[Gluster-users] gluster client mount fails on fedora 23

Thu Dec 24 08:12:25 UTC 2015

so what do I need to do to use cluster clients greater than
glusterfs:3.7.2:3.fc23:x86_64:(none):(none)
every version I have built on fedora 23 greater than this has mailed to
mount gluster volume.

Are there any best practices, for upgrading the server side, or do just
start from scratch, and copy tera bytes, peda bytes of data :-/ to the
new improved gluster server ?

gluster server
node 1 fedora 22
# rpmq ^gluster
glusterfs:3.6.7:1.fc22:x86_64:Fedora Project:Fedora Project
glusterfs-api:3.6.7:1.fc22:x86_64:Fedora Project:Fedora Project
glusterfs-cli:3.6.7:1.fc22:x86_64:Fedora Project:Fedora Project
glusterfs-fuse:3.6.7:1.fc22:x86_64:Fedora Project:Fedora Project
glusterfs-libs:3.6.7:1.fc22:x86_64:Fedora Project:Fedora Project
glusterfs-server:3.6.7:1.fc22:x86_64:Fedora Project:Fedora Project

nodes 2, 3, 4 centos 7.2
# rpmq ^gluster
glusterfs:3.6.6:1.el7.centos:x86_64:(none):(none)
glusterfs-api:3.6.6:1.el7.centos:x86_64:(none):(none)
glusterfs-cli:3.6.6:1.el7.centos:x86_64:(none):(none)
glusterfs-fuse:3.6.6:1.el7.centos:x86_64:(none):(none)
glusterfs-libs:3.6.6:1.el7.centos:x86_64:(none):(none)
glusterfs-server:3.6.6:1.el7.centos:x86_64:(none):(none)

working clients
all servers mount at least one volume from the server group

centos 6.7
# rpmq ^gluster
glusterfs:3.6.0.54:1.el6:x86_64:(none):CentOS BuildSystem <
http://bugs.centos.org>
glusterfs-api:3.6.0.54:1.el6:x86_64:(none):CentOS BuildSystem <
http://bugs.centos.org>
glusterfs-cli:3.6.0.54:1.el6:x86_64:(none):CentOS BuildSystem <
http://bugs.centos.org>
glusterfs-fuse:3.6.0.54:1.el6:x86_64:(none):CentOS BuildSystem <
http://bugs.centos.org>
glusterfs-libs:3.6.0.54:1.el6:x86_64:(none):CentOS BuildSystem <
http://bugs.centos.org>
glusterfs-rdma:3.6.0.54:1.el6:x86_64:(none):CentOS BuildSystem <
http://bugs.centos.org>

centos 7.2
# rpmq ^gluster
glusterfs:3.7.1:16.el7:x86_64:(none):CentOS BuildSystem <
http://bugs.centos.org>
glusterfs-api:3.7.1:16.el7:x86_64:(none):CentOS BuildSystem <
http://bugs.centos.org>
glusterfs-cli:3.7.1:16.el7:x86_64:(none):CentOS BuildSystem <
http://bugs.centos.org>
glusterfs-client-xlators:3.7.1:16.el7:x86_64:(none):CentOS BuildSystem
<http://bugs.centos.org>
glusterfs-fuse:3.7.1:16.el7:x86_64:(none):CentOS BuildSystem <
http://bugs.centos.org>
glusterfs-libs:3.7.1:16.el7:x86_64:(none):CentOS BuildSystem <
http://bugs.centos.org>

fedora 23
# rpmq ^gluster
glusterfs:3.7.2:3.fc23:x86_64:(none):(none)
glusterfs-api:3.7.2:3.fc23:x86_64:(none):(none)
glusterfs-client-xlators:3.7.2:3.fc23:x86_64:(none):(none)
glusterfs-fuse:3.7.2:3.fc23:x86_64:(none):(none)
glusterfs-libs:3.7.2:3.fc23:x86_64:(none):(none)
On Sun, 2015-12-06 at 03:06 -0600, Arno A. Karner wrote:
> the client side
> 
> /var/log/messages
> Dec  6 03:02:19 cli.dom.com: gvs.mount: Unit entered failed state.
> 
> # mount -a
> Mount failed. Please check the log file for more details.
> 
> # grep /gvs /etc/fstab
> srv.dom.com:/gvs /gvs glusterfs defaults 0 0
> 
> # tail /var/log/glusterfs/gvs.log
> [2015-12-06 08:40:34.245021] I [MSGID: 100030]
> [glusterfsd.c:2318:main]
> 0-/usr/sbin/glusterfs: Started running /usr/sbin/glusterfs version
> 3.7.6 (args: /usr/sbin/glusterfs --volfile-server=srv.dom.com -
> -volfile
> -id=/gvs /gvs)
> [2015-12-06 08:40:34.264452] I [MSGID: 101190] [event
> -epoll.c:632:event_dispatch_epoll_worker] 0-epoll: Started thread
> with
> index 1
> [2015-12-06 08:40:34.264919] W [socket.c:588:__socket_rwv] 0
> -glusterfs:
> readv on 10.255.255.4:24007 failed (No data available)
> [2015-12-06 08:40:34.265376] E [rpc-clnt.c:362:saved_frames_unwind] (
> -
> -> /lib64/libglusterfs.so.0(_gf_log_callingfn+0x1a3)[0x7f481a752183]
> (-
> -> /lib64/libgfrpc.so.0(saved_frames_unwind+0x1cf)[0x7f481a51d41f] (-
> ->
> /lib64/libgfrpc.so.0(saved_frames_destroy+0xe)[0x7f481a51d53e] (-->
> /lib64/libgfrpc.so.0(rpc_clnt_connection_cleanup+0x7e)[0x7f481a51ed0e
> ]
> (--> /lib64/libgfrpc.so.0(rpc_clnt_notify+0x88)[0x7f481a51f528] )))))
> 0
> -glusterfs: forced unwinding frame type(GlusterFS Handshake)
> op(GETSPEC(2)) called at 2015-12-06 08:40:34.264590 (xid=0x1)
> [2015-12-06 08:40:34.265399] E [glusterfsd
> -mgmt.c:1603:mgmt_getspec_cbk] 0-mgmt: failed to fetch volume file
> (key:/gvs)
> [2015-12-06 08:40:34.265427] W [glusterfsd.c:1236:cleanup_and_exit] (
> -
> ->/lib64/libgfrpc.so.0(saved_frames_unwind+0x1fa) [0x7f481a51d44a] -
> ->/usr/sbin/glusterfs(mgmt_getspec_cbk+0x498) [0x55e8a25b8068] -
> ->/usr/sbin/glusterfs(cleanup_and_exit+0x56) [0x55e8a25b2416] ) 0-:
> received signum (0), shutting down
> [2015-12-06 08:40:34.265446] I [fuse-bridge.c:5683:fini] 0-fuse:
> Unmounting '/gvs'.
> 
> On Sat, 2015-12-05 at 12:53 +0530, Atin Mukherjee wrote:
> > 
> > On 12/05/2015 06:10 AM, Arno A. Karner wrote:
> > > is there a trick when clients, and servers are not on the same
> > > version
> > > of glusterfs. I have clients and servers on gluserfs 3.6.6-1 both
> > > centos7, and fedora 22, and work fine. Now that fedora 23 has
> > > been
> > > released I'm trying to use it. The client I am have trouble with
> > > is
> > > on
> > > fedora 23, which comes with glusterfs 3.7.6-1. On nfs I can use
> > > options
> > > to control version protocol uses, does this exist for glusterfs
> > > as
> > > well.
> > mount log & glusterd log please, are you seeing UN-previleged port
> > requests entries in glusterd log file?
> > 
> > ~Atin
> > > 
> > > Thanks in advance, Arno.
> > > _______________________________________________
> > > Gluster-users mailing list
> > > Gluster-users at gluster.org
> > > http://www.gluster.org/mailman/listinfo/gluster-users
> > > 