概述
这两个概念容易混淆，简单来说：
- snapshot copy PDB是12c新增的功能，依赖于底层存储快照
- PDB snapshot是18c的功能，不必依赖底层存储快照
- 他们的区别，也可以认为是snapshot和snapshot copy的区别
snapshot copy PDB
按照文档的的说法：
A snapshot copy PDB does not include a complete copy of the source data files. Rather, Oracle Database creates a storage-level snapshot of the underlying file system, and then creates the clone PDB from the snapshot.
也就是说：
- snapshot copy PDB要求底层文件系统支持快照技术，例如ACFS
- snapshot copy PDB包含的是稀疏文件，只有修改的数据库才会被存储
示例如下，但由于我底层文件系统不支持存储快照，因此失败：
sql
SQL> create pluggable database orclpdb2_snapshot from orclpdb2 snapshot copy;
create pluggable database orclpdb2_snapshot from orclpdb2 snapshot copy
*
ERROR at line 1:
ORA-65169: error encountered while attempting to copy file
/u01/app/oracle/oradata/ORCL2/orclpdb2/users01.dbf
ORA-17525: Database clone using storage snapshot not supported on file
/u01/app/oracle/oradata/ORCL2/orclpdb2/users01.dbf
SQL> !oerr ora 17525
17525, 00000, "Database clone using storage snapshot not supported on file %s"
// *Cause: Cloning a database using storage level snapshot is not supported
// on the underlying storage.
// *Action: Use storage product that has snapshot based cloning support
// for Oracle.PDB snapshot
PDB snapshot是支持PDB Snapshot Carousel的底层技术。
按照文档的说法：
A PDB snapshot is a point-in-time copy of a PDB. The source PDB can be open read-only or read/write while the snapshot is created. You can create snapshots manually using the SNAPSHOT clause of CREATE PLUGGABLE DATABASE (or ALTER PLUGGABLE DATABASE), or automatically using the EVERY interval clause. If the storage system supports sparse clones, then the preceding command creates a sparse copy. Otherwise, the command creates a full copy.
也就是说，创建PDB snapshot时：
- 源PDB可以是只读或读写状态，
- 可以手工或自动创建
- 生成的PDB可以是完全拷贝或稀疏拷贝，后者需要底层存储系统支持。
PDB snapshot主要有2个作用：
- 克隆PDB以用于开发测试
- 用于时间点恢复
PDB snapshot是Exadata的特性，如果不是Exadata，则会报错：
sql
SQL> ALTER PLUGGABLE DATABASE SNAPSHOT;
ALTER PLUGGABLE DATABASE SNAPSHOT
*
ERROR at line 1:
ORA-12754: Feature PDB SNAPSHOT CAROUSEL is disabled due to missing capability如果确实不是Exadata，也可以设置隐含参数以模拟Exadata。
sql
connect / as sysdba
alter system set "_exadata_feature_on"=true scope=spfile;
shutdown immediate;
startup;如果SNAPSHOT_MODE不是NONE，则可以手工创建PDB snapshot。
sql
SET LINESIZE 150 TAB OFF
COLUMN pdb_name FORMAT A10
COLUMN snapshot_mode FORMAT A15
SELECT p.con_id,
p.pdb_name,
p.snapshot_mode,
p.snapshot_interval
FROM cdb_pdbs p
ORDER BY 1;
CON_ID PDB_NAME SNAPSHOT_MODE SNAPSHOT_INTERVAL
---------- ---------- --------------- -----------------
3 ORCLPDB1 MANUAL手工创建PDB snapshot：
sql
SQL> ALTER PLUGGABLE DATABASE SNAPSHOT;
Pluggable database altered.查询PDB snapshot：
sql
SET LINESIZE 150
COL CON_NAME FORMAT a9
COL ID FORMAT 99
COL SNAPSHOT_NAME FORMAT a17
COL SNAP_SCN FORMAT 9999999
COL FULL_SNAPSHOT_PATH FORMAT a61
SELECT CON_ID AS ID, CON_NAME, SNAPSHOT_NAME,
SNAPSHOT_SCN AS snap_scn, FULL_SNAPSHOT_PATH
FROM DBA_PDB_SNAPSHOTS
ORDER BY SNAP_SCN;
ID CON_NAME SNAPSHOT_NAME SNAP_SCN FULL_SNAPSHOT_PATH
--- --------- ----------------- -------- -------------------------------------------------------------
3 ORCLPDB2 SNAP_1677908289_1 3229763 /u01/app/oracle/oradata/ORCL2/orclpdb2/snap_1677908289_322976
154256623 3.pdb
sql
SET LINESIZE 200
SET PAGESIZE 50000
COL ID FORMAT 99
COL CON_NAME FORMAT a10
COL SNAPSHOT_NAME FORMAT a26
COL SNAPSHOT_SCN FORMAT 9999999999
COL FULL_SNAPSHOT_PATH FORMAT a70
SELECT CON_ID AS ID, CON_NAME, SNAPSHOT_NAME,
SNAPSHOT_SCN, FULL_SNAPSHOT_PATH
FROM DBA_PDB_SNAPSHOTS;
ID CON_NAME SNAPSHOT_NAME SNAPSHOT_SCN FULL_SNAPSHOT_PATH
--- ---------- -------------------------- ------------- ----------------------------------------------------------------------
3 ORCLPDB2 SNAP_1677908289_1154256623 3229763 /u01/app/oracle/oradata/ORCL2/orclpdb2/snap_1677908289_3229763.pdbpdb的后缀说明这时一个完全拷贝，而非稀疏拷贝。
下面的输出为0，也可以说明是完全拷贝。因为只有是稀疏拷贝时，DBA_PDB_SNAPSHOTFILE表才有行：
sql
SQL> SELECT COUNT(*) FROM DBA_PDB_SNAPSHOTFILE;
COUNT(*)
----------
0删除PDB snapshot：
sql
SQL> ALTER PLUGGABLE DATABASE DROP SNAPSHOT SNAP_1677908289_1154256623;
Pluggable database altered.