October 2021 - discuss - lists.mariadb.org

[Maria-discuss] MyISAM: single table GROUP BY plan changes based on LIMIT. BUG?
by Vassilis Virvilis 26 Nov '21

26 Nov '21

2 4

[Maria-discuss] How can I separate one test from mysql_test
by Lukas Javorsky 19 Nov '21

19 Nov '21

Hi guys, I've found few particular tests from the test suite that is failing due to the OpenSSL-3 version and I want to separate them and give them to the OpenSSL team in Red Hat. The idea behind this is to give them a set of commands/config options (reproducer) for each test and they will try to find out what's the exact problem there and help us develop the patch. The tests that I'm interested in are (there are more of them, I just need to see how to reproduce few of them and I'll do the same for the other ones as well): encryption.corrupted_during_recovery encryption.innodb-checksum-algorithm encryption.innodb-discard-import-change encryption.innodb_page_encryption_key_change binlog_encryption.rpl_loaddata_local binlog_encryption.rpl_parallel_free_deferred_event IMPORTANT: These tests are failing only with the 'cbc' cipher in use, the 'ctr' is passing. Could you please help me with how to make a reproducer for each test, so I can create them and give them to the OpenSSL team? Thank you for any help you can provide Lukas -- S pozdravom/ Best regards Lukáš Javorský Associate Software Engineer, Core service - Databases Red Hat <https://www.redhat.com> Purkyňova 115 (TPB-C) 612 00 Brno - Královo Pole ljavorsk(a)redhat.com <https://www.redhat.com>

3 14

[Maria-discuss] Incorrectly formatted XML files in mariadb-server sources
by Zuzana Miklankova 26 Oct '21

26 Oct '21

Hi, Fedora's rpminspect tool reports [1] some incorrectly formatted XMLs (most of them used for testsuite) within mariadb source files, namely they are: mysql-test/plugin/sphinx/sphinx/testdata.xml - undefined namespace mysql-test/std_data/loadxml.dat - different parser errors, seems to me like it is ill formed on purpose of the test mysql-test/plugin/connect/connect/std_data/nocs.xml - not proper utf-8 encoding mysql-test/std_data/loaddata/mdev9874.xml - not proper utf-8 encoding sql/share/mariadb/charsets/Index.xml - redefined attribute mysql-test/std_data/ldml/Index.xml - redefined attribute It does not seem (to me) that the ill-formatting breaks anything, and that in some cases it is desired. Specifically the loadxml.dat XML, where the test case counts with the form as it is, and Index.xml files, where modification to the XML format could change the meaning of the file. In other cases, it seems (to me) that the correction of the XMLs to a valid format could be possible. These are specifically the XMLs with encoding issues, where it should be sufficient to specify a used charset (ISO-8859-1 in both cases). This is also the case with the undefined namespace in testdata.xml, whose definition should be also possible. Would the correct approach to this issue be to suppress the rpminspect warnings, or should the specified XML files be modified to a valid XML form? Thanks for your opinions in advance. [1] https://osci-jenkins-1.ci.fedoraproject.org/job/fedora-ci/job/rpminspect-pi… Zuzana Miklánková Associate Software Engineer Red Hat <https://www.redhat.com> <https://www.redhat.com>

2 2

[Maria-discuss] MariaDB 10.4 shows a 2e308 cost on a optimization plan correctly preferred by 10.0
by David Strauss 25 Oct '21

25 Oct '21

I'm seeing a regression in query optimization after testing a database+query that was running well on MariaDB 10.0. The issue occurs on an installation of 10.4.18. Version 10.4 considers the plan that 10.0 used but dismisses it as having a cost of 2e308 [1]. The plan 10.4 chooses is highly non-optimal. Based on reviewing some source code, it looks like this sort of "max" cost should only result if summing/multiplying other costs would exceed that maximum. The usual steps of rebuilding tables and index statistics don't help, especially because the JSON output for the optimization plan already shows reasonable index use costs. (The 2e308 cost only starts appearing in a later section.) Because 10.0 lacks the same optimization tracing as 10.4, here is the EXPLAIN output for comparison [2]. Hopefully, there's just something simple amiss and posting here helps anyone else who encounters a 2e308 cost on query optimization plans. I tried to see if this may be fixed in 10.5+, but I haven't found much discussion online of this phenomenon. [1] https://gist.github.com/davidstrauss/becb03fa75ecac046cafe6f05da9f869#file-… [2] https://gist.github.com/davidstrauss/becb03fa75ecac046cafe6f05da9f869#file-…

1 0

[Maria-discuss] State of "alternative" engines and mariadb
by pslawek83 19 Oct '21

19 Oct '21

Hi Guys, what's the state of "alternative" engines (aria and rocks). AFAIR there were some plans to make aria much more advanced than myisam, transactional (transaction safe?) and more optimized for todays hardware. What's the status of RocksBD, basically on the mariadb website i can't find recent updates and most of sources on the internet are outdated. Is it 100% compatible with mariadb nowdays, is there a possibility to have "one-file-per-table" (one set of files) or at least per-database? If everything resides in shared space is there a possibility that one write error (eg. resulting from power loss) can make all rocks tables unreadable? Is it possible to efficiently drop eg. 100GB table partition if the table is using rocks? Last time i checked there were some issues with compaction for Rocks (some limit on DB size, which was rather small can't recall this currently). Is that resolved, and how much additional space is required to compact eg. 1tb database? How's the adoption and developement of both engines going? Are people using it instead of myisam/toku/inno? Is there a lot of effor put into making aria/rocks better? As i'll have to update many servers in the future what's generally the preferred way to go? Drop everything and just use innodb like Oracle did or move towards rocks/aria for everything which doesn't play with innodb well? Another interesting topic is that myisam is having performance issues because of spectre/meltdown mitigations and it seems to hit this engine the most (at least in 10.3, 10.4 as far as I recall). Are there some plans to optimize it for new CPUs (maybe it was already done) or all the effors are going into different storage engines and i should abandon myisam long term? Thanks

2 1