Solution to unreleased disk space after linux Files are deleted

Today, an ORACLE tablespace file is deleted, indicating that it is deleted from ORACLE according to the standard, but the physical file is sent to rm-rf,

The result shows that the directory size is reduced, but when you look at the partition information, the space still occupies so much space.

Later, I checked the solution on the Internet as follows:

Execute the command: lsof | grep the file name of the operating system you want to delete

The following information is displayed:

Oracle 12713 oracle 12u REG 22260228096, 1 19234817/data/oracle/.../pcds01.dbf (deleted)
Oracle 14899 oracle 22uW REG 253,1 22260228096 19234817/data/oracle/.../pcds01.dbf (deleted)
Oracle 14901 oracle 25u REG 253,1 22260228096 19234817/data/oracle/.../pcds01.dbf (deleted)
Oracle 14903 oracle 23u REG 253,1 22260228096 19234817/data/oracle/.../pcds01.dbf (deleted)
Oracle 14905 oracle 19u REG 22260228096, 1 19234817/data/oracle/.../pcds01.dbf (deleted)

The operating system is cleaning up the deletion,

Kill all (delete) States. Note that some processes of the oracle database should not be kill, and the space will be released after three minutes. use of the attached lsof command:

Actually, we only need to use ps ax | grep 12713

Check all processes and find

[Root @ beta1 beta1] # ps ax | grep 14899
14899? Ss 4: 55 ora_dbw0_pcds
16392 pts/6 S + grep 14899

In this case, the process then: kill-9 14899

Run df-h to check whether the space is released,

If ps ax | grep lists:

12713? Ss 0: 55 oraclepcds (DESCRIPTION = (LOCAL = YES) (ADDRESS = (PROTOCOL = beq )))

It cannot be deleted. This is an ORACLE process.