Testcases restart in InnoDB read only mode incorrectly | Intermittent errors in the error log on innodb.percona_log_archive_ro

[Laurynas Biveinis]

On 5.6 trunk, seen on ASan build:

innodb.percona_log_archive_ro w4 [ fail ] Found warnings/errors in server log file!
        Test ended at 2017-02-06 12:45:29
line
2017-02-06 20:45:29 26455 [Warning] InnoDB: Recovery skipped, --innodb-read-only set!
2017-02-06 20:45:29 26455 [ERROR] InnoDB: We scanned the log up to 0. A checkpoint was at 1632026 and the maximum LSN on a database page was 0. It is possible that the database is now corrupt!
2017-02-06 20:45:29 7fb1ac7c7780 InnoDB: Error: page 5 log sequence number 1632036
^ Found warnings in /mnt/workspace/percona-server-5.6-asan-param/BUILD_TYPE/release-asan/Host/ubuntu-xenial-64bit/build/mysql-test/var/4/log/mysqld.1.err
ok

[Laurynas Biveinis (laurynas-biveinis)]

InnoDB read only mode restart requires a preceding slow shutdown, which the current innodb.percona_log_archive_ro does not do. Another Percona testcase that does not do it is innodb.xtradb_compressed_columns_read_only

[Laurynas Biveinis (laurynas-biveinis)]

https://github.com/percona/percona-server/pull/1548
https://github.com/percona/percona-server/pull/1549

[Shahriyar Rzayev (rzayev-sehriyar)]

Percona now uses JIRA for bug reports so this bug report is migrated to: https://jira.percona.com/browse/PS-2234