[MDEV-15274] innodb_gis.rtree_recovery failed in buildbot with error on check table Created: 2018-02-11  Updated: 2023-12-11  Resolved: 2023-12-11

Status: Closed
Project: MariaDB Server
Component/s: GIS, Storage Engine - InnoDB, Tests
Affects Version/s: 10.2, 10.3, 10.4, 10.5
Fix Version/s: 10.5.2, 10.6.0

Type: Bug Priority: Major
Reporter: Elena Stepanova Assignee: Marko Mäkelä
Resolution: Won't Fix Votes: 1
Labels: None

Issue Links:
Relates
relates to MDEV-12353 Efficient InnoDB redo log record format Closed

 Description   

http://buildbot.askmonty.org/buildbot/builders/kvm-fulltest2/builds/11336/steps/mtr_nm/logs/stdio

innodb_gis.rtree_recovery 'innodb'       w1 [ fail ]
        Test ended at 2018-01-25 15:21:01
 
CURRENT_TEST: innodb_gis.rtree_recovery
--- /mnt/buildbot/build/mariadb-10.2.13/mysql-test/suite/innodb_gis/r/rtree_recovery.result	2018-01-25 11:38:02.000000000 +0200
+++ /mnt/buildbot/build/mariadb-10.2.13/mysql-test/suite/innodb_gis/r/rtree_recovery.reject	2018-01-25 15:21:00.000000000 +0200
@@ -19,7 +19,8 @@
 COMMIT;
 check table t1;
 Table	Op	Msg_type	Msg_text
-test.t1	check	status	OK
+test.t1	check	Warning	InnoDB: Index 'c2' contains 551 entries, should be 367.
+test.t1	check	error	Corrupt
 drop table t1;
 create table t1 (c1 int, c2 point not null, spatial index (c2))engine=innodb;
 CALL insert_t1(367);
 
mysqltest: Result length mismatch



 Comments   
Comment by Alice Sherepa [ 2020-09-08 ]

according to cross reference - test currently fails on 10.4 only (the last fail on 10.3 was 20.02.2019)
http://buildbot.askmonty.org/buildbot/builders/kvm-bintar-quantal-x86/builds/24256/steps/mtr/logs/stdio

innodb_gis.rtree_recovery 'innodb'       w3 [ fail ]
        Test ended at 2020-09-04 14:51:55
 
CURRENT_TEST: innodb_gis.rtree_recovery
--- /usr/local/mariadb-10.4.15-linux-i686/mysql-test/suite/innodb_gis/r/rtree_recovery.result	2020-09-04 12:44:03.000000000 +0300
+++ /usr/local/mariadb-10.4.15-linux-i686/mysql-test/suite/innodb_gis/r/rtree_recovery.reject	2020-09-04 14:51:55.111824835 +0300
@@ -21,7 +21,8 @@
 # restart
 check table t1;
 Table	Op	Msg_type	Msg_text
-test.t1	check	status	OK
+test.t1	check	Warning	InnoDB: Index 'c2' contains 551 entries, should be 367.
+test.t1	check	error	Corrupt
 drop table t1;
 create table t1 (c1 int, c2 point not null, spatial index (c2))engine=innodb;
 CALL insert_t1(367);
 
mysqltest: Result length mismatch
 
 - saving '/usr/local/mariadb-10.4.15-linux-i686/mysql-test/var/3/log/innodb_gis.rtree_recovery-innodb/' to '/usr/local/mariadb-10.4.15-linux-i686/mysql-test/var/log/innodb_gis.rtree_recovery-innodb/'
 
Retrying test innodb_gis.rtree_recovery, attempt(2/3)...
 
worker[3] > Restart  - not started
***Warnings generated in error logs during shutdown after running tests: innodb_gis.rtree_recovery
 
2020-09-04 14:51:52 8 [ERROR] InnoDB: Flagged corruption of `c2` in table `test`.`t1` in CHECK TABLE; Wrong count

Comment by Alice Sherepa [ 2021-06-24 ]

http://buildbot.askmonty.org/buildbot/builders/kvm-bintar-quantal-x86/builds/27170/steps/mtr/logs/stdio

Comment by Marko Mäkelä [ 2023-12-11 ]

I remember noticing and fixing some likely bugs in this area while working on MDEV-12353 (MariaDB Server 10.5).

In the cross-reference, there are plenty of failures of this test in 10.4 but none in 10.5 or later. Hence, I would think that this has been fixed in 10.5. This is not easily fixable in 10.4 without changing its old redo log format.

Generated at Thu Feb 08 08:20:02 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.