https://bugzilla.wikimedia.org/show_bug.cgi?id=34785
--- Comment #23 from Aaron Schulz <aschulz4...@gmail.com> 2012-05-22 20:12:09 UTC --- (In reply to comment #21) > They're using SQLite with no other options? That's stupid cause even if you > put > the DB on the fastest drive on the Earth, it will still suck because SQLite > had > never been intended for any serious level of concurrency, and it's not a bug > but a design decision. The concurrency limitations are probably dealt with by the fact that we shard the containers for the large wikis. This also mitigates the B-tree index update slowness (proportional to the no. of items in the container object list table). Since the container service and object service are running on the same boxes (and the later gets used way more and on way more data), no DB files can stay in the page cache. The random seeks can't be well handled by SATA drives. Asher also floating trying some hacks with http://www.kernel.org/doc/man-pages/online/pages/man2/posix_fadvise.2.html for the sake of investigation. -- Configure bugmail: https://bugzilla.wikimedia.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug. _______________________________________________ Wikibugs-l mailing list Wikibugs-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikibugs-l