Description:
sporadic pushbuild failure on windows in backup_backupdir:
main.backup_backupdir [ fail ]
CURRENT_TEST: main.backup_backupdir
mysqltest: At line 86: query 'BACKUP DATABASE bup_backupdir TO '../../../../../../../../../../../../../../../../../../bup_backupdir5.bak'' succeeded - should have failed with errno 1632...
The result from queries just before the failure was:
< snip >
backup_id
#
Ensure backup image file went to the correct location
Perform restore
RESTORE FROM '../bup_backupdir2.bak';
backup_id
#
Perform backup
BACKUP DATABASE bup_backupdir TO '../../bup_backupdir3.bak';
backup_id
#
Ensure backup image file went to the correct location
Reset backupdir with ending /
SET @@global.backupdir = '../../tmp/backup/';
Perform backup
BACKUP DATABASE bup_backupdir TO '../../bup_backupdir4.bak';
backup_id
#
Ensure backup image file went to the correct location
Try a backup to an invalid relative path.
More results from queries before failure can be found in e:\var-n_mix-100\2\log\backup_backupdir.log
- saving 'e:/var-n_mix-100/2/log/main.backup_backupdir/' to 'e:/var-n_mix-100/log/main.backup_backupdir/'
Retrying test, attempt(2/3)...
cygpath -m C:/cygwin/home/pushbuild/pb1/pb/bzr_mysql-6.0-rpl/79/mysql-6.0.8-alpha-pb79/sql/release/mysqld.exe
cygpath -m e:/var-n_mix-100/2/mysqld.1/mysqld.err
cygpath -m e:/var-n_mix-100/2/mysqld.1/mysqld.err
cygpath -m C:/cygwin/home/pushbuild/pb1/pb/bzr_mysql-6.0-rpl/79/mysql-6.0.8-alpha-pb79/client/release//mysqltest.exe
cygpath -m e:/var-n_mix-100/2/tmp/check-mysqld_1.err
cygpath -m C:/cygwin/home/pushbuild/pb1/pb/bzr_mysql-6.0-rpl/79/mysql-6.0.8-alpha-pb79/client/release//mysqltest.exe
cygpath -m e:/var-n_mix-100/2/log/current_test
How to repeat:
Has failed three times, the same push. Since it failed on different hosts (in fact, failed on all three windows hosts in this push), that seems to exclude host problems. The next push only enabled a previously disabled test case, which is run *after* backup_backupdir, and the test has run successfully after that. That seems to exclude something wrong with the push. So I can't think of any good explanation for why the failures are so closely located to each other, except witchcraft or chance.
https://intranet.mysql.com/secure/pushbuild/showpush.pl?dir=bzr_mysql-6.0-rpl&order=79 vm-win2003-32-a/n_mix
xref: http://tinyurl.com/6s9atd