Re: Dealock on Rollback Segment (I think)

ID1 corresponds to the undo (rollback)segment number and slot used. These match the xidusn and xidslot columns from v$transaction. It is easier to see if you convert ID1 to HEX though.

ID2 is the undo sequence number - the XIDSQN of v$transaction. These three numbers between them identify uniquely the transaction and how to find the related rollback entries.

To examine the chained rows you can use block dumps. The 8.1 syntax is:

    alter system dump datafile NNN block min XXX block max YYY;

Once you have dumped a block with a chained row in it, you will find that the row entry has an item called NRID, which is the 'next rowid' - the restricted rowid where the next section of the row starts, so you can then dump the next block down the chain.

--
Jonathan Lewis
Yet another Oracle-related web site:  http://www.jlcomp.demon.co.uk

Practical Oracle 8i:  Building Efficient Databases

Publishers:  Addison-Wesley
See a first review at:
http://www.ixora.com.au/resources/index.htm#practical_8i
More reviews at: http://www.jlcomp.demon.co.uk/book_rev.html



nickvysh_at_chat.ru wrote in message <92m6q5$ats$1_at_nnrp1.deja.com>...


>

>> a)  If a process has to wait for an ITL slot to become empty,

>>      you get a TX lock requested in mode 4 with an ID1/ID2

>>      that matches the TX lock of the session it is waiting for.

>

>What do ID1 and ID2 mean (for TX lock) Can get any usefull info from

>these numbers?

>

>> The result was

>> Row 1:

>>     varchar2( 3904), varchar2(4000), varchar2(90)

>> Row 2:

>>     varchar2(96) - next rowid, same block row 1.

>>

>> In other words the first 96 bytes of column 1 of

>> my row had been inserted as one row in the block,

>> with the rest of the row the target of a chain into

>> the same block !!

>

>That's definitely strange. But how can you see, how oracle chains a

>row? LIST CHAINED ROWS gives you only head rowids of chaind rows,

>nothing else?

>

>

>

>Sent via Deja.com

>http://www.deja.com/