[sheepdog-users] problem starting sheepdog

[Hitoshi Mitake]

At Mon, 13 Jul 2015 11:34:27 +0200,
Valerio Pachera wrote:
> 
> 2015-07-13 11:16 GMT+02:00 Hitoshi Mitake <mitake.hitoshi at gmail.com>:
> 
> >
> >
> > Does the cluster cannot see 192.168.2.46 even after 10:41:43? The time
> > stamp of cluser's last log entry is 10:41:07. I'd like to clarify it.
> >
> 
> First and last log because the cluster is new.
> 
> I run
> 
> sheep -n \
> -c zookeeper:192.168.2.44:2181,192.168.2.45:2181,192.168.2.46:2181 \
> /var/lib/sheepdog,/mnt/sheep/0
> 
> on all 4 nodes (at 10:41) but the third node doesn't join the others.
> 
> I tried now to kill sheep on the third node and run it again but it still
> doesn't join the cluster
> 
> root at testc:~# dog node list
>   Id   Host:Port         V-Nodes       Zone
>    0   192.168.2.46:7000         3  771926208
> 
> root at testc:~# cat /var/lib/sheepdog/sheep.log
> Jul 13 11:30:12   INFO [main] md_add_disk(368) /mnt/sheep/0, vdisk nr 9,
> total disk 1
> Jul 13 11:30:12   INFO [main] zk_init(1498) version 3.4.5, address
> 192.168.2.44:2181,192.168.2.45:2181,192.168.2.46:2181/sheepdog, timeout
> 30000
> Jul 13 11:30:12   INFO [main] zk_init(1503) the negociated session timeout
> is 30000
> Jul 13 11:30:13 NOTICE [main] get_local_addr(522) found IPv4 address
> Jul 13 11:30:13   INFO [main] send_join_request(1083) IPv4 ip:192.168.2.46
> port:7000 going to join the cluster
> Jul 13 11:30:13 NOTICE [main] nfs_init(645) nfs server service is not
> compiled
> Jul 13 11:30:13   INFO [main] check_host_env(539) allowed open files set to
> suggested 830000
> Jul 13 11:30:13   INFO [main] main(1040) sheepdog daemon (version
> 0.9.0_190_gaa4d885) started

Thanks for your detailed report, Valerio. I'm cc-ing to Ruoyu.

Ruoyu, have you ever seen this sort of partitioning problem?

Thanks,
Hitoshi