如何正确删除数据库分区

从DPF删除一个数据库分区决不是像直接编辑db2nodes.cfg那么简单，正确的做法是使用命令“db2stop drop partitionnum”。

如果你直接编辑db2nodes.cfg删除对应分区那行或者那些行，而恰好有数据库分区组(database partition group)还在使用被删除这个或者这些分区，那么会导致问题。 看下面的例子：

[db2inst1@testhost ~]$ cat sqllib/db2nodes.cfg

0 testhost 0

1 testhost 1

2 testhost 2

3 testhost 3

4 testhost 4

[db2inst1@testhost ~]$ db2 "select TBSP_NAME,DBPARTITIONNUM,TBSP_UTILIZATION_PERCENT from SYSIBMADM.TBSP_UTILIZATION order by TBSP_NAME with ur"

[db2inst1@testhost ~]$ db2 "create table t1(i int not null primary key, c char(10)) in TBS_TEST"

DB20000I  The SQL command completed successfully.

[db2inst1@testhost ~]$ db2 terminate

DB20000I  The TERMINATE command completed successfully.

[db2inst1@testhost ~]$./insert.sh

[db2inst1@testhost ~]$ db2stop

03/25/2018 18:54:41     1   0   SQL1064N  DB2STOP processing was successful.

03/25/2018 18:54:41     2   0   SQL1064N  DB2STOP processing was successful.

03/25/2018 18:54:42     0   0   SQL1064N  DB2STOP processing was successful.

03/25/2018 18:54:43     3   0   SQL1064N  DB2STOP processing was successful.

03/25/2018 18:54:43     4   0   SQL1064N  DB2STOP processing was successful.

SQL1064N  DB2STOP processing was successful.

[db2inst1@testhost ~]$ ipclean -a

Application ipclean: Removing all IPC resources for db2inst1

[db2inst1@testhost ~]$ db2_kill

rah: primary monitoring process for db2nkill is 61762

testhost: db2nkill: DB2 member 0 with PID 56900 does not exist.

testhost: db2nkill: DB2 member 1 with PID 57016 does not exist.

testhost: db2nkill: DB2 member 2 with PID 57113 does not exist.

testhost: db2nkill: DB2 member 3 with PID 57236 does not exist.

testhost: db2nkill: DB2 member 4 with PID 57269 does not exist.

testhost: db2nkill [] completed ok

[db2inst1@testhost ~]$ db2start

03/25/2018 19:06:23     3   0   SQL1063N  DB2START processing was successful.

03/25/2018 19:06:25     2   0   SQL1063N  DB2START processing was successful.

03/25/2018 19:06:25     1   0   SQL1063N  DB2START processing was successful.

03/25/2018 19:06:26     0   0   SQL1063N  DB2START processing was successful.

SQL1063N  DB2START processing was successful.

[db2inst1@testhost ~]$ db2 connect to testdb

   Database Connection Information

 Database server        = DB2/LINUXX8664 10.5.7

 SQL authorization ID   = DB2INST1

 Local database alias   = TESTDB

[db2inst1@testhost ~]$ db2 create table t2 like t1

DB21034E  The command was processed as an SQL statement because it was not a

valid Command Line Processor command.  During SQL processing it returned:

SQL1034C  The database was damaged, so all applications processing the

database were stopped.  SQLSTATE=58031

可以看到无法创建与t1一样的表。

[db2inst1@testhost ~]$ db2 "select * from t1"

SQL0901N  The SQL statement or command failed because of a database system

error. (Reason "sqlno_node_set_add        [100]:rc(     0)                ".) 

SQLSTATE=58004

x

2018-03-25-19.08.45.376940-420 I2972162E756          LEVEL: Info

PID     : 64558                TID : 140661319788288 PROC : db2sysc 0

INSTANCE: db2inst1             NODE : 000            DB   : TESTDB

APPHDL  : 0-62                 APPID: *N0.db2inst1.180326020755

AUTHID  : DB2INST1             HOSTNAME: testhost

EDUID   : 64                   EDUNAME: db2agent (TESTDB) 0

FUNCTION: DB2 UDB, access plan manager, sqlra_dumpEventsForType, probe:3317

DATA #1 : <preformatted>

Dumped APM/RDS event recorder information to file "/home/db2inst1/sqllib/db2dump/FODC_AppErr_2018-03-25-19.08.43.975256_64558_64_000/TESTDB.20180325190845.000.SQLRA.sqlraLOW.events.bin".

If IBM service is contacted to help resolve a recovery problem, this file may be requested.

2018-03-25-19.08.45.378378-420 I2972919E756          LEVEL: Info

PID     : 64558                TID : 140661319788288 PROC : db2sysc 0

INSTANCE: db2inst1             NODE : 000            DB   : TESTDB

APPHDL  : 0-62                 APPID: *N0.db2inst1.180326020755

AUTHID  : DB2INST1             HOSTNAME: testhost

EDUID   : 64                   EDUNAME: db2agent (TESTDB) 0

FUNCTION: DB2 UDB, access plan manager, sqlra_dumpEventsForType, probe:3317

DATA #1 : <preformatted>

Dumped APM/RDS event recorder information to file "/home/db2inst1/sqllib/db2dump/FODC_AppErr_2018-03-25-19.08.43.975256_64558_64_000/TESTDB.20180325190845.000.SQLRA.sqlraMED.events.bin".

If IBM service is contacted to help resolve a recovery problem, this file may be requested.

2018-03-25-19.08.45.379713-420 I2973676E757          LEVEL: Info

PID     : 64558                TID : 140661319788288 PROC : db2sysc 0

INSTANCE: db2inst1             NODE : 000            DB   : TESTDB

APPHDL  : 0-62                 APPID: *N0.db2inst1.180326020755

AUTHID  : DB2INST1             HOSTNAME: testhost

EDUID   : 64                   EDUNAME: db2agent (TESTDB) 0

FUNCTION: DB2 UDB, access plan manager, sqlra_dumpEventsForType, probe:3317

DATA #1 : <preformatted>

Dumped APM/RDS event recorder information to file "/home/db2inst1/sqllib/db2dump/FODC_AppErr_2018-03-25-19.08.43.975256_64558_64_000/TESTDB.20180325190845.000.SQLRA.sqlraHIGH.events.bin".

If IBM service is contacted to help resolve a recovery problem, this file may be requested.

2018-03-25-19.08.45.380027-420 I2974434E233       LEVEL: Severe

PID:64558 TID:140661319788288 NODE:000 Title: RDS DBCB

Dump File:/home/db2inst1/sqllib/db2dump/FODC_AppErr_2018-03-25-19.08.43.975256_64558_64_000/64558.64.000.dump.bin

……

重新创建在db2nodes.cfg中编辑恢复节点

下面是正确的方法，使用的是“db2stop drop partitionnum”，当然还有一些其它命令做为辅助。 主要是通过"drop dbpartitionnum verify"去确认当前分区不存在任何数据库的任何分区组(database partition group)上，如果存在，返回SQL6035W，否则返回SQL6034W。 如果返回的是SQL6035W，你则需要使用命令“redistribute database partition group”重分布分区组，去掉对当前分区的使用。 下面是具体过程：

[db2inst1@testhost ~]$ db2 "select * from syscat.dbpartitiongroupdef"

DBPGNAME          DBPARTITIONNUM IN_USE

-----------------         -------------------------- ----------

IBMCATGROUP                    0   Y    

IBMDEFAULTGROUP                0   Y     

IBMDEFAULTGROUP                1   Y    

IBMDEFAULTGROUP                2   Y    

IBMDEFAULTGROUP                3   Y    

IBMDEFAULTGROUP                4   Y    

DATAGROUP                      0   Y    

DATAGROUP                      1   Y    

DATAGROUP                      2   Y    

DATAGROUP                      3   Y     

  10 record(s) selected.

[db2inst1@testhost ~]$ export DB2NODE=4

[db2inst1@testhost ~]$ db2 terminate

DB20000I  The TERMINATE command completed successfully.

[db2inst1@testhost ~]$ db2 drop dbpartitionnum verify

SQL6035W  Database partition "4" is being used by database "TESTDB".

[db2inst1@testhost ~]$ export DB2NODE=0

[db2inst1@testhost ~]$ db2 terminate

[db2inst1@testhost ~]$ db2 connect to testdb

[db2inst1@testhost ~]$ db2 "redistribute database partition group DATAGROUP uniform drop dbpartitionnums (4)"

DB20000I  The REDISTRIBUTE NODEGROUP command completed successfully.

[db2inst1@testhost ~]$ db2 force applications all

DB20000I  The FORCE APPLICATION command completed successfully.

DB21024I  This command is asynchronous and may not be effective immediately.

[db2inst1@testhost ~]$ export DB2NODE=4

[db2inst1@testhost ~]$ db2 terminate

DB20000I  The TERMINATE command completed successfully.

[db2inst1@testhost ~]$ db2 drop dbpartitionnum verify

SQL6035W  Database partition "4" is being used by database "TESTDB".

[db2inst1@testhost ~]$ db2stop drop dbpartitionnum 4

SQL6076W  If you continue, this command will remove all database files for the specified database partition from the current instance. Before continuing, ensure that the specified database partition is not in use.

Do you want to continue ? (y/n)y     

03/25/2018 23:02:06     3   0   SQL1064N  DB2STOP processing was successful.

03/25/2018 23:02:07     1   0   SQL1064N  DB2STOP processing was successful.

03/25/2018 23:02:07     0   0   SQL1064N  DB2STOP processing was successful.

03/25/2018 23:02:07     2   0   SQL1064N  DB2STOP processing was successful.

03/25/2018 23:02:07     4   0   SQL1064N  DB2STOP processing was successful.

03/25/2018 23:02:11     3   0   SQL1063N  DB2START processing was successful.

03/25/2018 23:02:11     1   0   SQL1063N  DB2START processing was successful.

03/25/2018 23:02:11     4   0   SQL1063N  DB2START processing was successful.

03/25/2018 23:02:11     2   0   SQL1063N  DB2START processing was successful.

03/25/2018 23:02:11     0   0   SQL1063N  DB2START processing was successful.

03/25/2018 23:02:14     4   0   SQL6034W  Database partition "4" is not being used by any databases.

03/25/2018 23:02:17     0   0   SQL1064N  DB2STOP processing was successful.

03/25/2018 23:02:18     2   0   SQL1064N  DB2STOP processing was successful.

03/25/2018 23:02:19     1   0   SQL1064N  DB2STOP processing was successful.

03/25/2018 23:02:19     3   0   SQL1064N  DB2STOP processing was successful.

03/25/2018 23:02:19     4   0   SQL1064N  DB2STOP processing was successful.

03/25/2018 23:02:21     4   0   SQL1064N  DB2STOP processing was successful.

SQL1064N  DB2STOP processing was successful.

[db2inst1@testhost ~]$ cat sqllib/db2nodes.cfg

0 testhost 0

1 testhost 1

2 testhost 2

3 testhost 3

[db2inst1@testhost ~]$ db2start

03/25/2018 23:04:32     2   0   SQL1063N  DB2START processing was successful.

03/25/2018 23:04:32     1   0   SQL1063N  DB2START processing was successful.

03/25/2018 23:04:33     3   0   SQL1063N  DB2START processing was successful.

03/25/2018 23:04:34     0   0   SQL1063N  DB2START processing was successful.

SQL1063N  DB2START processing was successful.

[db2inst1@testhost ~]$ db2 connect to testdb

SQL1469N  Instance "db2inst1" (database partition number "0") does not have

node "4" specified in its db2nodes.cfg file.  SQLSTATE=08004

[db2inst1@testhost ~]$ export DB2NODE=0

[db2inst1@testhost ~]$ db2 terminate

DB20000I  The TERMINATE command completed successfully.

[db2inst1@testhost ~]$ db2 create table t2 like t1

DB20000I  The SQL command completed successfully.