Hi, I'm testing Slony-I 2.0.6. and, I noticed behavior has changed from before release(2.0.4). what I do is bellow. 1) I kill PostgreSQL process(postmaster) with SIGKILL on master. 2) and then I restart PostgreSQL. I checked log messages, appeartently behavior is different. a) log in Slony-I 2.0.6. -- 2011-03-09 00:40:27 JSTERROR remoteListenThread_1: "select con_origin, con_received, max(con_seqno) as con_seqno, max(con_timestamp) as con_timestamp from "_replication".sl_confirm where con_received <> 2 group by con_origin, con_received" 2011-03-09 00:40:33 JSTERROR remoteListenThread_1: "select ev_origin, ev_seqno, ev_timestamp, ev_snapshot, "pg_catalog".txid_snapshot_xmin(ev_snapshot), "pg_catalog".txid_snapshot_xmax(ev_snapshot), ev_type, ev_data1, ev_data2, ev_data3, ev_data4, ev_data5, ev_data6, ev_data7, ev_data8 from "_replication".sl_event e where (e.ev_origin = '1' and e.ev_seqno > '5000000025') order by e.ev_origin, e.ev_seqno limit 40" - no connection to the server -- *** Slony-I outputs above message endless. *** b) log in Slony-I 2.0.4 -- 2011-03-09 00:55:29 JSTERROR remoteListenThread_1: "select con_origin, con_received, max(con_seqno) as con_seqno, max(con_timestamp) as con_timestamp from "_replication".sl_confirm where con_received <> 2 group by con_origin, con_received" 2011-03-09 00:55:29 JSTCONFIG slon: child terminated status: 11; pid: 3154, current worker pid: 3154 2011-03-09 00:55:29 JSTCONFIG slon: restart of worker in 10 seconds -- *** In Slony-I 2.0.4, slon is restarted. It looks OK. *** the behavior in Slony-I 2.0.6 is intended ? If not so, please revise this. regards, -- NTT OSS Center Tomonari Katsumata
Please confirm if this change addresses the problem your describing? http://git.postgresql.org/gitweb?p=slony1-engine.git;a=commit;h=c2a06b4b36dc718fb849286de2dedffd4fc427a0
Hi, thank you for response. I apply the patch and test again. It's OK. slon could connect PostgreSQL soon. I'm looking forward to next release 2.0.7. --- NTT OSS Center Tomonari Katsumata