首先是查询等操作无法完成，一只没有反应，等待时间超长后gp会崩溃
[gpadmin@greenplum gpconfig]$ psql postgres gpadmin
select * from test；
2013-10-23 11:34:06.747775 CST,"gpadmin","postgres",p29544,th-1750079744,"[local]",,2013-10-23 11:05:23 CST,1426,con30,cmd24,seg-1,,dx100,x1426,sx1,"LOG","00000","statement: select * from test;",,,,,,"select * from test;",0,,"postgres.c",1535,
日志也无从下手
重启gp等都不可以，后来采取极端做饭，重新初始化GP，重启，还是不行
首先初始化的时候要停止端口5432和删除/data dpseg-1的内容
删除/tmp//tmp/.s.PGSQL.5432 和 /tmp/.s.PGSQL.5432.lock
20131023:18:38:39:016109 gpinitsystem:greenplum:gpadmin-[INFO]:-Checking Master host
20131023:18:38:39:016109 gpinitsystem:greenplum:gpadmin-[WARN]:-No lock file /tmp/.s.PGSQL.5432.lock but process running on port 5432
20131023:18:38:39:gpinitsystem:greenplum:gpadmin-[FATAL]:-Found indication of postmaster process on port 5432 on Master host Script Exiting!
[gpadmin@greenplum gpconfig]$ lsof -i:5432
COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
postgres 10738 gpadmin 3u IPv4 264510 0t0 TCP *:postgres (LISTEN)
postgres 10738 gpadmin 4u IPv6 264511 0t0 TCP *:postgres (LISTEN)
[gpadmin@greenplum gpconfig]$ kill -9 10738
后来初始化不知道怎么回事，重启报了一个错，这个错误信息很重要
20131023:18:41:13:031907 gpstart:greenplum:gpadmin-[INFO]:-Command pg_ctl reports Master greenplum instance active
20131023:18:41:13:031907 gpstart:greenplum:gpadmin-[WARNING]:-FATAL: DTM initialization: failure during startup/recovery, retry failed, check segment status (cdbtm.c:1351)
20131023:18:41:13:031907 gpstart:greenplum:gpadmin-[INFO]:-Check status of database with gpstate utility
20131023:18:41:13:021367 gpinitsystem:greenplum:gpadmin-[INFO]:-Completed restart of Greenplum instance in production mode
20131023:18:41:13:021367 gpinitsystem:greenplum:gpadmin-[INFO]:-Loading gp_toolkit...
psql: FATAL: DTM initialization: failure during startup/recovery, retry failed, check segment status (cdbtm.c:1351)
20131023:18:41:13:gpinitsystem:greenplum:gpadmin-[FATAL]:-Failed to retrieve rolname. Script Exiting!
原来是iptables没有关闭
首先stop iptables
[gpadmin@greenplum gpconfig]$ sudo service iptables stop
然后
[gpadmin@greenplum gpconfig]$ gpstop -u
20131024:10:00:00:017083 gpstart:greenplum:gpadmin-[INFO]:-Process results...
20131024:10:00:00:017083 gpstart:greenplum:gpadmin-[INFO]:-----------------------------------------------------
20131024:10:00:00:017083 gpstart:greenplum:gpadmin-[INFO]:- Successful segment starts = 2
20131024:10:00:00:017083 gpstart:greenplum:gpadmin-[INFO]:- Failed segment starts = 0
20131024:10:00:00:017083 gpstart:greenplum:gpadmin-[INFO]:- Skipped segment starts (segments are marked down in configuration) = 0
20131024:10:00:00:017083 gpstart:greenplum:gpadmin-[INFO]:-----------------------------------------------------
20131024:10:00:00:017083 gpstart:greenplum:gpadmin-[INFO]:-
20131024:10:00:00:017083 gpstart:greenplum:gpadmin-[INFO]:-Successfully started 2 of 2 segment instances
20131024:10:00:00:017083 gpstart:greenplum:gpadmin-[INFO]:-----------------------------------------------------
20131024:10:00:00:017083 gpstart:greenplum:gpadmin-[INFO]:-Starting Master instance greenplum directory /data/master_database/gpseg-1
20131024:10:00:01:017083 gpstart:greenplum:gpadmin-[INFO]:-Command pg_ctl reports Master greenplum instance active
20131024:10:00:01:017083 gpstart:greenplum:gpadmin-[INFO]:-Database successfully started
后来一切OK
gpadmin_db=# select * from test;
id | name
----+------
(0 rows)
gpadmin_db=# INSERT INTO test (id,name) VALUES (1,a);
ERROR: column "a" does not exist
LINE 1: INSERT INTO test (id,name) VALUES (1,a);
^
gpadmin_db=# INSERT INTO test (id,name) VALUES (1,'a');
INSERT 0 1
gpadmin_db=# select * from test ;
id | name
----+------
1 | a
(1 row)