南大通用数据库-Gbase-8a-学习-17-Gbase8a集群版本升级

一、测试环境

名称	值
cpu	Intel® Core™ i5-1035G1 CPU @ 1.00GHz
操作系统	CentOS Linux release 7.9.2009 (Core)
内存	4G
逻辑核数	3
Gbase-8a节点-IP	192.168.142.10
Gbase-8a数据库旧版本	8.6.2.43-R33.132743
Gbase-8a数据库新版本	8.6.2-R43.34.27468a27

二、旧版本环境

1、集群状态

[gbase@localhost gcinstall]$ gcadmin showdistributionDistribution ID: 1 | State: new | Total segment num: 1Primary Segment Node IP                           Segment ID         Duplicate Segment node IP
========================================================================================================================
|    192.168.142.12                              |       1          |                                                  |
========================================================================================================================

2、测试数据

gbase> create database czg;
Query OK, 1 row affected (Elapsed: 00:00:00.01)gbase> use czg;
Query OK, 0 rows affected (Elapsed: 00:00:00.00)gbase> create table sun (a int,b double,c varchar(100),d text,e blob,f longblob,g date,h timestamp);
Query OK, 0 rows affected (Elapsed: 00:00:00.16)gbase> insert into sun values(1,1.1,'czg','快乐的小天使','qwertasdsdfzxczxxv','gregergjsfishfuieehfuiew','1995-09-18','2022-08-03 09:24:00.000');
Query OK, 1 row affected (Elapsed: 00:00:00.14)gbase> insert into sun values(1,1.1,'czg','快乐的小天使','qwertasdsdfzxczxxv','gregergjsfishfuieehfuiew','1995-09-18','2022-08-03 09:24:00.000');
Query OK, 1 row affected (Elapsed: 00:00:00.14)gbase> insert into sun values(1,1.1,'czg','快乐的小天使','qwertasdsdfzxczxxv','gregergjsfishfuieehfuiew','1995-09-18','2022-08-03 09:24:00.000');
Query OK, 1 row affected (Elapsed: 00:00:00.02)gbase> insert into sun select * from sun;
Query OK, 3 rows affected (Elapsed: 00:00:00.21)
Records: 3  Duplicates: 0  Warnings: 0gbase> select * from sun;
+------+------+------+--------------------+--------------------+--------------------------+------------+---------------------+
| a    | b    | c    | d                  | e                  | f                        | g          | h                   |
+------+------+------+--------------------+--------------------+--------------------------+------------+---------------------+
|    1 |  1.1 | czg  | 快乐的小天使       | qwertasdsdfzxczxxv | gregergjsfishfuieehfuiew | 1995-09-18 | 2022-08-03 09:24:00 |
|    1 |  1.1 | czg  | 快乐的小天使       | qwertasdsdfzxczxxv | gregergjsfishfuieehfuiew | 1995-09-18 | 2022-08-03 09:24:00 |
|    1 |  1.1 | czg  | 快乐的小天使       | qwertasdsdfzxczxxv | gregergjsfishfuieehfuiew | 1995-09-18 | 2022-08-03 09:24:00 |
|    1 |  1.1 | czg  | 快乐的小天使       | qwertasdsdfzxczxxv | gregergjsfishfuieehfuiew | 1995-09-18 | 2022-08-03 09:24:00 |
|    1 |  1.1 | czg  | 快乐的小天使       | qwertasdsdfzxczxxv | gregergjsfishfuieehfuiew | 1995-09-18 | 2022-08-03 09:24:00 |
|    1 |  1.1 | czg  | 快乐的小天使       | qwertasdsdfzxczxxv | gregergjsfishfuieehfuiew | 1995-09-18 | 2022-08-03 09:24:00 |
+------+------+------+--------------------+--------------------+--------------------------+------------+---------------------+
6 rows in set (Elapsed: 00:00:00.01)

3、版本查看

gbase> select version();
+---------------------+
| version()           |
+---------------------+
| 8.6.2.43-R33.132743 |
+---------------------+
1 row in set (Elapsed: 00:00:00.00)

三、升级步骤

cexec命令的安装步骤看之前的写的文章里有，文章连接：南大通用数据库-Gbase-8a-学习-12-Gbase8a常用运维命令（持续更新哈）

1、确定业务是否停止

确保没有运行的Sql。

[gbase@localhost gcinstall]$ cexecs data: 'gncli -uroot -e "show processlist"|grep -v Sleep'
************************ data  ************************
--------- 192.168.142.12---------
Warning: Permanently added '192.168.142.12' (ECDSA) to the list of known hosts.
Id      User    Host    db      Command Time    State   Info
36      root    localhost       NULL    Query   0       NULL    show processlist

2、定时任务检查

（1）查看定时任务

[root@localhost opt]# cexecs data: 'cat /var/spool/cron/gbase'
************************ data  ************************
--------- 192.168.142.12---------
Warning: Permanently added '192.168.142.12' (ECDSA) to the list of known hosts.
0 */2 * * * /opt/gnode/server/bin/gc_syncpacklog.sh /opt

（2）注释定时任务

如果升级时间段内有定时任务需注释。

[root@localhost opt]# cexecs data: 'sed -i "s/^[^#]/#&/" /var/spool/cron/gbase'
************************ data  ************************
--------- 192.168.142.12---------
Warning: Permanently added '192.168.142.12' (ECDSA) to the list of known hosts.[root@localhost opt]# cexecs data: 'cat /var/spool/cron/gbase'
************************ data  ************************
--------- 192.168.142.12---------
Warning: Permanently added '192.168.142.12' (ECDSA) to the list of known hosts.
#0 */2 * * * /opt/gnode/server/bin/gc_syncpacklog.sh /opt

（3）解除定时任务注释（升级结束后执行）

[root@localhost opt]# cexecs data: 'sed -i "s/^[#]//" /var/spool/cron/gbase'
************************ data  ************************
--------- 192.168.142.12---------
Warning: Permanently added '192.168.142.12' (ECDSA) to the list of known hosts.[root@localhost opt]# cexecs data: 'cat /var/spool/cron/gbase'
************************ data  ************************
--------- 192.168.142.12---------
Warning: Permanently added '192.168.142.12' (ECDSA) to the list of known hosts.
0 */2 * * * /opt/gnode/server/bin/gc_syncpacklog.sh /opt

3、检查集群状态

如果有节点异常，先处理再升级。

[root@localhost opt]# gcadmin
CLUSTER STATE:  ACTIVE
CLUSTER MODE:   NORMAL=====================================================================
|               GBASE COORDINATOR CLUSTER INFORMATION               |
=====================================================================
|   NodeName   |       IpAddress       |gcware |gcluster |DataState |
---------------------------------------------------------------------
| coordinator1 |    192.168.142.12     | OPEN  |  OPEN   |    0     |
---------------------------------------------------------------------
=================================================================
|                GBASE DATA CLUSTER INFORMATION                 |
=================================================================
|NodeName |       IpAddress       |gnode |syncserver |DataState |
-----------------------------------------------------------------
|  node1  |    192.168.142.12     | OPEN |   OPEN    |    0     |
-----------------------------------------------------------------

4、检查EVENT

（1）查看命令

[root@localhost opt]# gcadmin showdmlevent
Event count:0[root@localhost opt]# gcadmin showddlevent
Event count:0[root@localhost opt]# gcadmin showdmlstorageevent
Event count:0

（2）日志排查

如果上述结果有不为0的，需查看日志中是否有报错日志，根据报错解决相应问题。

[gbase@localhost gcinstall]$ tail -f /opt/gcluster/log/gcluster/gc_recover.log
2022-08-23 09:39:09.242 [INFO ] <RECOVER-INFO>: Recover Thread(0x7fccba42c700) is been built success
2022-08-23 09:39:09.242 [INFO ] <RECOVER-INFO>: Recover Thread(0x7fccb9c2b700) is been built success
2022-08-23 09:39:09.242 [INFO ] <RECOVER-INFO>: Recover Thread(0x7fccb942a700) is been built success
2022-08-23 09:39:09.242 [INFO ] <RECOVER-INFO>: Recover Thread(0x7fccb8c29700) is been built success
2022-08-23 09:39:09.243 [INFO ] <RECOVER-INFO>: Recover Thread(0x7fccb8428700) is been built success
2022-08-23 09:39:09.243 [INFO ] <RECOVER-INFO>: Recover Thread(0x7fccb7c27700) is been built success
2022-08-23 09:39:09.243 [INFO ] <RECOVER-INFO>: Recover Thread(0x7fccb7426700) is been built success
2022-08-23 09:39:09.243 [INFO ] <RECOVER-INFO>: Recover Thread(0x7fccb6c25700) is been built success
2022-08-23 09:39:09.243 [INFO ] <RECOVER-INFO>: Recover Thread(0x7fccb6424700) is been built success
2022-08-23 09:39:09.243 [INFO ] <GCWare>: master node is: 210675904 192.168.142.12

5、集群版本检查（升级前后都需要做）

升级前后都需要记录，分别升级后做版本对比。

（1）集群服务版本

[gbase@localhost gcinstall]$ cexecs coor: "/opt/gcluster/server/bin/gclusterd -V"
************************ coor  ************************
--------- 192.168.142.12---------
Warning: Permanently added '192.168.142.12' (ECDSA) to the list of known hosts.
/opt/gcluster/server/bin/gclusterd ver 8.6.2.43-R33.132743 for unknown-linux-gnu on x86_64

（2）数据节点服务版本

[gbase@localhost gcinstall]$ cexecs data: "/opt/gnode/server/bin/gbased -V"
************************ data  ************************
--------- 192.168.142.12---------
Warning: Permanently added '192.168.142.12' (ECDSA) to the list of known hosts.
/opt/gnode/server/bin/gbased ver 8.6.2.43-R33.132743 for unknown-linux-gnu on x86_64

（3）gcadmin版本

[gbase@localhost gcinstall]$ cexecs "gcadmin -V"
************************ coor  ************************
--------- 192.168.142.12---------
Warning: Permanently added '192.168.142.12' (ECDSA) to the list of known hosts.
gcadmin 1.2.9 build 129123

（4）其他组件版本

gbase>  SHOW VARIABLES LIKE '%VERSION';
+----------------------------+---------------------+
| Variable_name              | Value               |
+----------------------------+---------------------+
| gbase_kafka_broker_version |                     |
| gbase_show_version         | 1                   |
| gcluster_hash_version      | 1                   |
| protocol_version           | 10                  |
| version                    | 8.6.2.43-R33.132743 |
+----------------------------+---------------------+
5 rows in set (Elapsed: 00:00:00.01)

6、停止集群服务

[root@localhost opt]# cexec all: "service gcware stop"
************************* all *************************
--------- 192.168.142.12---------
Warning: Permanently added '192.168.142.12' (ECDSA) to the list of known hosts.
Stopping GCMonit success!
Signaling GCRECOVER (gcrecover) to terminate: [  确定  ]
Waiting for gcrecover services to unload:...[  确定  ]
Signaling GCSYNC (gc_sync_server) to terminate: [  确定  ]
Waiting for gc_sync_server services to unload:[  确定  ]
Signaling GCLUSTERD  to terminate: [  确定  ]
Waiting for gclusterd services to unload:........[  确定  ]
Signaling GBASED  to terminate: [  确定  ]
Waiting for gbased services to unload:..[  确定  ]
Signaling GCWARE (gcware) to terminate: [  确定  ]
Waiting for gcware services to unload:..[  确定  ]

7、获取现有SCN和TABLEID

记录现有的SCN号和TABLEID号，用于升级意外失败的手工恢复。

[gbase@localhost gcinstall]$ python
Python 2.7.5 (default, Jun 28 2022, 15:30:04)
[GCC 4.8.5 20150623 (Red Hat 4.8.5-44)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import gcware
>>> gcware.getscn()
16
>>> gcware.gettableid()
8

8、集群配置文件备份

（1）集群层配置文件备份

[root@czg2 ~]# cexec coor: 'cp -a /opt/gcluster/ /home/gbase/bak/'

（2）数据层配置文件备份

[root@czg2 ~]# cexec coor: 'cp -a /opt/gnode/ /home/gbase/bak/'

（3）其他配置文件备份

[root@czg2 ~]# cexec coor: "cp -a /var/lib/gcware/ /home/gbase/bak/ "[root@czg2 ~]# cexec coor: "cp -a /etc/corosync /home/gbase/bak/ "

9、解压新版本

[root@localhost opt]# cd /opt/pkg/Gbase8a-34/[root@localhost Gbase8a-34]# ll
总用量 88300
-rw-rw-rw- 1 root root 90416620 8月  23 09:20 GBase8a_MPP_Cluster-NoLicense-8.6.2_build43-R34-2-redhat7-x86_64.tar.bz2[root@localhost Gbase8a-34]# tar -xvf GBase8a_MPP_Cluster-NoLicense-8.6.2_build43-R34-2-redhat7-x86_64.tar.bz2

10、修改配置文件

[root@localhost gcinstall]# cat demo.options
installPrefix= /opt
coordinateHost = 192.168.142.12
coordinateHostNodeID = 234,235,237
dataHost = 192.168.142.12
#existCoordinateHost =
#existDataHost =
loginUser= root
loginUserPwd = 'qwer1234'
#loginUserPwdFile = loginUserPwd.json
dbaUser = gbase
dbaGroup = gbase
dbaPwd = 'gbase'
rootPwd = 'qwer1234'
#rootPwdFile = rootPwd.json
dbRootPwd = ''
#mcastAddr = 226.94.1.39
mcastPort = 5493

11、集群升级

（1）升级命令

[gbase@czg2 gcinstall]$ python gcinstall.py --silent=demo.options -U
192.168.142.12          Install gcluster on host 192.168.142.12 successfully.
192.168.142.12          Install gcluster on host 192.168.142.12 successfully.
Starting all gcluster nodes...
Upgrade Successfully.

（2）升级日志

[root@localhost gcinstall]# tail -f gcinstall.log
2022-08-23 11:20:08,756-root-INFO Copying pexpect.py to host 192.168.142.12:/opt_prepare successfully
2022-08-23 11:20:08,757-root-INFO Copying BUILDINFO to host 192.168.142.12:/opt_prepare
2022-08-23 11:20:09,925-root-INFO Copying BUILDINFO to host 192.168.142.12:/opt_prepare successfully
2022-08-23 11:20:09,925-root-INFO Copying bundle.tar.bz2 to host 192.168.142.12:/opt_prepare
2022-08-23 11:20:11,997-root-INFO Copying bundle.tar.bz2 to host 192.168.142.12:/opt_prepare successfully
2022-08-23 11:20:11,998-root-INFO Copying bundle_data.tar.bz2 to host 192.168.142.12:/opt_prepare
2022-08-23 11:20:13,206-root-INFO Copying bundle_data.tar.bz2 to host 192.168.142.12:/opt_prepare successfully
2022-08-23 11:20:13,206-root-INFO Copy data files to host 192.168.142.12 successfully.
2022-08-23 11:20:13,206-root-INFO Upgrade gcluster.
2022-08-23 11:20:13,207-root-INFO send cmd: /usr/bin/python /opt_prepare/InstallTar.py --silent=/opt_prepare/192.168.142.12.options --IsCoordinator --IsData  -U  --ClusterUUID=4aa9025c-2292-11ed-9142-000c29c7ba0f --is_IPV4
2022-08-23 11:30:19,783-root-INFO Install gcluster on host 192.168.142.12 successfully.
2022-08-23 11:30:20,297-root-DEBUG set gc_syncpacklog on host 192.168.142.12
2022-08-23 11:30:21,647-root-INFO crontab -l > conf;sed -i '/gc_syncpacklog.sh/d' conf;sed -i '/^$/d' conf;echo '0 */2 * * * /opt/gnode/server/bin/gc_syncpacklog.sh /opt' >> conf;                crontab conf;rm -rf conf;
2022-08-23 11:30:22,270-root-DEBUG Starting all gcluster nodes...
2022-08-23 11:30:35,841-root-INFO start service successfull on host 192.168.142.12.
2022-08-23 11:30:36,345-root-DEBUG /bin/chown -R gbase:gbase gcChangeInfo.xml
2022-08-23 11:30:36,354-root-DEBUG source /home/gbase/.gbase_profile;gcadmin showcluster f
2022-08-23 11:30:42,437-root-INFO /opt/gcluster/server/bin/gbase -p'***' -uroot -e"select count(uid) from gbase.user where uid <> 0"

12、修改参数

V8.6.2build43R31及之后的版本推荐添加或修改如下参数。
所有管理节点都需要添加这两个参数，后续需重启数据库。

（1）corosync.conf参数

参数名	描述
inner_connect_read_write_timeout	gbase_inner_connect 连接、写、读的超时时间。
	参数值：整数型 1~5，单位：秒，默认值：1，最大值 5。

添加到gcware标签里。

[root@czg2 ~]# cat /etc/corosync/corosync.conf
totem {version: 2secauth: offinterface {member {memberaddr: 192.168.142.12}ringnumber: 0bindnetaddr: 192.168.142.12mcastport: 5493ttl: 1}transport: udputoken: 60000token_retransmits_before_loss_const:80send_join: 80join: 1500consensus:120000
}
logging {fileline: offto_stderr: noto_file: yesto_syslog: nologfile: /var/log/corosync.logcorosync_system_log: /opt/gcluster/log/gcluster/corosync_system.logdebug: offtimestamp: onlogger {ident: AMFdebug: offtags: enter|leave|trace1|trace2|trace3|trace4|trace6}
}gcware {persistent_interval: 5check_interval: 30whole_check_interval_num: 20cfg_connect_timeout: 5000gcluster_port: 5258gnode_port: 5050syncserver_port: 5288node_ssh_port: 22singlecastport: 5678check_coordinator_thread_num: 1check_dataserver_thread_num: 10check_tcp_only: 1inner_connect_read_write_timeout: 5coordinator {member {memberaddr: 192.168.142.12}}
}

（2）gcluster_shrink_to_rebalance

参数名	描述
gcluster_shrink_to_rebalance	rebalance 实现 shrink space 功能和原 shrink space 功能共存，用参数控制：gcluster_shrink_to_rebalance
	默认值 0；取值 1：打开 shrink to rebalance 功能；取值 0：关闭shrink to rebalance 功能。

在gbase_8a_gcluster.cnf中的模块[gbased]中添加gcluster_shrink_to_rebalance参数。

[root@czg2 ~]# cat /opt/gcluster/config/gbase_8a_gcluster.cnf
[client]
port=5258
socket=/tmp/gcluster_5258.sock
connect_timeout=43200
#default_character_set=gbk[gbased]
basedir = /opt/gcluster/server
datadir = /opt/gcluster/userdata/gcluster
socket=/tmp/gcluster_5258.sock
pid_file = /opt/gcluster/log/gcluster/gclusterd.pid#default_character_set=gbklog_error
port=5258
gcluster_gnode_port=5050
core_filedefault_storage_engine=express
default_time_zone='+8:00'
_gbase_query_path=0skip_name_resolve
query_cache_type = 0
query_cache_size = 0M
event_scheduler= 1thread_stack = 1048576sql_mode=PAD_CHAR_TO_FULL_LENGTH,PIPES_AS_CONCAT,ANSI_QUOTES,IGNORE_SPACE,NO_AUTO_CREATE_USER,NO_AUTO_VALUE_ON_ZERO,NO_ENGINE_SUBSTITUTION,STRICT_ALL_TABLES,NO_ZERO_IN_DATE,NO_ZERO_DATE,ONLY_FULL_GROUP_BY
lower_case_table_names=1max_connections = 10000
max_connect_errors=1000000
max_allowed_packet = 64M
net_write_timeout = 1000000
net_read_timeout = 1000000
connect_timeout = 1000000
interactive_timeout  =1000000
wait_timeout = 1000000
open_files_limit = 65535gbase_express_log = 1gcluster_connect_net_read_timeout  =  1000000
gcluster_connect_net_write_timeout =  1000000
gcluster_connect_timeout = 1000000
gcluster_wait_query_cancel_timeout = 200
gcluster_reconn_times = 3
gcluster_async_connect_timeout = 120gcluster_use_special_insert_method = 1
gcluster_use_special_materialized_table = 1
gcluster_special_insert_method_comment=tempgcluster_use_new_threadpool = 1
gcluster_max_thread_in_pool = 600  gcluster_use_conn_pool = 1
gcluster_max_conn_in_pool = 300
gcluster_conn_ping_expire = 0gcluster_local_max_conn_in_pool = 50gcluster_dynamic_cluster_node_status = 1
gcluster_lock_level = 2
gcluster_temp_table_engine='express nolock'
gcluster_dml_ddl_proxy_switch = 0
gcluster_adjust_nodes_before_redist = 0gcluster_starschema_optimize = 0
gcluster_starschema_join_estimate_optimize = 1
gcluster_hash_redistribute_groupby_optimize = 1
gcluster_hash_redistribute_join_optimize = 2
gcluster_crossjoin_use_hash_distribution = 1
gcluster_insertselect_use_values_optimize = 0
gcluster_union_optimize = 1
gcluster_count_optimize = 1
gcluster_insert_singlegrouppart_optimize = 0gcluster_empty_result_set_optimize = 0gcluster_special_correlated_optimize = 1
gcluster_support_hash_redist_combiner = 0
gcluster_order_by_limit_offset_optimize = 0gcluster_ha_event_monitor = 1
gcluster_ha_node_left_event_delay = 120000gcluster_sql_statistics = 0
gcluster_use_new_decimal = 1
gcluster_query_retry = 1
gcluster_insert_optimize_flag = 95gcluster_serial_exec_query = 0
#gcluster_special_correlated_optimize = 1
#gcluster_ddl_parallel_execute =1gbase_compression_str_method=5
gbase_compression_num_method=5back_log = 65535gcluster_hash_version = 1_gbase_transaction_disable = 1#const express can calculated before query
gcluster_prepare_const_express = 0
#disallow non printable char in char and varchar.
up_disallow_non_printable_char = 0gcluster_shrink_to_rebalance = 1[gbasedump]
max_allowed_packet = 64M[gbase]
max_allowed_packet = 64M
no_auto_rehash

13、重启集群

[root@czg2 ~]# cexec all: "service gcware restart"
************************* all *************************
--------- 192.168.142.12---------
Warning: Permanently added '192.168.142.12' (ECDSA) to the list of known hosts.
Stopping GCMonit success!
Signaling GCRECOVER (gcrecover) to terminate: [  确定  ]
Waiting for gcrecover services to unload:..[  确定  ]
Signaling GCSYNC (gc_sync_server) to terminate: [  确定  ]
Waiting for gc_sync_server services to unload:[  确定  ]
Signaling GCLUSTERD  to terminate: [  确定  ]
Waiting for gclusterd services to unload:.......[  确定  ]
Signaling GBASED  to terminate: [  确定  ]
Waiting for gbased services to unload:.....[  确定  ]
Signaling GCWARE (gcware) to terminate: [  确定  ]
Waiting for gcware services to unload:.[  确定  ]
Starting GCWARE (gcwexec): [  确定  ]
Starting GBASED    : [  确定  ]
Starting GCSYNC : [  确定  ]
Starting GCLUSTERD :   [  确定  ]
Starting GCRECOVER : [  确定  ]
Starting GCMonit success!

14、验证升级情况

重复：5、集群版本检查（升级前后都需要做）

[gbase@czg2 gcinstall]$ gccliGBase client 8.6.2-R43.34.27468a27. Copyright (c) 2004-2022, GBase.  All Rights Reserved.gbase> select version();
+-----------------------+
| version()             |
+-----------------------+
| 8.6.2-R43.34.27468a27 |
+-----------------------+
1 row in set (Elapsed: 00:00:00.00)

集群状态检查：

[gbase@czg2 gcinstall]$ gcadmin
CLUSTER STATE:  ACTIVE
CLUSTER MODE:   NORMAL=====================================================================
|               GBASE COORDINATOR CLUSTER INFORMATION               |
=====================================================================
|   NodeName   |       IpAddress       |gcware |gcluster |DataState |
---------------------------------------------------------------------
| coordinator1 |    192.168.142.12     | OPEN  |  OPEN   |    0     |
---------------------------------------------------------------------
=================================================================
|                GBASE DATA CLUSTER INFORMATION                 |
=================================================================
|NodeName |       IpAddress       |gnode |syncserver |DataState |
-----------------------------------------------------------------
|  node1  |    192.168.142.12     | OPEN |   OPEN    |    0     |
-----------------------------------------------------------------

15、升级失败回退

（1）方法一

先杀掉gbase用户下的残留进程。

[root@czg2 gcinstall]# ./Restore.py --help
Usage: Restore.py [options]Options:-h, --help            show this help message and exit-a                    do not prompt the user for confirmation--backupFile=BACKUPFILEbackup file.--silent=SILENTCONFIGuse the supplied properties file for a 'silent'restore--passwordInputMode=PASSWORDINPUTMODEThis process get password method, default:file. mode["file","pwdsame","pwddiff"]file:    get from config file[password withoutencrypted].                        pwdsame: getpassword from user input when all node password same.pwddiff: get password from user input for nodes one byone.

[gbase@localhost gcinstall]$ ./Restore.py --backupFile=/home/gbase/gcluster_backup_xxxx.tar.bz2 --silent=demo.options

gcluster_backup_xxxx.tar.bz2是执行升级脚本后，自动生成的backup 文件。最后出现Success to RestoreLocal的字样，表示回退成果。

（2）方法二

先停服务用新版本unInstall.py卸掉新版本，再用老版本gcinstall.py安装，需要和原来路径一致。再停服务，把 6、集群配置文件备份中的文件移动回原位置之前，最好把原位置的文件先删除，最后再启动服务即可。