[Sheepdog] Small sheepdog howto

[PCextreme B.V. - Wido den Hollander]

Hi,

I've just setup Sheepdog on four Ubuntu 9.10 AMD64 machines and wanted
to start testing it.

But i can't get the nodes to work ok, this is what i have done:

* Compiled sheepdog in /opt/sheepdog
* Compiled the patched qemu-kvm in /opt/qemu-kvm
* Configured corosync with:

        interface {
                ringnumber: 0
                bindnetaddr: 192.168.6.0
                mcastaddr: 226.94.1.1
                mcastport: 5405
        }

My subnet here is 192.168.6.0/24

But when i do "shepherd info -t dog" i get:

root at wido-desktop:~# shepherd info -t dog
  Idx	Node id (FNV-1a)    - Host:Port
--------------------------------------------------
  0	4c212f70289e9103 - 127.0.1.1:7000
  1	4c212f70289e9103 - 127.0.1.1:7000
  2	4c212f70289e9103 - 127.0.1.1:7000
  3	4c212f70289e9103 - 127.0.1.1:7000
* 4	4c212f70289e9103 - 127.0.1.1:7000
  5	4c212f70289e9103 - 127.0.1.1:7000
  6	4c212f70289e9103 - 127.0.1.1:7000
  7	4c212f70289e9103 - 127.0.1.1:7000
  8	4c212f70289e9103 - 127.0.1.1:7000
root at wido-desktop:~#

All the nodes claim to have the same nodeid? I have used
"start-sheepdog" to start the "collie" daemon.

On my nodes it looks like:

root at wido-desktop:~# ps aux|grep collie
root      4492  0.0  0.0  10284   592 ?        Ss   15:11
0:00 /usr/src/sheepdog/collie/collie --port 7000 /srv/sheepdog/0 -d
root      4493  0.0  0.2  65752 18132 ?        Ssl  15:11
0:00 /usr/src/sheepdog/collie/collie --port 7000 /srv/sheepdog/0 -d
root      4730  0.0  0.0   7336   888 pts/0    S+   16:18   0:00 grep
collie
root at wido-desktop:~#

So they see eachother, but all bind to 127.0.1.1:7000? I think i am
missing something here.

Would somebody be so kind to share a corosync.conf for review? I think i
am making a mistake there.

When i do:

shepherd mkfs --copies=3

It doesn't return any output.

collie segfaults with:
[12279.751092] collie[4812]: segfault at 3 ip 0000000000000003 sp
00007f34d5945fb8 error 14 in collie[400000+e000]

When i run collie on forground mode it doesn't segfault, and gives the
following output:

collie: worker_routine(175) started this thread 0
collie: worker_routine(175) started this thread 0
collie: worker_routine(175) started this thread 0
collie: worker_routine(175) started this thread 0
collie: sd_confch(653) confchg nodeid 1cb8d44
collie: sd_confch(655) 1 0 1
collie: sd_confch(659) [0] node_id: 30117188, pid: 4853, reason:
-1138069504
collie: sd_deliver(532) op: 1, done: 0, size: 41056, from:
127.0.1.1:7000
collie: __sd_deliver(468) op: 1, done: 0, size: 41056, from:
127.0.1.1:7000
collie: sd_deliver(532) op: 1, done: 1, size: 41056, from:
127.0.1.1:7000
collie: __sd_deliver(468) op: 1, done: 1, size: 41056, from:
127.0.1.1:7000
collie: print_node_list(244) l nodeid: 1cb8d44, pid: 4853, ip:
127.0.1.1:7000
collie: listen_handler(336) accepted a new connection, 8
collie: cluster_queue_request(177) 0x1abf940 19
collie: client_handler(297) closed a connection, 8
collie: listen_handler(336) accepted a new connection, 8
collie: cluster_queue_request(177) 0x1abf850 21
collie: sd_deliver(532) op: 2, done: 0, size: 136, from: 127.0.1.1:7000
collie: __sd_deliver(468) op: 2, done: 0, size: 136, from:
127.0.1.1:7000
collie: listen_handler(336) accepted a new connection, 10
collie: client_handler(297) closed a connection, 10
collie: listen_handler(336) accepted a new connection, 10
collie: client_handler(297) closed a connection, 10
collie: listen_handler(336) accepted a new connection, 10
collie: client_handler(297) closed a connection, 10
collie: sd_deliver(532) op: 2, done: 1, size: 136, from: 127.0.1.1:7000
collie: __sd_deliver(468) op: 2, done: 1, size: 136, from:
127.0.1.1:7000
collie: vdi_op_done(438) 3
collie: client_handler(297) closed a connection, 8

This morning i had sheepdog running on one node and it worked fine, i
had a KVM VM running on top of it without any troubles.

I made a checkout of the GIT repository this morning, so i am using the
latest version.

Any idea?


-- 
Met vriendelijke groet,

Wido den Hollander
Hoofd Systeembeheer / CSO
Telefoon Support Nederland: 0900 9633 (45 cpm)
Telefoon Support België: 0900 70312 (45 cpm)
Telefoon Direct: (+31) (0)20 50 60 104
Fax: +31 (0)20 50 60 111
E-mail: support at pcextreme.nl
Website: http://www.pcextreme.nl
Kennisbank: http://support.pcextreme.nl/
Netwerkstatus: http://nmc.pcextreme.nl