davisp commented on issue #1186: [WIP] Clustered purge mixed cluster upgrade URL: https://github.com/apache/couchdb/pull/1186#issuecomment-369707222 So thinking more on this, I think the best approach is to not upgrade the purge values to a tree until we get a purge request. Thus for the normal case when a customer has no purge activity before and after the upgrade/downgrade there won't be any issue. I'm also trying to decide if we should change the disk version for this instead of just altering the purge_docs field contents. The thing I'm most worried about at this point is what happens when we downgrade after purge requests have been issued over the new API. If we're doing this quickly then its likely that we'd have to invalidate all secondary indexes to be sure they're valid? Or else make some sort of check against all of the _local purge docs?
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
