How did you determine that the log file sync waits (oracle processes
waiting for LGWR) are due to LGWR writes being too slow? Did you try
changing the log files to different storage, and see if it made a
difference? Does the storage have a write back cache?
I doubt the cluster size has any impact. Log files are pre-allocated,
and do not extend at runtime.
Thanks,
Herbert.
On 11/21/11 9:17 PM, Pravin K Patil wrote:> Hi,
>
> We have ocfs2 1.4 and having high log file sync waits, which are
> mostly due to LGWR not able to complete writes fast enought for online
> redo logs. All of our online redo log file systems are created with
> default cluster size of 4K, If we get new file systems of cluster size
> 128 K or even of 1MB, will it help to speed up the writing to online
> redolog and reduce the i/o particular to this and help reduce log file
> sync waits? Indirectly we are looking to tune the heavy writes on the
> file system. which all properties of ocfs2 would help speed this up?
>
> Appreciate the response from all of your's experience..
>
> Regards,
> Pravin
>
>
> _______________________________________________
> Ocfs2-users mailing list
> Ocfs2-users at oss.oracle.com
> http://oss.oracle.com/mailman/listinfo/ocfs2-users
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://oss.oracle.com/pipermail/ocfs2-users/attachments/20111122/d55b0d51/attachment.html