recommendation increase size of undo segment, but in our scenario we have lot of free space in UNDO tablespace.
How can we find reason of problem, or is this any bug? On metalink I can't see same bugs.
SELECT DECODE(request,0,'Holder: ','Waiter: ')||sid sess,
id1, id2, lmode, request, type
WHERE (id1, id2, type) IN
(SELECT id1, id2, type FROM V$LOCK WHERE request>0)
ORDER BY id1, request