Bug#858495: glusterfs-client: glusterfs fails to mount storage

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

Bug#858495: glusterfs-client: glusterfs fails to mount storage

Ingo Jürgensmann-10
Hi!

I experienced issues with mounting glusterfs storage after I upgraded some clients from Jessie to Stretch, although I’m not using Infiniband at all:

[2017-06-24 12:51:53.240389] I [MSGID: 100030] [glusterfsd.c:2454:main] 0-/usr/sbin/glusterfs: Started running /usr/sbin/glusterfs version 3.8.8 (args: /usr/sbin/glusterfs --read-only --fuse-mountopts=nodev,noexec --volfile-server=192.168.254.254 --volfile-id=/le --fuse-mountopts=nodev,noexec /etc/letsencrypt.sh/certs)
[2017-06-24 12:51:54.534826] E [mount.c:318:fuse_mount_sys] 0-glusterfs-fuse: ret = -1

[2017-06-24 12:51:54.534896] I [mount.c:365:gf_fuse_mount] 0-glusterfs-fuse: direct mount failed (Invalid argument) errno 22, retry to mount via fusermount
[2017-06-24 12:51:56.668254] I [MSGID: 101190] [event-epoll.c:628:event_dispatch_epoll_worker] 0-epoll: Started thread with index 1
[2017-06-24 12:51:56.671649] E [glusterfsd-mgmt.c:1590:mgmt_getspec_cbk] 0-glusterfs: failed to get the 'volume file' from server
[2017-06-24 12:51:56.671669] E [glusterfsd-mgmt.c:1690:mgmt_getspec_cbk] 0-mgmt: failed to fetch volume file (key:/le)
[2017-06-24 12:51:57.014502] W [glusterfsd.c:1327:cleanup_and_exit] (-->/usr/lib/x86_64-linux-gnu/libgfrpc.so.0(rpc_clnt_handle_reply+0x90) [0x7fbea36c4a20] -->/usr/sbin/glusterfs(mgmt_getspec_cbk+0x494) [0x55fbbaed06f4] -->/usr/sbin/glusterfs(cleanup_and_exit+0x54) [0x55fbbaeca444] ) 0-: received signum (0), shutting down
[2017-06-24 12:51:57.014564] I [fuse-bridge.c:5794:fini] 0-fuse: Unmounting '/etc/letsencrypt.sh/certs'.
[2017-06-24 16:44:45.501056] I [MSGID: 100030] [glusterfsd.c:2454:main] 0-/usr/sbin/glusterfs: Started running /usr/sbin/glusterfs version 3.8.8 (args: /usr/sbin/glusterfs --read-only --fuse-mountopts=nodev,noexec --volfile-server=192.168.254.254 --volfile-id=/le --fuse-mountopts=nodev,noexec /etc/letsencrypt.sh/certs)
[2017-06-24 16:44:45.504038] E [mount.c:318:fuse_mount_sys] 0-glusterfs-fuse: ret = -1

[2017-06-24 16:44:45.504084] I [mount.c:365:gf_fuse_mount] 0-glusterfs-fuse: direct mount failed (Invalid argument) errno 22, retry to mount via fusermount

I tried to find a solution using Google and such, but in the end I joined #gluster on Freenode and asked there.

The answer was:

> JoeJulian
> ij__: debian breaks apart ipv4 and ipv6. You'll need to remove the ipv6 ::1 address from localhost in /etc/hosts or recombine your ip stack (it's a sysctl thing)
> JoeJulian
> It has to do with the decisions made by the debian distro designers. All debian versions should have that problem. (yes, server side).

The tip with disabling ::1 in /etc/hosts worked for Jessie server and Stretch clients, but when the servers got upgraded to Stretch as well, this didn’t help anymore.

What actually worked for me, was the tip in http://lists.gluster.org/pipermail/gluster-users/2017-February/029938.html to use "gluster volume set <volname> transport.address-family inet“

I can’t comment on the statement that there’s something strange in Debian with IPv4 and IPv6 that needs to get „recombined“ in sysctl.conf. However, my findings do support that suspicion that the error is caused by IPv6 as it immediatedly worked when just using IPv4 in glusterfs.

Would be nice if this issue could be fixed in Stretch.


--
Ciao...          //        http://blog.windfluechter.net
      Ingo     \X/     XMPP: [hidden email]
       
gpg pubkey:  http://www.juergensmann.de/ij_public_key.asc