I am not offended :)
I understand now what is wrong (I think).
To summerize what you want:
On servers 1-N export the directory's /glusterfsr1 data so it can be mounted by
clients (servers can be clients too) on mount point /export/glusterfs.
The /glusterfsr1 should be a raid-1 mirrored array.
Here is what you need:
/etc/glusterfs/glusterfsd.vol (the same as the last one I sent you)
**********************************
<snip of cmd line from volgen>
# TRANSPORT-TYPE tcp
# PORT 6996
volume posix
type storage/posix
option directory /glusterfs1
end-volume
volume locks
type features/locks
subvolumes posix
end-volume
volume brick
type performance/io-threads
option thread-count 8
subvolumes locks
end-volume
volume server
type protocol/server
option transport-type tcp
option auth.addr.brick.allow *
option listen-port 6996
subvolumes brick
end-volume
*************************
/etc/glusterfs/glusterfs.vol is
# RAID 1
# TRANSPORT-TYPE tcp
# PORT 6996
volume raidvol-0
type protocol/client
option transport-type tcp
option remote-host <name of server 0>
option remote-port 6996
option remote-subvolume brick
end-volume
volume raidvol-1
type protocol/client
option transport-type tcp
option remote-host <name of server 1>
option remote-port 6996
option remote-subvolume brick
end-volume
volume raidvol-2
type protocol/client
option transport-type tcp
option remote-host <name of server 2>
option remote-port 6996
option remote-subvolume brick
end-volume
. . .
volume raidvol-N
type protocol/client
option transport-type tcp
option remote-host <name of server N>
option remote-port 6996
option remote-subvolume brick
end-volume
volume mirror-0
type cluster/replicate
subvolumes raidvol-0 raidvol-1 raidvol-2 . . . raidvol-N
end-volume
volume writebehind
type performance/write-behind
option cache-size 4MB
subvolumes mirror-0
end-volume
volume io-cache
type performance/io-cache
option cache-size 1GB
subvolumes writebehind
end-volume
**********************************************
Then the mount command I gave you last time should be this (it can be used on
clients or servers):
/etc/glusterfs/glusterfs.vol /export/glusterfs glusterfs
defaults 0 0
^C
[email protected] wrote:
Someone who knows pipe up here, but I think that I am confused.
Lots of things look wrong to me, but I am not sure what you are trying to
do, and what is different between 2.x and 3.x.
I think what you want:
1 (or more) mount points/directories exported from your 2 servers
server002 and bee003 to be mounted by your N nodes.
You missed the <...>. I have 30 nodes. I've created /glusterfsr1 on each,
and I've created /export/glusterfs on each. The intent is that
/glusterfsr1 should be part of the glusterfs, and it should be mounted on
each on /export/glusterfs. ALL of them are servers.
<snip>
Are you exporting 2 mounts from 4 different servers?
One mount point, from 30 servers. One mount point on each server. At
least, this is what glusterfs-volgen created as
/etc/glusterfs/glusterfsr1-export.vol and /etc/glusterfsr1-mount.vol.
Rather than renaming them, I symlinked them to
/etc/glusterfs/glusterfsd.vol and /etc/glusterfs/glusterfs.vol,
respectively.
Are you really trying to create a raid0 array across all your client
nodes?
No, all the server nodes. I haven't done anything with a client node yet -
I can't start glusterfs on my servers.
Hey, no insult meant to you, Chad, but isn't there *anyone* else on this
list, who's been working with glusterfs longer? I mean, from the googling
I've been doing, it looks like 3.0 just came out last fall, and though
you've been doing your best to help, no one else seems to be stepping up.
mark
_______________________________________________
Gluster-users mailing list
[email protected]
http://gluster.org/cgi-bin/mailman/listinfo/gluster-users
