OS Upgrade Considerations For XFS Volume

OS Upgrade Considerations For XFS Volume

[Bharathiraja P]

[-- Attachment #1.1: Type: text/plain, Size: 1561 bytes --]

Hi there,

We recently upgraded our database system from OpenSuSE-11.4 to OpenSuSE-13.1

Post the upgrade we are noticing load spikes on our database system.

OpenSuSE 11.4 uses kernel-2.6.37 and OpenSuSE-13.1 uses kernel-3.11.10

There seems to be some changes related to XFS in new kernel and I'm getting
this error when I run xfs_logprint

==============
sudo /usr/sbin/xfs_logprint /dev/mapper/pgsql-data
raja's password:
xfs_logprint:
xfs_logprint: /dev/mapper/pgsql-data contains a mounted and writable
filesystem
    data device: 0xfd00
    log device: 0xfd00 daddr: 6442450880 length: 4173824

cycle: 879    version: 2        lsn: 879,3707616    tail_lsn: 879,3666528
length of Log Record: 32256    prev offset: 3707552        num ops: 120
uuid: 314e5417-3787-476d-9ee3-22a81732cb67   format: little endian linux
h_size: 32768
----------------------------------------------------------------------------
Oper (0): tid: f8163d32  len: 112  clientid: TRANS  flags: WAS_CONT END
Left over region from split log item
----------------------------------------------------------------------------
Oper (1): tid: f8163d32  len: 1664  clientid: TRANS  flags: none
xfs_logprint: unknown log operation type (b001)
**********************************************************************
* ERROR: data block=3707616                                           *
**********************************************************************
Bad data in log
==============


Do we need to do something for the xfs volume post os upgrade for better
performance?

--
Bharathiraja

[-- Attachment #1.2: Type: text/html, Size: 1854 bytes --]

[-- Attachment #2: Type: text/plain, Size: 121 bytes --]

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: OS Upgrade Considerations For XFS Volume

[Dave Chinner]

On Thu, Nov 27, 2014 at 07:59:59PM +0530, Bharathiraja P wrote:
> Hi there,
> 
> We recently upgraded our database system from OpenSuSE-11.4 to OpenSuSE-13.1
> 
> Post the upgrade we are noticing load spikes on our database system.

And ...?

Is there any actual performance problem you've measured?

> OpenSuSE 11.4 uses kernel-2.6.37 and OpenSuSE-13.1 uses kernel-3.11.10
> 
> There seems to be some changes related to XFS in new kernel and I'm getting
> this error when I run xfs_logprint
> 
> ==============
> sudo /usr/sbin/xfs_logprint /dev/mapper/pgsql-data
> raja's password:
> xfs_logprint:
> xfs_logprint: /dev/mapper/pgsql-data contains a mounted and writable

logprint is completely unreliable on an active and mounted
filesystem. It's not going to tell you anything useful about any
sort of performance issue - logprint is for debugging log recovery
problems....

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs

Re: OS Upgrade Considerations For XFS Volume

[Stan Hoeppner]

On 11/27/2014 08:29 AM, Bharathiraja P wrote:
> Hi there,
> 
> We recently upgraded our database system from OpenSuSE-11.4 to OpenSuSE-13.1
> 
> Post the upgrade we are noticing load spikes on our database system.
> 
> OpenSuSE 11.4 uses kernel-2.6.37 and OpenSuSE-13.1 uses kernel-3.11.10
> 
> There seems to be some changes related to XFS in new kernel and I'm
> getting this error when I run xfs_logprint

The two major XFS defaults changes here are:

1.  Delayed logging
2.  inode64 allocator

Read about each.  They should not be impacting performance of a DB
workload nor system load.

Stan

_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs