Re: management paranoia about 2 million rows updated in 1 commit

From: Dba DBA <oracledbaquestions_at_gmail.com>
Date: Wed, 22 Aug 2012 14:44:03 -0400
Message-ID: <CAE-dsOJpQJ6vZDn7+qW8Th9-zJR76cJ2t_t7EYWV0bXwFEKdPw_at_mail.gmail.com>



My question is basically, what do you say to managers when they say something like ;'we have done bulk updates before and found that if we have to kill them they sometimes take a really long time to rollback'. No idea who wrote, it what it was doing, what the situation was, etc... This kind of blanket statement is not uncommon.
In this case, we have an extended outage, plenty of undo space, and if this for some reason needs to rollback it won't block anything. At worst, we shutdown abort, and startup with a recovery. Note, if I mention 'shutdown abort' they will freak, so I probably won't say that.

This is for other people who have run into this kind of stuff from managers. What have you said to them that actually worked? I don't know how to return to their vague concern.

On Wed, Aug 22, 2012 at 11:53 AM, rjamya <rjamya_at_gmail.com> wrote:

> So you dont have a technical issue, but a management issue. I'd still
> go with single statement update. With multi sessions, batched commits,
> if it rolls back, you will spend hours ot find which ones were updated
> and which ones were not for the subsequent runs.
>
> Raj
>
> On Wed, Aug 22, 2012 at 11:41 AM, Dba DBA <oracledbaquestions_at_gmail.com>
> wrote:
> > any suggestions on what to say?
>

--
http://www.freelists.org/webpage/oracle-l
Received on Wed Aug 22 2012 - 13:44:03 CDT

Original text of this message