Re: [I] Server panics/aborts on EMFILE (Too many open files) during index rebuild instead of failing gracefully [iggy]

["numinnex (via GitHub)" <gi...@apache.org>]

numinnex commented on issue #3645:
URL: https://github.com/apache/iggy/issues/3645#issuecomment-6040981639

   Closing as fixed on master.
   
   The panic reported here came from the legacy server's bootstrap, where an index rebuild I/O error was unwrapped into a `panic!`. That code, including `core/server/src/compat/index_rebuilding`, was removed when server-ng became the server in #3856.
   
   Running out of file descriptors is now handled by #4318: the server raises its soft `RLIMIT_NOFILE` at startup, segment recovery returns structured errors instead of panicking (and writes a rebuilt index to a staging file before renaming it into place), and a boot that fails because descriptors ran out exits cleanly with status 4. At runtime the partition reconciler logs the failure and retries. `core/integration/tests/server/descriptor_exhaustion_vsr.rs` covers the end-to-end case.
   
   @jiengup the inverted subtraction you spotted in `index_rebuilder.rs` (`start_offset - header.offset`) was a real bug too, but it went away with that file.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@iggy.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org