Reverse Engineering very slow
Hi,
Firstly, I am running the ODI 11G client at the other end of a WAN, with the databases in an other country :-)
ODI Repository is on the DW located in the same datacentre as the source database.
But still, even when submitting Custom Reverse engineering via the Agent , its very slow.
Are there any suggestions to speeding this up somewhat, increasing the array fetch size or batch update size in the JDBC or hacking the RKM to use a db link (from DW to source db) to access the data dictionary?
Has anyone tried either approach, has anyone increased the JDBC sizes from the default to good effect ?
Im tempted to create all source tables locally, import them to a local ODI and then export / import the work repos at this rate. I have 30 models to RE and they are running at about 30 mins per model.
Cheers
Alastair