There is another way. What you are referring to is XPDL restrictions in ASE. As Mark pointed out, you could use an intermediate (e.g. a dev/test or QA) system of the same platform and load the dump/redump quiesced there.....however, you still have the nasty sp_post_xpload fun that can take hours...plus you still are restricted to the same pagesize & locking as the primary. Some customers have found it easier to simply do a bulkcopy method of bulk materialization - copy the schema, bulk load the data most of the data (e.g. historical), create the pkey indexes, and add the other indexes as necessary. They then simply use dsi_command_convert to globally put SRS into an autocorrect style mode and then start replication up. At the primary, they simply use rs_ticket and when the ticket arrives at the replicate, they can turn dsi_command_convert off.
↧