On Thursday, 31 October, 2019 10:01, Dominique Devienne <ddevie...@gmail.com> wrote:
>On Thu, Oct 31, 2019 at 4:10 PM Thomas Kurz <sqlite.2...@t-net.ruhr> wrote: >> Yes, but I'd expect that MariaDB has to do the same, but takes clearly >> less than 1 minute instead of 88 minutes... :confused: >Are we comparing apples to oranges here? >SQLite runs on the local machine. While MariaDB is client-server, so the >delete effectively runs on the server. >Is your MariaDB server local to the same machine you're running SQLite >on? >Are both using the same local disks? (as opposed to networked and/or >different disks) >Still can't account for 88 minutes though. Well, 88 minutes is a lot better than 56 hours, of which 24 hours was "waiting" time. How much of the 88 minutes is "waiting" time? If the elapsed time is much greater than the sum of user+sys time then I would suspect it is still I/O thrashing (or the machine is really busy doing something else -- those original timings represent only a 57% process dispatch rate, which is pretty low). >You haven't shown the "explain query plan" Keith asked for, which would >help understand what's going on here. Well, having added the missing index I doubt that there is anything else to be found there. -- The fact that there's a Highway to Hell but only a Stairway to Heaven says a lot about anticipated traffic volume. _______________________________________________ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
