Creating this retroactively.
We just got this affecting two hosts from s3 (db1175 and db1189):
[15:40:15] <+icinga-wm> PROBLEM - MariaDB Replica SQL: s3 #page on db1175 is CRITICAL: CRITICAL slave_sql_state Slave_SQL_Running: No, Errno: 1034, Errmsg: Error Index for table pagelinks is corrupt: try to repair it on query. Default database: mswiktionary. [Query snipped] https://wikitech.wikimedia.org/wiki/MariaDB/troubleshooting%23Depooling_a_replica [15:40:16] <+icinga-wm> PROBLEM - MariaDB Replica SQL: s3 #page on db1189 is CRITICAL: CRITICAL slave_sql_state Slave_SQL_Running: No, Errno: 1034, Errmsg: Error Index for table pagelinks is corrupt: try to repair it on query. Default database: mswiktionary. [Query snipped] https://wikitech.wikimedia.org/wiki/MariaDB/troubleshooting%23Depooling_a_replica
The issue was the same on both:
Last_SQL_Error: Error 'Index for table 'pagelinks' is corrupt; try to repair it' on query. Default database: 'mswiktionary'. Query: 'INSERT /* MediaWiki\Deferred\LinksUpdate\LinksTable::doWrites */ IGNORE INTO `pagelinks` (pl_from_namespace,pl_namespace,pl_title,pl_target_id,pl_from) VALUES (0,6,'Nl-bekijken.ogg',213370,50848),(0,0,'be-',29101,50848),(0,0,'kijken',213089,50848),(0,0,'lihat',14601,50848),(0,0,'tonton',16843,50848),(0,4,'Abjad_Fonetik_Antarabangsa',39952,50848),(0,100,'Sebutan_bahasa_Belanda',39978,50848),(0,102,'Bahasa_Belanda/ɛi̯kən',213088,50848)'
It was fixed by issuing the following on each host:
cumin2024@db1175.eqiad.wmnet[(none)]> set session sql_log_bin=0; Query OK, 0 rows affected (0.001 sec) cumin2024@db1175.eqiad.wmnet[mswiktionary]> alter table pagelinks engine=innodb,force; Query OK, 0 rows affected (4.077 sec) Records: 0 Duplicates: 0 Warnings: 0 cumin2024@db1175.eqiad.wmnet[mswiktionary]> stop slave; start slave;