This happens when a portmap request and RPC reconfigure fails on a
client. The client should connect to glusterd to get the brick port,
and then reconnect to the brick using the port. But this fails (for
some reason), leaving the client connected to glusterd instead of the
brick. When the client no
Have you checked the brick logs to see if there's anything unusual there?
Regards,
Vijay
On Thu, Sep 15, 2016 at 5:24 PM, Danny Lee wrote:
> Hi,
>
> Environment:
> Gluster Version: 3.8.3
> Operating System: CentOS Linux 7 (Core)
> Kernel: Linux 3.10.0-327.28.3.el7.x86_64
> Architecture: x86-64
>
On Friday 16 September 2016, Danny Lee wrote:
> Hi,
>
> Environment:
> Gluster Version: 3.8.3
> Operating System: CentOS Linux 7 (Core)
> Kernel: Linux 3.10.0-327.28.3.el7.x86_64
> Architecture: x86-64
> Replicated 3-Node Volume
> ~400GB of around a million files
>
> Description of Problem:
> One
Hi,
Environment:
Gluster Version: 3.8.3
Operating System: CentOS Linux 7 (Core)
Kernel: Linux 3.10.0-327.28.3.el7.x86_64
Architecture: x86-64
Replicated 3-Node Volume
~400GB of around a million files
Description of Problem:
One of the brick dies. The only suspect log I see is in the
etc-glusterf