Archives of the TeradataForum

Message Posted: Wed, 05 Jul 2006 @ 12:54:10 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

<-- Anonymously Posted: Wednesday, July 05, 2006 07:55 -->

Hi,

1) Each Teradata Database table has an associated subtable called a table header. One copy of each table header is stored on each AMP in the system, having one row per table per AMP. Table headers are composed of a row header, one fixed-length field, and six variable-length Fields. Field 4 of the table header contains the Multiload,Fastload and Archive/Recovery information.

Following are the uses of table header for Multiload.

Table header will be updating at each phase of load job. This allows DBS to restore the job if it fails in midway.

Utility locks are applied to the target table headers. This locks disallow any ddl to the table (except drop).

Locks placed on table header alerts other utilities that a multiload is in session for this table.

If the load is successful, table header locks are released across all AMPs and the field4 of the table header is cleared. Table headers will be restored to their original status across all AMPs.

2) NUSIs are implemented on an AMP-local basis. Each AMP manages the NUSI subtable rows that correspond to base table rows located on that AMP only. Since NUSIs allow duplicate index values and are based on different columns than the PI, data rows matching the supplied NUSI value could appear on any AMP.

In a NUSI retrieval a message is sent to all AMPs to see if they have an index row for the supplied value. The AMPs with matching rows use the "pointers" in the NUSI subtable row to access their corresponding base table rows. Any AMP that does not have an index row for the NUSI value will not access the base table to extract rows.

Thanks,

Anomy.

	<Prev	Next>		<<First	<Prev	Next>	Last>>