"Stephen B" <stephen.bell_at_cgi.ca> wrote in message news:<Coz38.13816$KC1.2314744_at_news20.bellglobal.com>...
> Oh yeah...hardly a day goes by when I don't do that :)
>
> "Mladen Gogala" <mgogala_at_earthlink.net> wrote in message
> news:pan.2002.01.23.03.00.36.167637.1171_at_earthlink.net...
> > On Wed, 23 Jan 2002 02:15:26 -0500, Thomas wrote:
> >
> > > Hi everybody,
> > >
> > > I have a simple question. How can i test weather a record in a table is
> > > locked by another user without causing another lock.
> > >
> > > regards
> > > thomas
> >
> >
> > That is very simple and I'm sure that the most of us here do that every
> day:
> > you find the database block for that record (rowid contains file and block
> > id), dump the block from oradebug and see whether the row header contains
> > the pointer to a valid ITL in that block. The very first byte of the row
> > header is the offset into the ITL table. ITLs are 23 bytes in size and
> > contain a XID of the transaction that is interested in the block. After
> > you do that, ask yourself whether the obtained information was worth the
> > trouble. Oh yes, and exact descriptions on the database blocks can be
> > obtained only in the Scott Gosset's "top gun" class.
> > Oh, did I forget to mention that you have to check whether the block is in
> > SGA (v$bh) and if it is, then you have to look there. ITLs in the disk
> > block may contain invalid information and be scheduled for "delayed
> > cleanup". It's a very simple thing indeed.
Could we make our answers a little bit clear and simpler for
those who may really need our help.?
Thanks
Received on Wed Jan 23 2002 - 12:46:11 CST
