[Gluster-users] CMA: unable to get RDMA device list

[Alan Hartless]

Hi Anoop,

So I think I found the issue. Looks like the server's IP changed. What's
the best way to recover from an IP change?

Thanks!
Alan

On Tue, Jun 28, 2016 at 2:41 AM Anoop C S <anoopcs at redhat.com> wrote:

> On Tue, 2016-06-28 at 06:01 +0000, Alan Hartless wrote:
> > Hi Anoop,
> >
> > 1. I didn't configure anything specific with RDMA so I assume TCP.
> > When I run gluster volume status all on the working pod, it has 0 for
> > the RDMA Port. The server that won't start Gluster is a Docker image
> > running Ubuntu 14.04 with Gluster 3.7 installed from Gluster's apt
> > repositories.
> >
>
> Fine. So we can take out the involvement of RDMA from the picture.
>
> > 2. There are no error messages for the bricks. No brick logs at all
> > on the server since it won't even start Glusterd due to the RDMA
> > error. On the working server, there are no errors recorded for the
> > bricks.
> >
>
> Can you also please check errors inside glusterd log file? If you can't
> find any error messages there, try increasing the log level to DEBUG by
> adding "-L DEBUG" option while starting glusterd and check for any
> abnormal messages?
>
> > Thanks,
> > Alan
> >
> > On Mon, Jun 27, 2016 at 5:52 AM Anoop C S <anoopcs at redhat.com> wrote:
> > > On Mon, 2016-06-27 at 15:20 +0530, Anoop C S wrote:
> > > > On Sun, 2016-06-26 at 14:48 +0000, Alan Hartless wrote:
> > > > >
> > > > > I had a glusterd server restart and now glusterfsd refused to
> > > > > start.
> > > > > When I try to start glusterd with
> > > > > /usr/sbin/glusterd -N -p /var/run/glusterd.pid
> > > > >
> > > > > I get this:
> > > > > librdmacm: couldn't read ABI version.
> > > > > librdmacm: assuming: 4
> > > > > CMA: unable to get RDMA device list
> > > > >
> > > > Some queries:
> > > > 1. With what transport type have you created the volume? On which
> > > > platform are your glusterfs servers running and in case you are
> > > using
> > > > rdma transport, what packages were installed for setting up the
> > > IB
> > > > stack?
> > > >
> > > > 2. Can you please share the error messages from brick logs(under
> > > > /var/log/gluster/bricks/) and glusterd logs(under
> > >
> > > Small correction. Brick logs are located at
> > > /var/log/glusterfs/bricks.
> > >
> > > > /var/log/glusterfs/)
> > > > for better clarity?
> > > >
> > > > >
> > > > > A google search has been unsuccessful on how to resolve this.
> > > Any
> > > > > ideas?
> > > > >
> > > > > Thanks!
> > > > > Alan
> > > > > _______________________________________________
> > > > > Gluster-users mailing list
> > > > > Gluster-users at gluster.org
> > > > > http://www.gluster.org/mailman/listinfo/gluster-users
> > > > _______________________________________________
> > > > Gluster-users mailing list
> > > > Gluster-users at gluster.org
> > > > http://www.gluster.org/mailman/listinfo/gluster-users
> > >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.gluster.org/pipermail/gluster-users/attachments/20160628/666f90f5/attachment.html>