[Gluster-users] does your samba work with 4.1.x (centos 7.5)
Anoop C S
anoopcs at cryptolab.net
Thu Nov 15 13:03:56 UTC 2018
On Wed, 2018-11-14 at 22:19 -0500, Diego Remolina wrote:
> Hi,
>
> Please download the logs from:
>
> https://www.dropbox.com/s/4k0zvmn4izhjtg7/samba-logs.tar.bz2?dl=0
[2018/11/14 22:01:31.974084, 10, pid=7577, effective(1009, 513), real(1009, 0)]
../source3/smbd/smb2_server.c:2279(smbd_smb2_request_dis
patch)
smbd_smb2_request_dispatch: opcode[SMB2_OP_FLUSH] mid = 9303
[2018/11/14 22:01:31.974123, 4, pid=7577, effective(1009, 513), real(1009, 0)]
../source3/smbd/uid.c:384(change_to_user)
Skipping user change - already user
[2018/11/14 22:01:31.974163, 10, pid=7577, effective(1009, 513), real(1009, 0)]
../source3/smbd/smb2_flush.c:134(smbd_smb2_flush_send)
smbd_smb2_flush: Test-Project_backup/preview.1957.dat - fnum 2399596398
I see the above flush request(SMB2_OP_FLUSH) without a response being logged compared to other
request-response pairs for SMB2_OP_QUERY_DIRECTORY, SMB2_OP_IOCTL, SMB2_OP_WRITE, SMB2_OP_READ which
brings me to the following bug:
https://bugzilla.samba.org/show_bug.cgi?id=13297
Is it also possible for you to collect network traces using wireshark from Windows client to samba
server along with corresponding samba logs?
> These options had to be set in the [global] section:
> kernel change notify = no
> kernel oplocks = no
Well, 'kernel oplocks' and 'posix locking' are listed (S) in man smb.conf(5) which can also be used
in [global] section meaning it affects all services. Anyway this should be fine.
> I also set log level = 10
Good one.
> I renamed the file as Test-Project.rvt for simplicity I opened revit
> and from revit I opened this file. I then started attempting to save
> it as a central at around 22:01:30s The file save then got stuck and
> at around 22:05 it finally failed saying that two dat files did not
> exist.
Do you experience problems with other applications and/or with other file types? Is this specific to
Revit?
> Diego
> On Tue, Nov 13, 2018 at 8:46 AM Anoop C S <anoopcs at autistici.org> wrote:
> >
> > On Tue, 2018-11-13 at 07:50 -0500, Diego Remolina wrote:
> > > >
> > > > Thanks for explaining the issue.
> > > >
> > > > I understand that you are experiencing hang while doing some operations on files/directories
> > > > in
> > > > a
> > > > GlusterFS volume share from a Windows client. For simplicity can you attach the output of
> > > > following
> > > > command:
> > > >
> > > > # gluster volume info <volume>
> > > > # testparm -s --section-name global
> > >
> > > gluster v status export
> > > Status of volume: export
> > > Gluster process TCP Port RDMA Port Online Pid
> > > ------------------------------------------------------------------------------
> > > Brick 10.0.1.7:/bricks/hdds/brick 49153 0 Y 2540
> > > Brick 10.0.1.6:/bricks/hdds/brick 49153 0 Y 2800
> > > Self-heal Daemon on localhost N/A N/A Y 2912
> > > Self-heal Daemon on 10.0.1.6 N/A N/A Y 3107
> > > Self-heal Daemon on 10.0.1.5 N/A N/A Y 5877
> > >
> > > Task Status of Volume export
> > > ------------------------------------------------------------------------------
> > > There are no active volume tasks
> > >
> > > # gluster volume info export
> > >
> > > Volume Name: export
> > > Type: Replicate
> > > Volume ID: b4353b3f-6ef6-4813-819a-8e85e5a95cff
> > > Status: Started
> > > Snapshot Count: 0
> > > Number of Bricks: 1 x 2 = 2
> > > Transport-type: tcp
> > > Bricks:
> > > Brick1: 10.0.1.7:/bricks/hdds/brick
> > > Brick2: 10.0.1.6:/bricks/hdds/brick
> > > Options Reconfigured:
> > > diagnostics.brick-log-level: INFO
> > > diagnostics.client-log-level: INFO
> > > performance.cache-max-file-size: 256MB
> > > client.event-threads: 5
> > > server.event-threads: 5
> > > cluster.readdir-optimize: on
> > > cluster.lookup-optimize: on
> > > performance.io-cache: on
> > > performance.io-thread-count: 64
> > > nfs.disable: on
> > > cluster.server-quorum-type: server
> > > performance.cache-size: 10GB
> > > server.allow-insecure: on
> > > transport.address-family: inet
> > > performance.cache-samba-metadata: on
> > > features.cache-invalidation-timeout: 600
> > > performance.md-cache-timeout: 600
> > > features.cache-invalidation: on
> > > performance.cache-invalidation: on
> > > network.inode-lru-limit: 65536
> > > performance.cache-min-file-size: 0
> > > performance.stat-prefetch: on
> > > cluster.server-quorum-ratio: 51%
> > >
> > > I had sent you the full smb.conf, so no need to run testparm -s
> > > --section-name global, please reference:
> > > http://termbin.com/y4j0
> >
> > Fine.
> >
> > > >
> > > > > This is the test samba share exported using vfs object = glusterfs:
> > > > >
> > > > > [vfsgluster]
> > > > > path = /vfsgluster
> > > > > browseable = yes
> > > > > create mask = 660
> > > > > directory mask = 770
> > > > > write list = @Staff
> > > > > kernel share modes = No
> > > > > vfs objects = glusterfs
> > > > > glusterfs:loglevel = 7
> > > > > glusterfs:logfile = /var/log/samba/glusterfs-vfsgluster.log
> > > > > glusterfs:volume = export
> > > >
> > > > Since you have mentioned path as /vfsgluster I hope you are sharing a subdirectory under
> > > > root of
> > > > the
> > > > volume.
> > >
> > > Yes, vfsgluster is a directory at the root of the export volume.
> >
> > Thanks for the confirmation.
> >
> > > It is also currently mounted in /export so that the rest of the files can be
> > > exported via samba with fuse mounts:
> > >
> > > # mount | grep export
> > > 10.0.1.7:/export on /export type fuse.glusterfs
> > > (rw,relatime,user_id=0,group_id=0,allow_other,max_read=131072)
> > >
> > > # ls -ld /export/vfsgluster
> > > drwxrws---. 4 dijuremo Staff 4096 Nov 12 20:24 /export/vfsgluster
> > >
> > > >
> > > > > Full smb.conf
> > > > > http://termbin.com/y4j0
> > > >
> > > > I see the "clustering" parameter set to 'yes'. How many nodes are there in the cluster? Out
> > > > of
> > > > those
> > > > how many are running as samba and/or gluster nodes?
> > > >
> > >
> > > There are a total of 3 gluster peers, but only two have bricks. The
> > > other is just present, but not even configured as an arbiter. Two of
> > > the nodes with bricks run ctdb and samba.
> >
> > OK. So basically a two node Samba-CTDB cluster.
> >
> > > > > /var/log/samba/glusterfs-vfsgluster.log
> > > > > http://termbin.com/5hdr
> > > > >
> > > > > Please let me know if there is any other information I can provide.
> > > >
> > > > Are there any errors in /var/log/samba/log.<IP/hostname>? IP/hostname = Windows client
> > > > machine
> > > >
> > >
> > > I do not currently have the log file directive enabled in smb.conf, I
> > > would have to enable it. Do you need me to repeat the process with it?
> >
> > Yes, preferably after adding the following parameters to [vfsgluster] share section(and of
> > course a
> > restart):
> >
> > kernel change notify = no
> > kernel oplocks = no
> > posix locking = no
> >
More information about the Gluster-users
mailing list