Big update makes Oracle stop responding

Hello there,
we have a table named TBL_TREE having 11 columns (10*number, 1*date) with 4 millions rows (about 400MB). Table has 5 indexes (together 800MB). Sometimes we need to delete about 30000 rows from the table or insert such amount of rows into it. When this operation starts database stopps responding to other queries. It's even impossible to connect to instance or (when already connected) perform query which works with real tables. V$ views works so so.

Database is 9i SE running on Linux (SuSE) server with 1GB memory, multiprocessor Intel architecture, archiving is on, redologs have 40MB each (5 groups), logswitch each 10-12 minutes when load is high, 40-60 minutes when load is normal. Redo buffer 512KB.

I understand the database has to do a lot of IOs when modifying large table and its indexes but I expect other sessions still can execute their queries (not as fast as usually, of course). Unfortuntely, database looks like dead for 10 minutes.

Can you give me a clue what's wrong? What should I check? Received on Mon Dec 15 2003 - 13:19:13 CST