Archives of the TeradataForum

Message Posted: Thu, 09 Nov 2001 @ 03:44:47 GMT

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

I'm looking for someone who has some experience with doing restores after a hardware upgrade.

We're in the process of doing a hardware upgrade that involves replacing a 1 node cabinet with a two node cabinet - which means a sysinit. Yikes!

I'm trying to get a feel for how long my restores are going to take. NCR has suggested 5x to 10x longer than the archive. I think it's extreme. I can't have a 60 hour restore (some of my archives are 6 hours.

Last year we did an upgrade that required a sysint (you think I'd be good t them by now). My restores took about the same amount of time as my archives, but were running on a machine 3 times faster (1 node to 3 nodes).

The restores from the last upgrade had messages as follows:
*** Warning 3803:Table 'B_MB_DETAIL' already exists.
NEED TO RE-HASH ROWS
STARTING TO RESTORE TABLE "HBCADM"."B_MB_DETAIL"

I've been doing some archives and restores between my development and production boxes. I expected to see messages about rehashing when doing a copy onto one box that was archived on the other. (My thinking being, I could get an estimate on the rehash time).

I don't see the rehash message I expected.

I would have expected the same kind of message restoring from one system to another - after all they have a different hash map.

I see no time difference between restoring a table from an archive of tself to restoring a table from an archive from a different system. I expected it to take longer coming from another box.

Does anyone have any experience with re-hashing times on a restore?

Ruth Fenwick

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