Hey all, We've got a good 30ish sites on Umbraco cloud ranging from a single V8 site all the way to V13 sites. On our V9+ sites, we are noticing an increased downtime and all of their logs appear to be showing:

System.ObjectDisposedException: this Directory is disposed.
Object name: 'Lucene.Net.Store.MMapDirectory'.
   at Lucene.Net.Store.BaseDirectory.EnsureOpen()
   at Lucene.Net.Store.FSDirectory.CreateOutput(String name, IOContext context)
   at Lucene.Net.Index.SegmentInfos.Write(Directory directory)
   at Lucene.Net.Index.SegmentInfos.PrepareCommit(Directory dir)
   at Lucene.Net.Index.IndexWriter.StartCommit(SegmentInfos toSync)
   at Lucene.Net.Index.IndexWriter.PrepareCommitInternal()
   at Lucene.Net.Index.IndexWriter.CommitInternal()
   at Lucene.Net.Index.IndexWriter.CloseInternal(Boolean waitForMerges, Boolean doFlush)
   at Lucene.Net.Index.IndexWriter.Dispose(Boolean waitForMerges)
   at Examine.Lucene.Providers.LuceneIndex.Dispose(Boolean disposing)

or

System.ObjectDisposedException: Cannot access a disposed object.
Object name: 'snapshot'.
   at Umbraco.Cms.Infrastructure.PublishedCache.ContentStore.Snapshot.Get(Int32 id)
   at Umbraco.Cms.Infrastructure.PublishedCache.ContentCache.GetById(Boolean preview, Int32 contentId)
(Truncated because of length)

Has something changed infrastructure wise like hot swapping of the sites or am I isolated in this issue?

Have the same issue with v13 sites today. @Chancer did you ever solve this?

Hey @Mark Drake We never actually found out the cause but indeed the solution was to nuke the Indexes and force a rebuild. We've only had it a handful of times since this particular day, but the issue certainly is still creeping around in the background. Wish we could figure out why it happens

Did you figure what happens? Im experiencing it on a azure setup not umbraco cloud. we think it happens everytime we swap the new deploy and we need a restart

Just a FED guy. But Umbraco Cloud for one of our sites has been relatively unstable the moment we hit 20k+ content nodes. What seems to happen is the index becomes corrupt and never rebuilds, or server runs out of memory during rebuild and thus the indexes fail until we delete everything (through Kudu) and restart the application (through Umbraco IO). We solved this, ironically, by moving a large portion of content into a custom DB table… and no longer being editable through the CMS. There’s not enough info in the server logs for me to really understand the issue better I’m afraid.

Hi @Mark Drake Thanks for the quick answer 🙂 We are not running on the cloud but on our own Azure setup, so we could probably do something with the server configuration. I would be happy not to move the content nodes to a custom DB table. We have 12k+ content nodes because it's a multi-site solution. We did not have this problem before we upgraded from 8 to 13. We suspect that the issue is caused by the production mode.