Archives of the TeradataForum

Message Posted: Mon, 01 Mar 2004 @ 13:38:55 GMT


	<Prev	Next>		<<First	<Prev	Next>	Last>>

Subj:		Re: Delete Query on a huge table taking too much time

From:		Victor Sokovin

I had run this query earlier also, it takes rime but this time it has been running for last 32 hours. PF state is being shown as DISPATCHING. I don't want to kill this job as I am afraid if I kill the job, rollback will take more time. Please suggest What to do in such scenario and what could be the reason behind this.

You have basically two options:

1. Wait until the query is over.

2. Kill the jobs, then kill the rollback.

Just a few words about them. To find out how long the query will take monitor the number of rows deleted per hour. If you see that the waiting time is too long, then consider killing the job. The rollback will kick off after that. You are right that it will take a very long time, perhaps 50-60 hours in your case.

The good news is that it is possible to kill the rollback! In V2R5, it is made easier for a DBA to do that but even in V2R4 your NCR contact should be able to help. Please note that this is not an advertisement for their services but I know they can help with such issues very efficiently so I thought I should mention that. As far as I know they do not guarantee that the data stored in the table before the delete or the SI will be OK but often they will be. If it is damaged, then you could restore the table from the last backup, if available. The restoration of the table should take just a fraction of the rollback time, perhaps just a few hours on your system. You should be back in business the same day.

Regards,

Victor


	<Prev	Next>		<<First	<Prev	Next>	Last>>

Attachments

Library

Quick Reference