Hi,
The issue Niels is mentioning was present in 3.8.0, which I believe is
the version you have installed. It was fixed in 3.8.1. Also, he is right
about the fact that the issue made mounting completely non-functional,
and there was no delay. To know the issue further could you please
answer the following questions about your setup.
1. You have two servers, one running 3.7.11, and the other 3.8.0 right?
Are you mounting the client on one of these servers, if yes then which
one and whch server's IP are you using in the mount point.
2. Also can you describe your volume's configuration. Is it a replicate
or dist-replicate volume. This will help us confirm that, the client
translator is getting the child-up from the node running 3.8.0, and
hence is operational after some delay.
3. Can you also confirm, if after the delay, you are able to see data
being written to the server running 3.7.11.
Regards,
Avra
On 10/21/2016 04:13 PM, Niels de Vos wrote:
On Fri, Oct 21, 2016 at 12:14:15PM +0200, Josep Manel Andrés wrote:
Hi again,
I have two servers SLES 12 SP 0 with
Information for package glusterfs:
----------------------------------
Repository: GlusterFS
Name: glusterfs
Version: 3.7.11-101.1
Arch: x86_64
Vendor: obs://build.opensuse.org/home:kkeithleatredhat
and now I decided to upgrade one server to SP1 and install
Information for package glusterfs:
----------------------------------
Repository: GlusterFS-3.8 (SLE_12_SP1)
Name: glusterfs
Version: 3.8.0-100.1
Arch: x86_64
Vendor: obs://build.opensuse.org/home:kkeithleatredhat
but what happens now is that when trying to mount a volume with 3.8 it takes
so long...over 6 or 7 seconds, and here is the moment where gets stacked in
the logs:
[2016-10-21 10:05:18.285837] I [MSGID: 108005]
[afr-common.c:4137:afr_notify] 0-volume1-replicate-0: Subvolume
'volume1-client-1' came back up; going online.
[2016-10-21 10:05:18.285872] I [MSGID: 114035]
[client-handshake.c:201:client_set_lk_version_cbk] 0-volume1-client-1:
Server lk version = 1
[2016-10-21 10:05:18.286023] W [MSGID: 114007]
[client-handshake.c:1176:client_setvolume_cbk] 0-volume1-client-0: failed to
find key 'child_up' in the options
[2016-10-21 10:05:29.273913] I [fuse-bridge.c:5241:fuse_graph_setup] 0-fuse:
switched to graph 0
[2016-10-21 10:05:29.274151] I [fuse-bridge.c:4153:fuse_init]
0-glusterfs-fuse: FUSE inited with protocol versions: glusterfs 7.24 kernel
7.22
[2016-10-21 10:05:29.275426] I [MSGID: 108031]
[afr-common.c:1908:afr_local_discovery_cbk] 0-volume1-replicate-0: selecting
local read_child volume1-client-1
Any idea about what happens? Should I upgrade de other server?
This looks related to https://bugzilla.redhat.com/1350326 . It has been
addressed in glusterfs-3.8.0, so I do not know if the delay in mounting
is expected. Without the fix mounting seemed to have been completely
non-functional. Avra should be able to explain the details a little
more.
Niels
_______________________________________________
Gluster-users mailing list
Gluster-users@xxxxxxxxxxx
http://www.gluster.org/mailman/listinfo/gluster-users
