CRC difference between conv89 and load df

[D. Verschuuren]

Hi,

From one db I dumped the df from the version 8.3 database and loaded the df in the new version 9 db.

The other db I used the conv89 tool from progress.

Between the two methods I saw the CRC difference.
I encountered a CRC difference on the SYSCOLUMNS table.

Did anyone else have this problem, or have seen this not expected hidden feature from Progress??

Many TIA, Dennis

 

[toby.Harman]

This is a problem that many people encounter!

In the earlier versions of Progress, deleted fields and indices caused CRC's to be different on tables when compared to newly loaded df's.

There were many discussions revolving around the use of a "golden Database" or "schema holder". Progress have addressed this by increasing the granularity on CRCs to reduce the likelihood of encountering this problem.

If you are going from v8 to v9 I would recommend a full dump and reload, once you have fully analysed your table usage (i/o) and record sizes. You'll be able to tune your database and I/O far better with records per block and areas.

Hope this helps