How does elasticsearch handle the case where IO just freezes to a single disk and the path doesn't disappear (cached entries can be read, some writes may succeed until they are fsynced etc)?
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.