We are running Sitecore 9.0.0 (Initial Release) on Azure with SolrCloud (6.6.2) and getting close to go-live.
We are using Sitecore's
SwitchOnRebuild functionality for indexing so that the indexes should be refreshed in the background. However, we are seeing an issue with the swapping of those indexes which means that the indexes are not updated correctly and various 404 errors are presented to the end-user because of failed lookups.
We can also see clear differences between the index and rebuild indexes via a direct query in Solr.
Initially, Sitecore thought this was a timeout issue (because there was a clear error in the logs relating to this) and we've increased the period from the default to 30 minutes. We are no longer seeing the timeout.
We are also using
EnforceAliasCreation and have multiple SolrCloud nodes, both of which are highlighted as potential causes of this behaviour and fixed in 9.0.2. Sitecore has suggested however that the release notes on these two issues are misleading.
Has anyone seen/solved this before? Or any general pointers would be greatly appreciated - Sitecore support seem a bit stumped so far.