Subject | Re: [Firebird-Architect] Re: Can we, can we, can we????... |
---|---|
Author | Ann W. Harrison |
Post date | 2005-06-17T16:50:41Z |
Alexander Klenin wrote:
that could be improved by quotas. I can certainly imagine situations
where quotas make managing an application more difficult, so I need a
very concrete reason to go there - not just because we can.
and kill those it considers too long.
runaway queries.
Regards,
Ann
>Granted - but I'd need to hear from somebody with a real application
> 1) Quotas are NOT a replacement for on-demand request cancelling, they
> are different, although related, features appropriate for different
> use cases.
that could be improved by quotas. I can certainly imagine situations
where quotas make managing an application more difficult, so I need a
very concrete reason to go there - not just because we can.
>A Web application can (should?) do its own monitoring of user queries
> 2) ...
> More dangerous are run-away queries initiated by users, ...
> The "cancel" button is good, but not a perfect solution
> a) it does not work in web interfaces
and kill those it considers too long.
> b) there may be situation when user is content to wait a few minutes,And probably we need a way for an administrator to identify and stop
> but the query is so heavy that it blocks all other database activity
runaway queries.
> c) there may be security implications in trusting users to cancelSuch as?
> their own requests
>The kill-my-request call works fine in that case.
> 3) Lastly, even in the case of developer/dba, who is for some reason
> forced to run untested requests againts production database...
Regards,
Ann