At Mon, 29 Aug 2011 15:38:52 +0200, Valerio Pachera wrote: > > after some time (no idea how much), i get this from node1: > > # collie cluster info > running > Ctime Epoch Nodes > 2011-08-23 23:32:17 4 [192.168.2.27:7000, 192.168.2.59:7000 > > The same commando run on node2 (192.168.2.59)...hangs up. > > This is the last part of sheepdog.log on node1. > Aug 29 12:09:48 __recover_one(1181) cannot find target node, a34c67000007e6 > Aug 29 12:09:48 __recover_one(1176) recover obj a34c67000007e6 from epoch 3 > Aug 29 12:09:48 find_tgt_node(1107) 27, 64, 56, 128, 0 > Aug 29 12:09:48 find_tgt_node(1114) 27, 0, 56, 128 > Aug 29 03:34:23 listen_handler(595) accepted a new connection, 12 > Aug 29 03:34:23 start_cpg_event_work(1345) 0 2 > Aug 29 03:34:23 cluster_queue_request(243) 0x4cb0010 87 > Aug 29 03:34:23 client_handler(545) closed a connection, 12 > > > On node3 still the wrong date from 'collie cluster info' but sheep.log > seems fine > Aug 29 12:13:08 listen_handler(523) accepted a new connection, 11 > Aug 29 12:13:08 start_cpg_event_work(1448) 0 2 > Aug 29 12:13:08 cluster_queue_request(255) 0x37d8960 87 > Aug 29 12:13:08 client_handler(484) closed a connection, 11 > Aug 29 03:34:07 listen_handler(523) accepted a new connection, 11 > Aug 29 03:34:07 start_cpg_event_work(1448) 0 2 > Aug 29 03:34:07 cluster_queue_request(255) 0x37d8960 87 > Aug 29 03:34:07 client_handler(484) closed a connection, 11 > Aug 29 03:34:35 listen_handler(523) accepted a new connection, 11 > Aug 29 03:34:35 start_cpg_event_work(1448) 0 2 > Aug 29 03:34:35 cluster_queue_request(255) 0x37d8960 87 > Aug 29 03:34:35 client_handler(484) closed a connection, 11 This line should be "client_handler(545)" if you have the same revision of Sheepdog installed on node2 and node3. Can you update all the sheeps to the same version? Thanks, Kazutaka |