Wondering what folks think of backporting some classloading deadlock-related fixes: https://issues.apache.org/jira/browse/SOLR-16165 and possibly upgrade to Lucene 9.4 https://issues.apache.org/jira/browse/SOLR-16442 particularly considering: https://github.com/apache/lucene/pull/11718
The main concerns for all the above are classloading issues. The first one (SlotAcc) I think should be backported, since we know it has manifested in the wild and it's a very self-contained fix. I gather based on SOLR-15254 that the deadlock that would be fixed by Lucene 9.4 upgrade has also manifested; in fact I'm surprised that we haven't seen this issue manifesting more frequently -- maybe it has, but it doesn't get reported/diagnosed because it's an issue that's mostly likely to manifest on startup (when actual people are likely to be paying attention), and is easily "fixed" in the short-term by restarting (again)?. There's some discussion at: https://issues.apache.org/jira/browse/SOLR-16167 and https://issues.apache.org/jira/browse/SOLR-15254 I wouldn't want to do anything that would substantially delay 9.1.1 though... Thoughts? On Fri, Dec 30, 2022 at 6:17 PM David Smiley <dsmi...@apache.org> wrote: > > Still looking for a PR approval for the protobuf upgrade: > https://github.com/apache/solr/pull/1252 --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@solr.apache.org For additional commands, e-mail: dev-h...@solr.apache.org
