We currently use the Thinkwise 2023.3 product. It seems that in this version all indexes are being rebuilt when doing a smart upgrade.
In the current version of our application only one new table is introduced, but it takes much longer than in the previous version of the SF (2023.1) we used. Hundreds if of indexes are created, so the upgrade takes longer than expected.
Anybody else have experienced this? Is this new in this version and if so, why? Can't it be “smarter”?
Hi Henri,
The Smart upgrade checks the archive and compares it to the current state of the indexes. If a index is mutated in any way, the index will be recreated. It may have happened that the archive is not entirely complete and result in some false positives. This does indeed cause the upgrade deploy to run longer but will not cause any errors; it may even boost performance because the indexes are up-to-date similar to what the tsf_optimize procedure does.
Was this the first time of upgrading in the 2023.3 after the platform upgrade from 2023.1?
Yes. For this application it was the first time we installed a new release with version 2023.3 of the SF (using smart upgrade). So the next time(s) it wil be a reduced set of indexes, depending on changed tables I guess?
If so, then never mind my question….
That's what I would expect yes. The archive will be in a stable place from that point forward.
Enter your E-mail address. We'll send you an e-mail with instructions to reset your password.