[sheepdog-users] Unaspected output from collie node recovery

[MORITA Kazutaka]

At Wed, 19 Dec 2012 08:54:55 +0100,
Valerio Pachera wrote:
> 
> 2012/12/19 Liu Yuan <namei.unix at gmail.com>:
> >> [main] connect_to(244) failed to connect to 192.168.2.215:7000:
> >> Connection refused
> 
> > Wait for some minutes and retry the command?
> 
> Still the same.
> 
> I see no network/disk activity on any of the nodes.
> The node that has die is 192.168.2.43.
> 
> sheep.log (last rows)
> Dec 19 08:47:51 [io 10455] do_process_work(1238) 97, 0, 7
> Dec 19 08:47:51 [main] client_handler(815) 4, rx 0, tx 0
> Dec 19 08:47:51 [main] finish_tx(705) connection from: 23, 192.168.2.43:39963
> Dec 19 08:47:51 [main] client_handler(815) 19, rx 0, tx 3
> Dec 19 08:47:51 [main] clear_client_info(744) connection seems to be dead
> Dec 19 08:47:51 [main] clear_client_info(765) refcnt:0, fd:23,
> 192.168.2.43:39963
> Dec 19 08:47:51 [main] destroy_client(735) connection from: 192.168.2.43:39963
> Dec 19 08:49:53 [main] listen_handler(880) accepted a new connection: 23
> Dec 19 08:49:53 [main] client_handler(815) 1, rx 0, tx 0
> Dec 19 08:49:53 [main] finish_rx(618) 23, 192.168.2.43:39968
> Dec 19 08:49:53 [main] queue_request(355) STAT_RECOVERY, 1
> Dec 19 08:49:53 [io 10456] do_process_work(1238) 97, 0, 7
> Dec 19 08:49:53 [main] client_handler(815) 4, rx 0, tx 0
> Dec 19 08:49:53 [main] finish_tx(705) connection from: 23, 192.168.2.43:39968
> Dec 19 08:49:53 [main] client_handler(815) 19, rx 0, tx 3
> Dec 19 08:49:53 [main] clear_client_info(744) connection seems to be dead
> Dec 19 08:49:53 [main] clear_client_info(765) refcnt:0, fd:23,
> 192.168.2.43:39968
> Dec 19 08:49:53 [main] destroy_client(735) connection from: 192.168.2.43:39968

Can you give us the output of 'collie cluster info' on 192.168.2.41?

The gone node is 192.168.2.43, but 192.168.2.41 cannot connect to
192.168.2.215.  Is that correct?

Thanks,

Kazutaka