[Gluster-devel] Operations will block untill timeout is happened,when a node's I/O is very high.

Vijay Bellur vbellur at redhat.com
Mon Apr 15 07:55:49 UTC 2013


On 04/12/2013 12:06 PM, 高佳乐 wrote:
> Hi Everyone:
>
>      I’m using glusterfs-3.4.0alpha.There has always been a high
> iowait node in the cluster.Some
> opearations such as lookup,mkdir,rename...will block until timeout (1800s) is happened.
>
>      Is there a solution to this or an alternative that you can offer?
>
>

Can you please provide more details on the volume type and the 
workload/nature of operations being done on the mount point?

Also, looking into the brick log files might provide you with some 
information on what is happening.

-Vijay





More information about the Gluster-devel mailing list