 |
 |
Archives of the TeradataForum
Message Posted: Tue, 21 Nov 2006 @ 23:44:09 GMT
| Subj: | | Re: Problem with first production use of PPI & JI |
| |
| From: | | Anomy Anom |
<-- Anonymously Posted: Tuesday, November 21, 2006 18:43 -->
It seems as though the entire process is taking longer, each step has extra time which leads me to believe it's a system wide issue. The
Fetch/Join into the staging table, the Merge into the target table and then the Merge into the Join Index on the Target table are all taking
longer. Systemusage shows that CPU and I/O consumed by the job are climbing week over week, even though the amount of data we load each day
really isn't changing that much. ResNode data doesn't show an system wide increase over the past two months.
It reminds me of the duplicate row check, the bigger the table gets, the more painful that dup row check is. This target table is a Mulitset
table so that shouldn't be an issue. What about the internal table that is the Join Index? Is it possible that it would have some duplicate row
check in it that is causing this extension of the run time? I can't help but think it has something to do with the PPI and the Join Index.
| |
