GP failover，primary和mirror切換實驗
http://blog.sina.com.cn/s/blog_9869114e0101k1nc.html
一、恢復失敗的segment
出現segment節點down，恢復GP的思路：
if
"segment down" then
case "determine case(errer,message,log,gpstate and so on)"
when "data error[missing or corrupted data directory]"
then
if "mirror segment是否可用" then
"gprecoverseg --從mirror segment中恢復primary segment丟失的部分"
else
"rebuild gp database system from backup file"
end if
when "進程錯誤--postmaster process not is running"
then
if "mirror segment是否可用" then
"gprecoverseg --從mirror segment中恢復primary segment丟失的部分"
else
"重啟GP數據庫--gpstop gpstart"
end if
when "主機出現故障--dns,network,host down and so on"
then
if "主機故障是否恢復" then
"故障恢復，切換主機在線"
if "mirror segment是否可用" then
"gprecoverseg --從mirror segment中恢復primary segment丟失的部分"
else
"重啟GP數據庫--gpstop gpstart"
end if
elsif "mirror segment是否可用"
then
"從備機上恢復primary segment丟失的數據 --gprecoverseg -p | -s"
else
"rebuild gp database system from backup file"
end if
end case
end if
a、當發生主機出現故障，主機所有的primary和mirror節點都會標注為down，並踢出集群。如果GP集群沒有部署mirror segment，那么整個GP集群將會宕機。
b、通過gprecoverseg 工具恢復segment丟失的數據部分，恢復期間，數據庫寫的操作將暫時的延緩。
c、gpreceverseg完成之后，數據庫將進行重新同步模式，並開始復制改變的文件。這些操作都都是通過后台進程來完成，在這個階段數據庫是可以進行正常的服務。
d、當重新同步完成，系統編程同步化狀態，可以通過gpstate -m查詢
e、主備切換--切換segment到預設角色（primary，mirror）。當出現segment故障時，mirror segment會替換故障的primary segment，成為primary segment，之前故障的primary segment在使用gprecoverseg恢復后，之前的primary不會切回primary segment，而是變為mirror segment。
二、Preferred role和current role的概念
當配置了mirror segment時，存在Preferred role和current role的概念，當primary segment節點down了，mirror segment節點會接管primary的位置，成為新的primary segment，之前down掉的primary segment在online后會成為mirror segment。那么預設角色是指正常情況下primary segment各種的位置，當前角色是只出現failover后，mirror取代之前的primary，當前的segment的角色。
# 顯示primary、mirror segment存在的問題
gpstate -e
#顯示mirror segment實例的當前角色，同步狀態等信息
gpstate -m
#可以通過gprecover -r將處於當前角色的segment切回預設角色。
gprecoverseg -r
# 顯示primary、mirror segment存在的問題
gpstate -e
案例1：
3個primary segment+3個mirror segment+ 1個master server模擬一個主機宕機的情況。
分析可能出現的問題：
一個主機host2宕機，上面相應的1個primary segment和1個mirror segment都
會down掉。由於做了mirror，host2 primary對於的mirror在host1上，會被激活，成為新的primary節點,當前的角色也從mirror切換成primary。故障的primary節點在online后，會成功新的mirror節點，當前角色從primary角色切換成mirror角色。
模擬故障前的信息確認：
$ gpstate --查看gp數據庫的總體狀態
- Master instance = Active --master狀態
- Master standby = No master standby configured
- Total segment instance count from metadata = 6 --segment節點數
-----------------------------------------------------
- Primary Segment Status
-----------------------------------------------------
- Total primary segments = 3
- Total primary segment valid (at master) = 3
- Total primary segment failures (at master) = 0
- Total number of postmaster.pid files missing = 0
- Total number of postmaster.pid files found = 3
- Total number of postmaster.pid PIDs missing = 0
- Total number of postmaster.pid PIDs found = 3
- Total number of /tmp lock files missing = 0
- Total number of /tmp lock files found = 3
- Total number postmaster processes missing = 0
- Total number postmaster processes found = 3
-----------------------------------------------------
- Mirror Segment Status
-----------------------------------------------------
- Total mirror segments = 3
- Total mirror segment valid (at master) = 3
- Total mirror segment failures (at master) = 0
- Total number of postmaster.pid files missing = 0
- Total number of postmaster.pid files found = 3
- Total number of postmaster.pid PIDs missing = 0
- Total number of postmaster.pid PIDs found = 3
- Total number of /tmp lock files missing = 0
$ gpstate -m --查看mirror節點情況
-Obtaining Segment details from master...
--------------------------------------------------------------
--Current GPDB mirror list and status
--Type = Spread
--------------------------------------------------------------
- Mirror Datadir Port Status Data Status
- gp02 /gpsegment2/gpseg0 50000 Passive Synchronized
- gp03 /gpsegment2/gpseg1 50000 Passive Synchronized
- gp01 /gpsegment2/gpseg2 50000 Passive Synchronized
--------------------------------------------------------------
$ gpstate -e --顯示primary、mirror segment存在的問題
-Gathering data from segments... ----------------------------------------------------- -Segment Mirroring Status Report ----------------------------------------------------- -All segments are running normally
模擬gp02的主機故障：
[gp02]# killall -u gpadmin --殺掉gp02的gpadmin下所有的進程
[root@gp02 ~]# killall -u gpadmin
[root@gp02 ~]# ps -ef | grep gpadmin
root 6733 3155 0 03:26 pts/0 00:00:00 grep gpadmin
查詢故障后，gp數據庫的狀態：
$ gpstate -e
-----------------------------------------------------
-Segment Mirroring Status Report
-----------------------------------------------------
-Segments with Primary and Mirror Roles Switched
- Current Primary Port Mirror Port
- gp03 50000 gp02 40000
-----------------------------------------------------
-Primaries in Change Tracking
- Current Primary Port Change tracking size Mirror Port
- gp01 40000 100 MB gp02 50000
- gp03 50000 100 MB gp02 40000
--gp02的primary segment對於的mirror segment的位置：gp03的/gpsegment2/gpseg1。
顯示mirror segment實例的當前角色，同步狀態等信息
$ gpstate -m
[INFO]:-------------------------------------------------------------- [INFO]:--Current GPDB mirror list and status [INFO]:--Type = Spread
[INFO]:-------------------------------------------------------------- [INFO]:- Mirror Datadir Port Status Data Status [WARNING]:-gp02 /gpsegment2/gpseg0 50000 Failed
[INFO]:- gp03 /gpsegment2/gpseg1 50000 Acting as Primary Change Tracking [INFO]:- gp01 /gpsegment2/gpseg2 50000 Passive Synchronized [INFO]:-------------------------------------------------------------- [WARNING]:-1 segment(s) configured as mirror(s) are acting as primaries [WARNING]:-1 segment(s) configured as mirror(s) have failed [WARNING]:-1 mirror segment(s) acting as primaries are in change tracking
查詢結果gpseg0 的mirror segment 狀態為failed，gpseg1的mirror成為了primary節點。
[INFO]:----------------------------------------------------- [INFO]:- Segment Info [INFO]:- Hostname = gp02 [INFO]:- Address = gp02 [INFO]:- Datadir = /gpsegment2/gpseg0 [INFO]:- Port = 50000 [INFO]:- Mirroring Info [INFO]:- Current role = Mirror [INFO]:- Preferred role = Mirror [WARNING]:- Mirror status = Out of Sync <<<<<<<< [INFO]:- Status [WARNING]:- PID = Not found <<<<<<<< [WARNING]:- Configuration reports status as = Down <<<<<<<< [WARNING]:- Segment status = Down in configuration <<<<<<<< [INFO]:----------------------------------------------------- [INFO]:- Segment Info [INFO]:- Hostname = gp03 [INFO]:- Address = gp03 [INFO]:- Datadir = /gpsegment2/gpseg1 [INFO]:- Port = 50000 [INFO]:- Mirroring Info [INFO]:- Current role = Primary [INFO]:- Preferred role = Mirror [INFO]:- Mirror status = Change Tracking [INFO]:- Change Tracking Info [INFO]:- Change tracking data size = 100 MB [INFO]:- Status [INFO]:- PID = 5901 [INFO]:- Configuration reports status as = Up [INFO]:- Database status = Up [INFO]:----------------------------------------------------- [INFO]:- Segment Info [INFO]:- Hostname = gp02 [INFO]:- Address = gp02 [INFO]:- Datadir = /gpsegment1/gpseg1 [INFO]:- Port = 40000 [INFO]:- Mirroring Info [INFO]:- Current role = Mirror [INFO]:- Preferred role = Primary [WARNING]:- Mirror status = Out of Sync <<<<<<<< [INFO]:- Status [WARNING]:- PID = Not found <<<<<<<< [WARNING]:- Configuration reports status as = Down <<<<<<<< [WARNING]:- Segment status = Down in configuration <<<<<<<< [INFO]:-----------------------------------------------------
確定主機無故障后，恢復丟失的數據：
[gpadmin@gp01 gpsegment2]$ gprecoverseg
20140310:04:09:19:021098 gprecoverseg:gp01:gpadmin-[INFO]:-Starting gprecoverseg with args:
20140310:04:09:19:021098 gprecoverseg:gp01:gpadmin-[INFO]:-local Greenplum Version: 'postgres (Greenplum Database) 4.2.2.4 build 1 Community Edition'
20140310:04:09:19:021098 gprecoverseg:gp01:gpadmin-[INFO]:-master Greenplum Version: 'PostgreSQL 8.2.15 (Greenplum Database 4.2.2.4 build 1 Community Edition) on x86_64-unknown-linux-gnu, compiled by GCC gcc (GCC) 4.4.2 compiled on Oct 17 2012 11:52:28'
20140310:04:09:19:021098 gprecoverseg:gp01:gpadmin-[INFO]:-Obtaining Segment details from master...
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:-Greenplum instance recovery parameters
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:----------------------------------------------------------
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:-Recovery type = Standard
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:----------------------------------------------------------
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:-Recovery 1 of 2
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:----------------------------------------------------------
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Synchronization mode = Incremental
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Failed instance host = gp02
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Failed instance address = gp02
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Failed instance directory = /gpsegment2/gpseg0
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Failed instance port = 50000
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Failed instance replication port = 51000
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Recovery Source instance host = gp01
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Recovery Source instance address = gp01
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Recovery Source instance directory = /gpsegment1/gpseg0
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Recovery Source instance port = 40000
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Recovery Source instance replication port = 41000
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Recovery Target = in-place
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:----------------------------------------------------------
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:-Recovery 2 of 2
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:----------------------------------------------------------
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Synchronization mode = Incremental
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Failed instance host = gp02
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Failed instance address = gp02
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Failed instance directory = /gpsegment1/gpseg1
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Failed instance port = 40000
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Failed instance replication port = 41000
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Recovery Source instance host = gp03
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Recovery Source instance address = gp03
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Recovery Source instance directory = /gpsegment2/gpseg1
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Recovery Source instance port = 50000
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Recovery Source instance replication port = 51000
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:- Recovery Target = in-place
20140310:04:09:20:021098 gprecoverseg:gp01:gpadmin-[INFO]:----------------------------------------------------------
Continue with segment recovery procedure Yy|Nn (default=N):
> y
20140310:04:09:23:021098 gprecoverseg:gp01:gpadmin-[INFO]:-2 segment(s) to recover
20140310:04:09:23:021098 gprecoverseg:gp01:gpadmin-[INFO]:-Ensuring 2 failed segment(s) are stopped
updating flat files
20140310:04:09:26:021098 gprecoverseg:gp01:gpadmin-[INFO]:-Updating configuration with new mirrors
20140310:04:09:27:021098 gprecoverseg:gp01:gpadmin-[INFO]:-Updating mirrors
.
20140310:04:09:28:021098 gprecoverseg:gp01:gpadmin-[INFO]:-Starting mirrors
20140310:04:09:28:021098 gprecoverseg:gp01:gpadmin-[INFO]:-Commencing parallel primary and mirror segment instance startup, please wait...
...
20140310:04:09:31:021098 gprecoverseg:gp01:gpadmin-[INFO]:-Process results...
20140310:04:09:31:021098 gprecoverseg:gp01:gpadmin-[INFO]:-Updating configuration to mark mirrors up
20140310:04:09:31:021098 gprecoverseg:gp01:gpadmin-[INFO]:-Updating primaries
20140310:04:09:31:021098 gprecoverseg:gp01:gpadmin-[INFO]:-Commencing parallel primary conversion of 2 segments, please wait...
..
20140310:04:09:33:021098 gprecoverseg:gp01:gpadmin-[INFO]:-Process results...
20140310:04:09:33:021098 gprecoverseg:gp01:gpadmin-[INFO]:-Done updating primaries
20140310:04:09:33:021098 gprecoverseg:gp01:gpadmin-[INFO]:-******************************************************************
20140310:04:09:33:021098 gprecoverseg:gp01:gpadmin-[INFO]:-Updating segments for resynchronization is completed.
20140310:04:09:33:021098 gprecoverseg:gp01:gpadmin-[INFO]:-For segments updated successfully, resynchronization will continue in the background.
20140310:04:09:33:021098 gprecoverseg:gp01:gpadmin-[INFO]:-
20140310:04:09:33:021098 gprecoverseg:gp01:gpadmin-[INFO]:-Use gpstate -s to check the resynchronization progress.
20140310:04:09:33:021098 gprecoverseg:gp01:gpadmin-[INFO]:-******************************************************************
a、gprecoverseg才用 Synchronization mode恢復gp02的segment節點丟失的數據。
b、采用增量同步的方式恢復/gpsegment1/gpseg1丟失的數據。--gp02 primary
c、采用增量同步的方式恢復/gpsegment2/gpseg0丟失的數據。--gp01 primary的mirror
恢復后查詢gp數據庫的狀態：
$ gpstate -e --查詢mirror存在的錯誤，查詢顯示gp02的primary節點被mirror替換掉，目前的primary的位置在gp03上。
[INFO]:-----------------------------------------------------
[INFO]:-Segment Mirroring Status Report
[INFO]:-----------------------------------------------------
[INFO]:-Segments with Primary and Mirror Roles Switched
[INFO]:- Current Primary Port Mirror Port
[INFO]:- gp03 50000 gp02 40000
$ gpstate -m
[INFO]:-Obtaining Segment details from master..
[INFO]:--------------------------------------------------------------
[INFO]:--Current GPDB mirror list and status
[INFO]:--Type = Spread
[INFO]:--------------------------------------------------------------
[INFO]:- Mirror Datadir Port Status Data Status
[INFO]:- gp02 /gpsegment2/gpseg0 50000 Passive Synchronized
[INFO]:- gp03 /gpsegment2/gpseg1 50000 Acting as Primary Synchronized
[INFO]:- gp01 /gpsegment2/gpseg2 50000 Passive Synchronized
[INFO]:--------------------------------------------------------------
[WARNING]:-1 segment(s) configured as mirror(s) are acting as primaries
切換current role到prefered role
[gpadmin@gp01 gpsegment2]$ gprecoverseg -r
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:-Starting gprecoverseg with args: -r
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:-local Greenplum Version: 'postgres (Greenplum Database) 4.2.2.4 build 1 Community Edition'
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:-master Greenplum Version: 'PostgreSQL 8.2.15 (Greenplum Database 4.2.2.4 build 1 Community Edition) on x86_64-unknown-linux-gnu, compiled by GCC gcc (GCC) 4.4.2 compiled on Oct 17 2012 11:52:28'
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:-Obtaining Segment details from master...
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:-Greenplum instance recovery parameters
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:----------------------------------------------------------
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:-Recovery type = Rebalance
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:----------------------------------------------------------
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:-Unbalanced segment 1 of 2
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:----------------------------------------------------------
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:- Unbalanced instance host = gp03
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:- Unbalanced instance address = gp03
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:- Unbalanced instance directory = /gpsegment2/gpseg1
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:- Unbalanced instance port = 50000
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:- Unbalanced instance replication port = 51000
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:- Balanced role = Mirror
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:- Current role = Primary
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:----------------------------------------------------------
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:-Unbalanced segment 2 of 2
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:----------------------------------------------------------
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:- Unbalanced instance host = gp02
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:- Unbalanced instance address = gp02
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:- Unbalanced instance directory = /gpsegment1/gpseg1
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:- Unbalanced instance port = 40000
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:- Unbalanced instance replication port = 41000
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:- Balanced role = Primary
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:- Current role = Mirror
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[INFO]:----------------------------------------------------------
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[WARNING]:-This operation will cancel queries that are currently executing.
20140310:04:35:51:021852 gprecoverseg:gp01:gpadmin-[WARNING]:-Connections to the database however will not be interrupted.
Continue with segment rebalance procedure Yy|Nn (default=N):
> y
20140310:04:36:01:021852 gprecoverseg:gp01:gpadmin-[INFO]:-Getting unbalanced segments
20140310:04:36:01:021852 gprecoverseg:gp01:gpadmin-[INFO]:-Stopping unbalanced primary segments...
..
20140310:04:36:03:021852 gprecoverseg:gp01:gpadmin-[INFO]:-Triggering segment reconfiguration
20140310:04:36:07:021852 gprecoverseg:gp01:gpadmin-[INFO]:-Starting segment synchronization
............
20140310:04:36:20:021852 gprecoverseg:gp01:gpadmin-[INFO]:-******************************************************************
20140310:04:36:20:021852 gprecoverseg:gp01:gpadmin-[INFO]:-The rebalance operation has completed successfully.
20140310:04:36:20:021852 gprecoverseg:gp01:gpadmin-[INFO]:-There is a resynchronization running in the background to bring all
20140310:04:36:20:021852 gprecoverseg:gp01:gpadmin-[INFO]:-segments in sync.
20140310:04:36:20:021852 gprecoverseg:gp01:gpadmin-[INFO]:-
20140310:04:36:20:021852 gprecoverseg:gp01:gpadmin-[INFO]:-Use gpstate -s to check the resynchronization progress.
20140310:04:36:20:021852 gprecoverseg:gp01:gpadmin-[INFO]:-******************************************************************
查詢切換后的狀態：
$ gpstate -e
[INFO]:-Starting gpstate with args: -e
[INFO]:-local Greenplum Version: 'postgres (Greenplum Database) 4.2.2.4 build 1 Community Edition'
[INFO]:-master Greenplum Version: 'PostgreSQL 8.2.15 (Greenplum Database 4.2.2.4 build 1 Community Edition) on x86_64-unknown-linux-gnu, compiled by GCC gcc (GCC) 4.4.2 compiled on Oct 17 2012 11:52:28'
[INFO]:-Obtaining Segment details from master...
[INFO]:-Gathering data from segments...
[INFO]:-----------------------------------------------------
[INFO]:-Segment Mirroring Status Report
[INFO]:-----------------------------------------------------
[INFO]:-Segment Pairs in Resynchronization
[INFO]:- Current Primary Port Resync mode Est. resync progress Data synced Est. total to sync Est. resync end time Change tracking size Mirror Port
[INFO]:- gp02 40000 Incremental 100% 224 kB Sync complete; awaiting config change 100 MB gp03 50000
再次查詢gp狀態：
$ gpstate -e
[INFO]:-Starting gpstate with args: -e
[INFO]:-local Greenplum Version: 'postgres (Greenplum Database) 4.2.2.4 build 1 Community Edition'
[INFO]:-master Greenplum Version: 'PostgreSQL 8.2.15 (Greenplum Database 4.2.2.4 build 1 Community Edition) on x86_64-unknown-linux-gnu, compiled by GCC gcc (GCC) 4.4.2 compiled on Oct 17 2012 11:52:28'
[INFO]:-Obtaining Segment details from master...
[INFO]:-Gathering data from segments...
[INFO]:-----------------------------------------------------
[INFO]:-Segment Mirroring Status Report
[INFO]:-----------------------------------------------------
[INFO]:-All segments are running normally
查詢prefer role和current role：
[INFO]:-----------------------------------------------------
[INFO]:--Master Configuration & Status
[INFO]:-----------------------------------------------------
[INFO]:- Master host = gp01
[INFO]:- Master postgres process ID = 12423
[INFO]:- Master data directory = /gpmaster/gpseg-1
[INFO]:- Master port = 5432
[INFO]:- Master current role= dispatch
[INFO]:- Greenplum initsystem version = 4.2.2.4 build 1 Community Edition
[INFO]:- Greenplum current version= PostgreSQL 8.2.15 (Greenplum Database 4.2.2.4 build 1 Community Edition) on x86_64-unknown-linux-gnu, compiled by GCC gcc (GCC) 4.4.2 compiled on Oct 17 2012 11:52:28
[INFO]:- Postgres version = 8.2.15
[INFO]:- Master standby = No master standby configured
[INFO]:-----------------------------------------------------
[INFO]:-Segment Instance Status Report
[INFO]:-----------------------------------------------------
[INFO]:- Segment Info
[INFO]:-Hostname = gp01
[INFO]:-Address = gp01
[INFO]:-Datadir = /gpsegment1/gpseg0
[INFO]:-Port= 40000
[INFO]:- Mirroring Info
[INFO]:-Current role = Primary
[INFO]:-Preferred role = Primary
[INFO]:-Mirror status = Synchronized
[INFO]:- Status
[INFO]:-PID = 12390
[INFO]:-Configuration reports status as = Up
[INFO]:-Database status = Up
[INFO]:-----------------------------------------------------
[INFO]:- Segment Info
[INFO]:-Hostname = gp02
[INFO]:-Address = gp02
[INFO]:-Datadir = /gpsegment2/gpseg0
[INFO]:-Port= 50000
[INFO]:- Mirroring Info
[INFO]:-Current role = Mirror
[INFO]:-Preferred role = Mirror
[INFO]:-Mirror status = Synchronized
[INFO]:- Status
[INFO]:-PID = 10689
[INFO]:-Configuration reports status as = Up
[INFO]:-Segment status = Up
[INFO]:-----------------------------------------------------
[INFO]:- Segment Info
[INFO]:-Hostname = gp02
[INFO]:-Address = gp02
[INFO]:-Datadir = /gpsegment1/gpseg1
[INFO]:-Port= 40000
[INFO]:- Mirroring Info
[INFO]:-Current role = Primary
[INFO]:-Preferred role = Primary
[INFO]:-Mirror status = Synchronized
[INFO]:- Status
[INFO]:-PID = 10695
[INFO]:-Configuration reports status as = Up
[INFO]:-Database status = Up
[INFO]:-----------------------------------------------------
[INFO]:- Segment Info
[INFO]:-Hostname = gp03
[INFO]:-Address = gp03
[INFO]:-Datadir = /gpsegment2/gpseg1
[INFO]:-Port= 50000
[INFO]:- Mirroring Info
[INFO]:-Current role = Mirror
[INFO]:-Preferred role = Mirror
[INFO]:-Mirror status = Synchronized
[INFO]:- Status
[INFO]:-PID = 10946
[INFO]:-Configuration reports status as = Up
[INFO]:-Segment status = Up
[INFO]:-----------------------------------------------------
[INFO]:- Segment Info
[INFO]:-Hostname = gp03
[INFO]:-Address = gp03
[INFO]:-Datadir = /gpsegment1/gpseg2
[INFO]:-Port= 40000
[INFO]:- Mirroring Info
[INFO]:-Current role = Primary
[INFO]:-Preferred role = Primary
[INFO]:-Mirror status = Synchronized
[INFO]:- Status
[INFO]:-PID = 5907
[INFO]:-Configuration reports status as = Up
[INFO]:-Database status = Up
[INFO]:-----------------------------------------------------
[INFO]:- Segment Info
[INFO]:-Hostname = gp01
[INFO]:-Address = gp01
[INFO]:-Datadir = /gpsegment2/gpseg2
[INFO]:-Port= 50000
[INFO]:- Mirroring Info
[INFO]:-Current role = Mirror
[INFO]:-Preferred role = Mirror
[INFO]:-Mirror status = Synchronized
[INFO]:- Status
[INFO]:-PID = 14134
[INFO]:-Configuration reports status as = Up
[INFO]:-Segment status = Up
current role切換回prefered role。