Re: ZODB storage for cassandra
Jim Fulton <[email protected]>
| Newsgroups | gmane.comp.web.zope.zodb |
|---|---|
| Message-ID | <CAPDm-FggLmUXrEs4U6VgaUd6f6ShhfV1-c5nc3gEBXY5JaKjpg@mail.gmail.com> |
On Tue, May 16, 2017 at 3:30 AM, Sanjay Ra ... > I fully understand the point raised here, but I have a problem. I have an > existing zope application(lots of browser, SOAP, xmlrpc views) which my > employer does not want to rewrite. In this application we need massive > scaling at database writing. > OK, no changes and massive scaling. Check. We can give back seat to consistency or can handle consistency at > application level. > That would be a very bad idea. > Can you suggest any alternate solution ? > I already did, a while ago. It's unlikely with the sort of information you can post to a mailing list that people are going to be able to suggest the best solution to your problems. In the past however, you've mentioned having major problems with BTree conflicts, which I interpreted to be associated with indexing. I suggested switching to Newt DB and moving your indexes to Postgres. This would greatly decrease the amount of data you need to write and the likelihood of conflicts. Recently, I released newt.qbe, https://github.com/newtdb/qbe, which can help with converting catalog-based searches to Postgres. It's not a drop-in replacement for catalogs, but makes writing one straightforward. I'm in the process of converting a catalog-based app to use newt.db and newt.qbe. This app uses repoze catalogs and I've written a replacement that uses qbe to provide the same API they're using with their existing catalog. https://github.com/karlproject/karl/blob/master/ karl/models/newtqbe.py Jim -- Jim Fulton http://jimfulton.info -- You received this message because you are subscribed to the Google Groups "zodb" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.