big-data/hadoop
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Revert "HADOOP-13946. Document how HDFS updates timestamps in the FS spec; compare with object stores. Contributed by Steve Loughran"

Browse Source 
This reverts commit 451efb08fe.
This commit is contained in:
Mingliang Liu 2017-01-03 14:44:26 -08:00
parent ebdd2e03b7
commit 88731c731a
1 changed files with 0 additions and 33 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

33
hadoop-common-project/hadoop-common/src/site/markdown/filesystem/introduction.md
View File

@ -392,36 +392,3 @@ Object stores with these characteristics, can not be used as a direct replacemen
for HDFS. In terms of this specification, their implementations of the for HDFS. In terms of this specification, their implementations of the
specified operations do not match those required. They are considered supported specified operations do not match those required. They are considered supported
by the Hadoop development community, but not to the same extent as HDFS. by the Hadoop development community, but not to the same extent as HDFS.
#### Timestamps
The HDFS filesystem does not update the modification time while it is being written to.
Specifically
* `FileSystem.create()` creation: a zero-byte file is listed; creation and modification time is
set to the current time as seen on the NameNode.
* Writes to a file via the output stream returned in the `create()` call: the modification
time *does not change*.
* When `OutputStream.close()` is called, all remaining data is written, the file closed and
the NameNode updated with the final size of the file. The modification time is set to
the time the file was closed.
* Opening a file for appends via an `append()` operation does not change the modification
time of the file until the `close()` call is made on the output stream.
* `FileSystem.setTimes()` can be used to explicitly set the time on a file.
* The rarely used operations: `FileSystem.concat()`, `createSnapshot()`, `createSymlink()` and
`truncate()` all update the modification time.
Other filesystems may have different behaviors.
Object stores have a significantly simpler view of time:
* The file only becomes visible at the end of the write operation; this also sets
the creation time of the file.
* The timestamp is likely to be in UTC or the TZ of the object store. If the
client is in a different timezone, the timestamp may be ahead or behind that
of the client.
* A file's modification time is always the same as its creation time.
* The `FileSystem.setTimes()` operation to set file timestamps will generally be ignored.
* If `FileSystem.append()` is supported, the changes and modification time
are likely to only become visible after the output stream is closed.