Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Row Locks?

Re: Row Locks?

From: Jim Kennedy <kennedy-family_at_home.com>
Date: Wed, 06 Jun 2001 05:59:43 GMT
Message-ID: <jFjT6.118223$p33.2475894@news1.sttls1.wa.home.com>

If Oracle detected that the session crashed then Oracle performs a rollback on that session. Only reasonable thing it can do. In that case the lock would be released.
Jim

"Sunday Driver" <jane_austen_at_geocities.com> wrote in message news:3b1d3c3b.83223438_at_client.ne.news.psi.net...
> "Sybrand Bakker" <postbus_at_sybrandb.demon.nl> is rumoured to have said:
>
> >
> >"Sunday Driver" <jane_austen_at_geocities.com> wrote in message
> >news:3b1ced7e.63066474_at_client.ne.news.psi.net...
> >> Hi
> >>
> >> If 2 users try to update the same row at the same time, the one that
> >> gets there first gets to save and the other gets an error about the
> >> row having been modified by another user, right?
> >>
> >> My (other) question is:
> >> Has anyone ever heard of Oracle leaving some kind of permanent lock on
> >> the row so that all users after that get the error msg about the row
> >> having been modified by another user?
> >>
> >> If not, why would all future users get the error when they try to
> >> modify these rows?
> >>
> >> TIA
> >> Bryn
> >>
> >> "In such cases as these,
> >> a good memory is unpardonable."
> >
> >If a row is locked others will get the error message above.
> >The lock will remain until the first user commits
> >And before you start asking: there is no such thing as a timeout, so if
 the
> >first user goes out to lunch, the lock will remain virtually forever.
> >
> >Hth,
> >
> >
> >Sybrand Bakker, Oracle DBA
> >
> >The usual adminitions to read the manual apply
> >
> >
> >
> >
>
> Thanks, I appreciate the info. This lock has (according to my user)
> been in place for months, we're going to hunt for a computer under the
> stairs. :-)
>
> What would happen if the first user didn't ever commit the changes but
> rather crashed out somehow, leaving the row uncommitted yet changed?
> Is this even possible? Would that result in the error presenting to
> all users after that?
>
> Thanks again for your help.
> Bryn
>
> "In such cases as these,
> a good memory is unpardonable."
Received on Wed Jun 06 2001 - 00:59:43 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US