Hi all, Here is a summary of what has happened since this thread has been created. Three problems reported on this thread have been solved and resulted in different commits for early lock lookups: - VACUUM FULL, patched on 12~: https://www.postgresql.org/message-id/20180812222142.ga6...@paquier.xyz Commit a556549: Improve VACUUM and ANALYZE by avoiding early lock queue - TRUNCATE, patched on 12~: https://www.postgresql.org/message-id/20180806165816.ga19...@paquier.xyz Commit f841ceb: Improve TRUNCATE by avoiding early lock queue - REINDEX, patched on 11~: https://www.postgresql.org/message-id/20180805211059.ga2...@paquier.xyz Commit 661dd23: Restrict access to reindex of shared catalogs for non-privileged users
Please note that I have been very conservative with the different fixes as v11 is getting very close to release. The patch for REINDEX is a behavior change which will not get further down anyway. It would still be nice to get a second lookup at the code and look if there are other suspicious calls of relation_open or such which could allow non-privileged users to pile up locks and cause more DOS problems. Thanks, -- Michael
signature.asc
Description: PGP signature