[sheepdog-users] Cluster not starting

[Valerio Pachera]

2013/1/3 Valerio Pachera <sirio81 at gmail.com>:
> After enjoying some holiday, I restarted my cluster.

Of 3 nodes, only one is starting.

Jan 03 10:38:38 [main] jrnl_recover(230) opening the directory
/mnt/sheepdog/journal/
Jan 03 10:38:38 [main] jrnl_recover(235) starting journal recovery
Jan 03 10:38:38 [main] jrnl_recover(291) journal recovery complete
Jan 03 10:38:38 [main] send_join_request(1014) IPv4 ip:192.168.2.42 port:7000
Jan 03 10:38:38 [main] main(616) sheepdog daemon (version
0.5.5_23_gfacdf48) started
Jan 03 10:44:30 [main] cluster_wait_for_join_check(531) joining node
epoch too small: 10 vs 12
Jan 03 10:44:30 [main] sd_check_join_cb(974) 192.168.2.41:7000: ret =
0x1, cluster_status = 0x4
Jan 03 10:45:31 [main] cluster_wait_for_join_check(531) joining node
epoch too small: 10 vs 12
Jan 03 10:45:31 [main] sd_check_join_cb(974) 192.168.2.41:7000: ret =
0x1, cluster_status = 0x4
Jan 03 10:51:17 [main] jrnl_recover(230) opening the directory
/mnt/sheepdog/journal/
Jan 03 10:51:17 [main] jrnl_recover(235) starting journal recovery
Jan 03 10:51:17 [main] jrnl_recover(291) journal recovery complete
Jan 03 10:51:17 [main] send_join_request(1014) IPv4 ip:192.168.2.42 port:7000
Jan 03 10:51:17 [main] main(616) sheepdog daemon (version
0.5.5_23_gfacdf48) started
Jan 03 10:51:17 [main] update_cluster_info(799) status = 4, epoch =
12, finished: 0
Jan 03 10:51:59 [main] cluster_wait_for_join_check(531) joining node
epoch too small: 11 vs 12
Jan 03 10:51:59 [main] sd_check_join_cb(974) 192.168.2.43:7000: ret =
0x1, cluster_status = 0x4