Hello:

We're migrating our engine from Lucene7 to 9.
We're using Lucene9.8 and Microsoft JDK17

After the migration, we observed performance degradation on some hosts for
some applications. Profiling shows this is caused by the JIT deoptimization.
The scale is uncertain. For small applications, no such problem. For a
large application, only a few hosts, and restarting them can fix the
issue.  However for another larger application, the problem happens more
frequently; restarting would only work once, after another deployment the
problem can occur again.

Since there is no such issue while on Lucene7, it seems this problem has a
strong correlation with Lucene9.

I'm just wondering if anyone in the community has any experience on this,
and any suggestions?

Thanks

Reply via email to