lp:1055547 repeatable on XFS

[Raghavendra D Prabhu]

I am able to repeat this on XFS during prepare:

====================================
120629 6:12:19 InnoDB: Shutdown completed; log sequence number 3351590558
InnoDB: Error: tried to read 2048 bytes at offset 0 0.
InnoDB: Was only able to read 0.
120629 6:12:19 InnoDB: Operating system error number 22 in a file operation.
InnoDB: Error number 22 means 'Invalid argument'.
InnoDB: Some operating system error numbers are described at
InnoDB: http://dev.mysql.com/doc/refman/5.5/en/operating-system-error-codes.html
InnoDB: File operation call: 'read'.
InnoDB: Cannot continue operation.
innobackupex: Error:
==========================================

Does PXB testsuite run tests on XFS?

[Alexey Kopytov (akopytov)]

It's not clear why it doesn't fail before prepare, as exactly the same I/O requests are issued. I think the fix is to disable O_DIRECT when manipulating xtrabackup_logfile before and after recovery.

[Raghavendra D Prabhu (raghavendra-prabhu)]

The details of XFS filesystem where this happened (for those reproducing this):

xfs_info /dev/sda7
==========================

meta-data=/dev/sda7 isize=256 agcount=4, agsize=1387072 blks
         = sectsz=512 attr=2
data = bsize=4096 blocks=5548288, imaxpct=25
         = sunit=0 swidth=0 blks
naming =version 2 bsize=4096 ascii-ci=0
log =internal bsize=4096 blocks=2709, version=2
         = sectsz=512 sunit=0 blks, lazy-count=1
realtime =none extsz=4096 blocks=0, rtextents=0

Pretty much standard settings.

[Shahriyar Rzayev (rzayev-sehriyar)]

Percona now uses JIRA for bug reports so this bug report is migrated to: https://jira.percona.com/browse/PXB-640