|Subject:||hangs on index rebuilt|
|Posted by:||Marco Vannini (m.vanni…@tinCANCELLA.it)|
|Date:||Mon, 19 Apr 2004|
Session hang with a 'waited too long for a row cache enqueue lock!' during
an index rebuilt after a direct load (with skip_index_maintenance) into a
large subpartition-composite table. I realized that the problem concerned
2 newly created function-based bitmap indexes, built on the same field:
index1: 'create bitmap index fbx_tab1_1 on tab1(substr((fieldx),1,1)) local
index 2: 'create bitmap index fbx_tab1_2 on tab1(substr((fieldx),2,1))
local tablespace ts_index;'
I didn't get any problem in creating them, but when I load new data in a new
empty partition and then perform 'alter table tab1 modify subpartition
new_subpartition rebuild unusable local indexes' the process hangs and the
alert reports the message above. After having dropped those indexes
everything it's ok; but since it would be useful to have them, how can
avoid this problem?