- discuss - lists.mariadb.org

[Maria-discuss] mariadb 10.4.13-lp152.1.1 to 10.5.12-3.9.1 upgrade failure
by Marc Chamberlin 25 Nov '21

25 Nov '21

Hello - I have ran into a problem that maybe someone else has hit and knows a solution for it. I recently upgraded my OpenSuSE 15.2 x64 version to OpenSuSE 15.3 x64 and installed all additional updates and patches that were suggested for 15.3. Most of the upgrade succeeded, all except Mariadb. Google has not been helpful either. My understanding of Mariadb is that after a new install of a new version it is suppose to automagically update and fix the databases from the older version, when the server is restarted, but I think this is failing. I found the upgrade log file for Mariadb and it appears to be reporting some kind of corruption has occurred. This is preventing the mariadb (mysql) server from even starting, so it is not possible to use queries to find and repair the database. - > bigbang:/var/tmp/mysql-protected.lBP2YQ # more log_upgrade_run.err > 2021-10-31 12:11:36 0 [Note] InnoDB: !!! innodb_force_recovery is set > to 1 !!! > 2021-10-31 12:11:36 0 [Note] InnoDB: Uses event mutexes > 2021-10-31 12:11:36 0 [Note] InnoDB: Compressed tables use zlib 1.2.11 > 2021-10-31 12:11:36 0 [Note] InnoDB: Number of pools: 1 > 2021-10-31 12:11:36 0 [Note] InnoDB: Using crc32 + pclmulqdq instructions > 2021-10-31 12:11:36 0 [Note] mysqld: O_TMPFILE is not supported on > /tmp (disabling future attempts) > 2021-10-31 12:11:36 0 [Note] InnoDB: Using Linux native AIO > 2021-10-31 12:11:36 0 [Note] InnoDB: Initializing buffer pool, total > size = 134217728, chunk size = 134217728 > 2021-10-31 12:11:36 0 [Note] InnoDB: Completed initialization of > buffer pool > 2021-10-31 12:11:36 0 [ERROR] InnoDB: corrupted TRX_NO 5f737461748dea16 > 2021-10-31 12:11:36 0 [Note] InnoDB: Retry with innodb_force_recovery=5 > 2021-10-31 12:11:36 0 [ERROR] InnoDB: Plugin initialization aborted > with error Data structure corruption > 2021-10-31 12:11:36 0 [Note] InnoDB: Starting shutdown... > 2021-10-31 12:11:36 0 [ERROR] Plugin 'InnoDB' init function returned > error. > 2021-10-31 12:11:36 0 [ERROR] Plugin 'InnoDB' registration as a > STORAGE ENGINE failed. > 2021-10-31 12:11:36 0 [Note] Plugin 'FEEDBACK' is disabled. > 2021-10-31 12:11:36 0 [ERROR] Unknown/unsupported storage engine: InnoDB > 2021-10-31 12:11:36 0 [ERROR] Aborting On a hunch, I decided to install the OpenSuSE OSS repository for 15.2 and this allowed me to drop back and install version 10.4.13-lp152.1.1 of the mariadb server, client, errormessages, and tools packages. (OpenSuSE 15.3 wants to install version 10.5.12-3.9.1 of these mariadb packages.) And lo and behold the previous 10.4.13-lp152.1.1 package works fine! So what gives? Any idea why the upgraded version of mariadb packages are failing to upgrade and support my databases? As I said I have tried Googling for an answer but no joy finding one. A lot of the proposed solutions, Google found, assume the mariadb server is up and running (hump!) so the example solutions given, use mysql queries and commands to fix corrupt databases and tables. I should also point out that AFAIK most, if not all of my database structures are using InnoDB and not the older MYISAM structures (although the mysql database itself appears to have both). I did make a mariadb backup of all the databases but without the ability to start the mariadb server I don't know how to make use of this backup file to reconstruct all the databases and their tables etc. Thanks in advance for any and all advice on how to upgrade my mariadb server. Marc Chamberlin -- *"The Truth is out there" - Spooky* -- *_ _ . . . . . . _ _ . _ _ _ _ . . . . _ . . . . _ _ . _ _ _ . . . . _ _ . _ . . _ . _ _ _ _ . _ . _ . _ . _ . * Computers: the final frontier. These are the voyages of the user Marc. His mission: to explore strange new hardware. To seek out new software and new applications. To boldly go where no Marc has gone before! (/This email is digitally signed. My public key for sending encrypted email to me can be found at - https://keys.openpgp.org/search?q=marc@domesweetdome.us.com or just ask me for it and I will send it to you as an attachment. If you don't understand, no worries, just ignore it and/or ask me to explain it further./)

5 12

[Maria-discuss] Failed DDL on Galera
by William Edwards 23 Nov '21

23 Nov '21

Hi, A customer ran an 'ALTER TABLE' query, which failed. This caused Galera slaves to shut themselves down: 2021-11-23 9:35:25 1 [ERROR] Slave SQL: Column 2 of table 'inserve_fpfvxiqm.invoices' cannot be converted from type 'date' to type 'int(10) unsigned', Internal MariaDB error code: 1677 2021-11-23 9:35:25 1 [Warning] WSREP: RBR event 17 Update_rows_v1 apply warning: 3, 28891485 2021-11-23 9:35:25 1 [Warning] WSREP: Failed to apply app buffer: seqno: 28891485, status: 1 at /home/buildbot/buildbot/build/galera/src/trx_handle.cpp:apply():353 [ 3 more times ] 2021-11-23 9:35:25 1 [Warning] WSREP: RBR event 17 Update_rows_v1 apply warning: 3, 28891485 2021-11-23 9:35:26 1 [ERROR] WSREP: Failed to apply trx: source: 41c479ab-0527-11ec-9fcc-9f02b1141194 version: 4 local: 0 state: APPLYING flags: 1 conn_id: 13475392 trx_id: 63620823 seqnos (l: 312653, g: 28891485, s: 28891454, d: 28891449, ts: 10603037948851154) 2021-11-23 9:35:26 1 [ERROR] WSREP: Failed to apply trx 28891485 4 times 2021-11-23 9:35:26 1 [ERROR] WSREP: Node consistency compromised, aborting... The node on which the query was executed, crashed ("mysqld got signal 11"). At first sight, this seems to be the result of "WSREP: Send action {(nil), 528, TORDERED} returned -107 (Transport endpoint is not connected)" - as all the slaves were down at this point. As 'ALTER' is a DDL statement that does not support rollback, I understand that the only option for Galera may be to abort. But is there a way to handle something like this more gracefully? -- With kind regards, William Edwards

1 0

[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] Mariadb 10.2 - 10.6 mysqld_multi problem
by Tom Dangler 18 Nov '21

18 Nov '21

We're working on upgrading a mariadb-10.1 server to mariadb-10.6. We run multiple instances of Mariadb on this server, and some run different versions of Mariadb for different reasons. When we attempted to upgrade to mariadb-10.2, we noticed that we can no longer specify mysqld=/path/to/mysqld without generating an error and not starting the instance. We get [ERROR] /opt/mariadb10_2/bin/mysqld: unknown variable 'mysqld=/opt/mariadb10_2/bin/mysqld'. We tested this with 10.2 and 10.6 binaries. When using 10.1 binaries for the [mysqld_multi] group, it accepts the mysqld= argument just fine, and starts each instance with the version defined in mysql= in each option group. The documentation still shows that mysqld= is still a valid configuration, but it does not seem to work from 10.2 to current. Is there a different configuration for this now? We would like to continue using mysqld_multi with different versions of Mariadb. Using 10.1 binaries for mysqld_multi it works [mysqld_multi] mysqld = /opt/mariadb10_1/bin/mysqld_safe mysqladmin = /opt/mariadb10_1/bin/mysqladmin user = multi_admin password = ourpassword Using 10.2 - 10.6 binaries for mysqld_multi we get errors about unknown variable mysqld [mysqld_multi] mysqld = /opt/mariadb10_6/bin/mysqld_safe mysqladmin = /opt/mariadb10_6/bin/mysqladmin user = multi_admin password = ourpassword [mysqld1] mysqld = /opt/mariadb10_1/bin/mysqld datadir = /opt/mysql_data/mariadb10_1 port = 3306 socket = /var/run/mysqld/mariadb_10_1.sock pid-file = /var/run/mysqld/mariadb_10_1.pid user = mysql [mysqld2] mysqld = /opt/mariadb10_2/bin/mysqld datadir = /opt/mysql_data/mariadb10_2 port = 3307 socket = /var/run/mysqld/mariadb_10_2.sock pid-file = /var/run/mysqld/mariadb_10_2.pid user = mysql [mysqld6] mysqld = /opt/mariadb10_6/bin/mysqld datadir = /opt/mysql_data/mariadb10_6 port = 3308 socket = /var/run/mysqld/mariadb_10_6.sock pid-file = /var/run/mysqld/mariadb_10_6.pid user = mysql -- Tom Dangler IT Architect and Systems Manager T 214.778.3158 | M 214.493.0386 | F 888.811.1566 tom.dangler(a)oakwoodsoftware.com Oakwood Software Insurance Solutions 6201 W Plano Parkway | Suite 225 | Plano, TX 75093 Facebook <https://www.facebook.com/OakwoodSoftware> | LinkedIn <https://www.linkedin.com/company/oakwoodsoftware> | Twitter <https://twitter.com/TheRealOakwood> | oakwoodsoftware.com <http://www.oakwoodsoftware.com/> An Amwins Company

1 0

[Maria-discuss] JOB | Permanent DBA in Quebec
by James Tobin 15 Nov '21

15 Nov '21

Contact me using "jamesbtobin (at) gmail (dot) com" to learn more

1 0

[Maria-discuss] Timestamp column not inserting UTC
by William Edwards 13 Nov '21

13 Nov '21

Hi, I have a `timestamp` column, because I would like to insert timestamps in UTC. The documentation says that the inserted value should be converted from my session's time zone to UTC, but I can't reproduce that behaviour. I'm on MariaDB 10.5.9. I have a column of type `timestamp`: `created_at` timestamp NULL DEFAULT NULL The documentation at https://mariadb.com/kb/en/timestamp/ says: "If a column uses the TIMESTAMP data type, then any inserted values are converted from the session's time zone to Coordinated Universal Time (UTC) when stored, and converted back to the session's time zone when retrieved." However, I cannot reproduce this intended behaviour. I set my time zone as follows. (The value was 'SYSTEM', which should also be '+01:00', but I'm setting it explicitly just in case.) -- MariaDB [db]> set time_zone = '+01:00'; Query OK, 0 rows affected (0.006 sec) MariaDB [db]> select now(); +---------------------+ | now() | +---------------------+ | 2021-11-13 14:56:09 | +---------------------+ 1 row in set (0.001 sec) MariaDB [db]> SELECT @@global.time_zone, @@session.time_zone; +--------------------+---------------------+ | @@global.time_zone | @@session.time_zone | +--------------------+---------------------+ | SYSTEM | +01:00 | +--------------------+---------------------+ 1 row in set (0.000 sec) -- Then, I update the values in the `timestamp` column to the timestamp of my session's time zone: -- MariaDB [db]> update objs set created_at=now(); Query OK, 2 rows affected (0.001 sec) Rows matched: 2 Changed: 2 Warnings: 0 -- ... I would expect the inserted value to be converted from my session's time zone ('+01:00') to UTC. However, the inserted value is actually a timestamp of my session's time zone: -- MariaDB [db]> select created_at from objs; +---------------------+ | created_at | +---------------------+ | 2021-11-13 14:56:42 | | 2021-11-13 14:56:42 | +---------------------+ 2 rows in set (0.000 sec) -- I would not be surprised if I'm looking over something very obvious, but I haven't figured it out yet. Hopefully someone sees what I'm missing. TIA! -- With kind regards, William Edwards

1 1

[Maria-discuss] Clarification on Memory utilization of Mariadb
by ragul rangarajan 10 Nov '21

10 Nov '21

Hi Team, Need clarification on Memory utilization of Mariadb. I have a java application with MariaDB that runs on the same Rhel 7 server with physical memory(RAM) allocated as 12GB where I would see the normal utilization at the start as below. *sys> free -h* * total used free shared buff/cache available* *Mem: 11G 7.8G 1.6G 9.0M 2.2G 3.5G* *Swap: 0B 0B 0B* Over a period of time say >6month, I could see the RAM utilization is almost full, and available space is very less but still, the application continues to work fine without any issues. *sys> free -h* * total used free shared buff/cache available* *Mem: 11G 11G 134M 17M 411M 240M* *Swap: 0B 0B 0B* I don't see any OOM issue or high usage of RSS at the java application level but I could see MariaDB occupies huge memory.... apart from that overall application runs fine. To ensure this we observed memory utilization at the large-scale application servers where we deployed MariaDB and Java applications into different servers with 128GB each and could observe that the system with MariaDB uses full RAM size whereas the Java application consumption looks stable. *Server with Java application* sys> free total used free shared buff/cache available Mem: 128755 125171 606 118 2977 2694 Swap: 65535 9746 55789 *Server with MariaDB* sys> free total used free shared buff/cache available Mem: 128755 125171 606 118 2977 2694 Swap: 65535 9746 55789 *MariaDb conf:* innodb_buffer_pool_size=102G innodb_buffer_pool_instances=8 innodb_io_capacity=3000 innodb_log_buffer_size=8M innodb_log_file_size=2047MB innodb_log_files_in_group=2 Is this expected that MariaDB consumes overall memory and manages within RAM size.... regardless of buffer size configured? Do we need to optimize any configuration to control the MariaDB consumption apart from buffer pool size? Thanks and Regards, Ragul

3 3

[Maria-discuss] Troubleshooting Statement That Crashes MariaDB (mysqld got signal 11)
by Michael Caplan 04 Nov '21

04 Nov '21

Hello, I have a nothing fancy SQL statement that crashes MariaDB 10.3.31. The error log reports "mysqld got signal 11" with the below included mini core dump. The statement caused no issues with my prior install of 10.2.xx. With 10.3, the statement runs the majority of the time. The vast majority of the time. I have never dealt with such an issue with MariaDB, and am unsure where to start with troubleshooting this. Any tips? FYI, there is a corresponding bug recorded here: https://jira.mariadb.org/browse/MDEV-26835 Thanks, Mike 211014 16:13:53 [ERROR] mysqld got signal 11 ; This could be because you hit a bug. It is also possible that this binary or one of the libraries it was linked against is corrupt, improperly built, or misconfigured. This error can also be caused by malfunctioning hardware. To report this bug, see https://mariadb.com/kb/en/reporting-bugs We will try our best to scrape up some info that will hopefully help diagnose the problem, but since we have already crashed, something is definitely wrong and this may fail. Server version: 10.3.31-MariaDB-1:10.3.31+maria~focal-log key_buffer_size=134217728 read_buffer_size=131072 max_used_connections=587 max_threads=752 thread_count=81 It is possible that mysqld could use up to key_buffer_size + (read_buffer_size + sort_buffer_size)*max_threads = 1784362 K bytes of memory Hope that's ok; if not, decrease some variables in the equation. Thread pointer: 0x7f13e83b3768 Attempting backtrace. You can use the following information to find out where mysqld died. If you see no messages after this, something went terribly wrong... stack_bottom = 0x7f1644e96dd8 thread_stack 0x49000 /usr/sbin/mysqld(my_print_stacktrace+0x32)[0x55d5a2f6f3a2] /usr/sbin/mysqld(handle_fatal_signal+0x55d)[0x55d5a2a46e8d] /lib/x86_64-linux-gnu/libpthread.so.0(+0x153c0)[0x7f27569fc3c0] /usr/sbin/mysqld(+0x978f4f)[0x55d5a2bdbf4f] /usr/sbin/mysqld(_ZN7handler10ha_rnd_posEPhS0_+0x50)[0x55d5a2a4c6d0] /usr/sbin/mysqld(_Z16rr_from_pointersP11READ_RECORD+0x3e)[0x55d5a2b7cb2e] /usr/sbin/mysqld(_Z10sub_selectP4JOINP13st_join_tableb+0x18e)[0x55d5a28869be] /usr/sbin/mysqld(_ZN4JOIN10exec_innerEv+0xbee)[0x55d5a28b0dce] /usr/sbin/mysqld(_ZN4JOIN4execEv+0x37)[0x55d5a28b1137] /usr/sbin/mysqld(_Z12mysql_selectP3THDP10TABLE_LISTjR4ListI4ItemEPS4_jP8st_orderS9_S7_S9_yP13select_resultP18st_select_lex_unitP13st_select_lex+0xf2)[0x55d5a28b1282] /usr/sbin/mysqld(_Z13handle_selectP3THDP3LEXP13select_resultm+0x14b)[0x55d5a28b1bcb] /usr/sbin/mysqld(+0x5de9f1)[0x55d5a28419f1] /usr/sbin/mysqld(_Z21mysql_execute_commandP3THD+0x5414)[0x55d5a284f6e4] /usr/sbin/mysqld(_Z11mysql_parseP3THDPcjP12Parser_statebb+0x1f3)[0x55d5a2852253] /usr/sbin/mysqld(_Z16dispatch_command19enum_server_commandP3THDPcjbb+0x105d)[0x55d5a285462d] /usr/sbin/mysqld(_Z10do_commandP3THD+0x12d)[0x55d5a28565dd] /usr/sbin/mysqld(_Z24do_handle_one_connectionP7CONNECT+0x246)[0x55d5a293f186] /usr/sbin/mysqld(handle_one_connection+0x3f)[0x55d5a293f35f] /lib/x86_64-linux-gnu/libpthread.so.0(+0x9609)[0x7f27569f0609] /lib/x86_64-linux-gnu/libc.so.6(clone+0x43)[0x7f2756915293] Trying to get some variables. Some pointers may be invalid and cause the dump to abort. Query (0x7f13e84a8b70): SELECT * FROM (( # # get postings from last statement and postings not on statements # SELECT a.*, COALESCE(ci.id, i.id) as invoice_id FROM ddx_lab_801317.postings a LEFT JOIN ddx_lab_801317.cases_invoices ci ON a.ref_type = 'CASE_INVOICE' AND a.ref_id = ci.id LEFT JOIN ddx_lab_801317.invoices i ON a.ref_type = 'INVOICE' AND a.ref_id = i.id WHERE ( a.statement_date IS NULL OR a.statement_date = '2021-09-27' ) AND a.practice_id='19686' ORDER BY post_date ) UNION ( # # get postings that have remaining balances # SELECT DISTINCT a.*, COALESCE(ci.id, i.id) as invoice_id FROM ddx_lab_801317.postings a LEFT JOIN ddx_lab_801317.cases_invoices ci ON a.ref_type = 'CASE_INVOICE' AND a.ref_id = ci.id LEFT JOIN ddx_lab_801317.invoices i ON a.ref_type = 'INVOICE' AND a.ref_id = i.id LEFT JOIN ddx_lab_801317.postings b ON a.ref_type = b.ref_type AND a.ref_id = b.ref_id AND a.id != b.id WHERE NOT (a.type = 'CREDIT' OR a.type = 'PAYMENT') AND ((SELECT SUM(b.amount)) IS NULL OR (SELECT SUM(a.amount)) + (SELECT SUM(b.amount)) > 0) AND a.practice_id='19686' GROUP BY a.id ORDER BY post_date ) UNION ( # # get partial payments and credits from older statements # SELECT a.*, COALESCE(ci.id, i.id) as invoice_id FROM ddx_lab_801317.postings a LEFT JOIN ddx_lab_801317.cases_invoices ci ON a.ref_type = 'CASE_INVOICE' AND a.ref_id = ci.id LEFT JOIN ddx_lab_801317.invoices i ON a.ref_type = 'INVOICE' AND a.ref_id = i.id WHERE ( a.statement_date < '2021-09-27' ) AND (a.type='PAYMENT' OR a.type='CREDIT') AND a.practice_id='19686' )) AS t1 ORDER BY post_date ASC, invoice_id ASC Connection ID (thread ID): 21014956 Status: NOT_KILLED Optimizer switch: index_merge=on,index_merge_union=on,index_merge_sort_union=on,index_merge_intersection=on,index_merge_sort_intersection=off,engine_condition_pushdown=off,index_condition_pushdown=on,derived_merge=on,derived_with_keys=on,firstmatch=on,loosescan=on,materialization=on,in_to_exists=on,semijoin=on,partial_match_rowid_merge=on,partial_match_table_scan=on,subquery_cache=on,mrr=off,mrr_cost_based=off,mrr_sort_keys=off,outer_join_with_cache=on,semijoin_with_cache=on,join_cache_incremental=on,join_cache_hashed=on,join_cache_bka=on,optimize_join_buffer_size=off,table_elimination=on,extended_keys=on,exists_to_in=on,orderby_uses_equalities=on,condition_pushdown_for_derived=on,split_materialized=on The manual page at https://mariadb.com/kb/en/how-to-produce-a-full-stack-trace-for-mysqld/ contains information that should help you find out what is causing the crash. Writing a core file... Working directory at /var/lib/mysql Resource Limits: Limit Soft Limit Hard Limit Units Max cpu time unlimited unlimited seconds Max file size unlimited unlimited bytes Max data size unlimited unlimited bytes Max stack size 8388608 unlimited bytes Max core file size 0 unlimited bytes Max resident set unlimited unlimited bytes Max processes unlimited unlimited processes Max open files 1048576 1048576 files Max locked memory unlimited unlimited bytes Max address space unlimited unlimited bytes Max file locks unlimited unlimited locks Max pending signals 514974 514974 signals Max msgqueue size 819200 819200 bytes Max nice priority 0 0 Max realtime priority 0 0 Max realtime timeout unlimited unlimited us Core pattern: core

2 2

[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