High-Availability Storage With GlusterFS - Mac Client

I followed your wonderful how to "High-Availability Storage With GlusterFS On Debian Lenny - Automatic File Replication Across Two Storage Servers" and it works perfectly on Linux.

I was trying to make a Mac OSX Snow Leopard Client working, but I cannot yet...

I have already installed MacFUSE as required and glusterfs compiled from source as described in the official website (http://www.gluster.com/community/documentation/index.php/GlusterFS_on_MAC_OS_X), I have also created the /etc/glusterfs/glusterfs.vol file and its content is exactly the same as other Kubuntu client. When I mount the glusterfs with the command

mount -t glusterfs /etc/glusterfs/glusterfs.vol /mnt/glusterfs

it works, and mount and df command show the correct glusterfs size, but the mount point on Mac is empty!

If I try to list the directory from user I receive

"Permession denied"

and if I try to do the same by root I receive:

"Device not configured"

but mount and df says that the gluster "share" is mounted and show the correct volume size...can somebody help me please?!

 

I am looking at the log file in real time, and the fact is that when I try to list the directory content I have a sort of lag during about 1 minute, and after that in the moment I receive the "Permession denied" error, in the log file I see that fuse has unmounted the mount point itselfs...maybe this can help?

 

Any errors in your logs (on the servers and the client)?

 

On the client side I see this

Code:

[2010-01-27 19:04:37] W [xlator.c:555:validate_xlator_volume_options] writebehind: option 'window-size' is deprecated, preferred is 'cache-size', continuing with correction
[2010-01-27 19:04:37] N [glusterfsd.c:1306:main] glusterfs: Successfully started
[2010-01-27 19:04:37] N [client-protocol.c:5733:client_setvolume_cbk] remote: Connected to 192.168.1.1:6996, attached to remote volume 'brick'.
[2010-01-27 19:04:37] N [afr.c:2194:notify] replicate: Subvolume 'remote' came back up; going online.
[2010-01-27 19:04:37] N [client-protocol.c:5733:client_setvolume_cbk] remote: Connected to 192.168.1.1:6996, attached to remote volume 'brick'.
[2010-01-27 19:04:37] N [afr.c:2194:notify] replicate: Subvolume 'remote' came back up; going online.
[2010-01-27 19:04:37] N [client-protocol.c:5733:client_setvolume_cbk] remote2: Connected to 192.168.1.2:6996, attached to remote volume 'brick'.
[2010-01-27 19:04:37] N [client-protocol.c:5733:client_setvolume_cbk] remote2: Connected to 192.168.1.2:6996, attached to remote volume 'brick'.
[2010-01-27 19:05:37] N [fuse-bridge.c:2577:fuse_thread_proc] fuse: unmounting /glusterfs
[2010-01-27 19:05:37] W [glusterfsd.c:898:cleanup_and_exit] glusterfs: shutting down

So no Errors, only "normal" Warning, now I'll post the Server side log

 

On the server side I see this:

Code:

[2010-01-27 15:55:12] N [server-protocol.c:7065:mop_setvolume] server: accepted client from 192.168.1.10:1023
[2010-01-27 15:55:12] N [server-protocol.c:7065:mop_setvolume] server: accepted client from 192.168.1.10:1022
[2010-01-27 15:56:05] N [server-protocol.c:7825:notify] server: 192.168.1.10:1023 disconnected
[2010-01-27 15:56:05] N [server-protocol.c:7825:notify] server: 192.168.1.10:1022 disconnected
[2010-01-27 15:56:05] N [server-helpers.c:779:server_connection_destroy] server: destroyed connection of Alessio-K-1043-2010/01/27-15:55:09:665816-remote
[2010-01-27 15:57:07] N [server-protocol.c:7065:mop_setvolume] server: accepted client from 192.168.1.10:1023
[2010-01-27 15:57:12] N [server-protocol.c:7065:mop_setvolume] server: accepted client from 192.168.1.10:1021
[2010-01-27 17:11:15] N [server-protocol.c:7065:mop_setvolume] server: accepted client from 192.168.1.50:1023
[2010-01-27 17:11:15] N [server-protocol.c:7065:mop_setvolume] server: accepted client from 192.168.1.50:1022
[2010-01-27 17:14:49] N [server-protocol.c:7825:notify] server: 192.168.1.50:1022 disconnected
[2010-01-27 17:14:49] N [server-protocol.c:7825:notify] server: 192.168.1.50:1023 disconnected
[2010-01-27 17:14:49] N [server-helpers.c:779:server_connection_destroy] server: destroyed connection of Mac-Pro-di-Alessio.local-17314-2010/01/27-17:11:14:617401-remote

192.168.1.10 is a Ubuntu client with no problem, and the Mac is the 192.168.1.50, I see no differences between theme, so I think the log is not telling us nothing...am I wrong?!

 

I wonder where this comes from?

 

I really do not know...but as I tought, the problem is on client side...

On the Ubuntu Pc, the log file looks like this:

Code:

[2010-01-21 08:49:28] W [xlator.c:555:validate_xlator_volume_options] writebehind: option 'window-size' is deprecated, preferred is 'cache-size', continuing with correction
[2010-01-21 08:49:28] N [glusterfsd.c:1306:main] glusterfs: Successfully started
[2010-01-21 08:49:29] N [client-protocol.c:5733:client_setvolume_cbk] remote: Connected to 192.168.1.1:6996, attached to remote volume 'brick'.
[2010-01-21 08:49:29] N [afr.c:2194:notify] replicate: Subvolume 'remote' came back up; going online.
[2010-01-21 08:49:29] N [client-protocol.c:5733:client_setvolume_cbk] remote: Connected to 192.168.1.1:6996, attached to remote volume 'brick'.
[2010-01-21 08:49:29] N [afr.c:2194:notify] replicate: Subvolume 'remote' came back up; going online.
[2010-01-21 08:49:29] N [client-protocol.c:5733:client_setvolume_cbk] remote2: Connected to 192.168.1.2:6996, attached to remote volume 'brick'.
[2010-01-21 08:49:29] N [client-protocol.c:5733:client_setvolume_cbk] remote2: Connected to 192.168.1.2:6996, attached to remote volume 'brick'.
[2010-01-21 09:08:51] W [glusterfsd.c:898:cleanup_and_exit] glusterfs: shutting down
[2010-01-21 09:08:51] N [fuse-bridge.c:2879:fini] fuse: Unmounting '/mnt/glusterfs'.
[2010-01-21 09:08:51] W [glusterfsd.c:898:cleanup_and_exit] glusterfs: shutting down

so even my Ubuntu does the same thing, but it disconnect when I unmount or halt the Pc, so the real question is: "why Mac decide itself to disconnect? what's the matter?!"

 

The difference between Ubuntu (which works well) and Mac (which does not), is that Ubuntu does one more thing before unmounting:

Code:

W [glusterfsd.c:898:cleanup_and_exit] glusterfs: shutting down

while the Mac directly unmount.

The problem is that the log file does not say anything to investigate...where I can look for something?! Maybe the MacFUSE client? Even if I used a 2.x version like glusterfs suggest con the Mac guide posted earlier.

 

I don't know why this is happening - unfortunately I don't have a Mac to test with...

 

...an Hackintosh to make some test?!?! I really do not know where search. GlusterFS project does not have a forum or email address...