If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below.

 
Go Back  dBforums > Database Server Software > DB2 > Index Creation During LOAD In UDB ESE w/DPF (aka EEE)

Reply
 
LinkBack Thread Tools Search this Thread Display Modes
  #1 (permalink)  
Old 05-28-04, 13:35
rubystep rubystep is offline
Registered User
 
Join Date: Sep 2003
Posts: 85
Question Index Creation During LOAD In UDB ESE w/DPF (aka EEE)

Since the LOAD runs external to the database and it builds the indexes externally, does the LOAD use any containers (be it from the coordinator node or any other node) such as TEMP tablespace?

The reason that I ask is that it seem sthat the build index phase seems to be taking more time than we'd expect.
Reply With Quote
  #2 (permalink)  
Old 05-28-04, 14:37
bmujeeb bmujeeb is offline
Registered User
 
Join Date: Mar 2004
Posts: 448
During the Build phase ,indexes are created on the index keys that were
collected and sorted duing the load phase.The using directory clause should
be used because the default sort path is sqllib/tmp.

for load you can use the db2split command to create many split files.
then use that files at that partition, performance is better in that case.

regards,

mujeeb
Reply With Quote
  #3 (permalink)  
Old 06-01-04, 09:20
rubystep rubystep is offline
Registered User
 
Join Date: Sep 2003
Posts: 85
Perhaps you could elaborate on where you know about sqllib/tmp for the sort?

What I have found in the Data Movement Utilities pdf is:

ALLOW READ ACCESS also supports the following modifiers:
USE tablespace-name
If the indexes are being rebuilt, a shadow copy of the index is
built in table space tablespace-name and copied over to the
original tablespace at the end of the load during an INDEX
COPY PHASE. Only system temporary table spaces can be
used with this option. If not specified then the shadow index
will be created in the same table space as the index object. If
the shadow copy is created in the same table space as the
index object, the copy of the shadow index object over the old
index object is instantaneous. If the shadow copy is in a
different table space from the index object a physical copy is
performed. This could involve considerable I/O and time. The
copy happens while the table is offline at the end of a load
during the INDEX COPY PHASE.
Without this option the shadow index is built in the same
table space as the original. Since both the original index and
shadow index by default reside in the same table space
simultaneously, there may be insufficient space to hold both
indexes within one table space. Using this option ensures that
you retain enough table space for the indexes.
This option is ignored if the user does not specify INDEXING
MODE REBUILD or INDEXING MODE AUTOSELECT. This
option will also be ignored if INDEXING MODE
AUTOSELECT is chosen and load chooses to incrementally
update the index.

I am also not specifying ALLOW READ ACCESS and I am issuing a LOAD ... REPLACE with indexes being auto-rebuilt and I'm gathering statistics.

Ruby
Reply With Quote
Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On