Skip to content

Commit

Permalink
HADOOP-11183. Memory-based S3AOutputstream. (Thomas Demoor via stevel)
Browse files Browse the repository at this point in the history
  • Loading branch information
steveloughran committed Mar 4, 2015
1 parent e17e5ba commit 15b7076
Show file tree
Hide file tree
Showing 7 changed files with 570 additions and 17 deletions.
2 changes: 2 additions & 0 deletions hadoop-common-project/hadoop-common/CHANGES.txt
Expand Up @@ -667,6 +667,8 @@ Release 2.7.0 - UNRELEASED
HADOOP-11620. Add support for load balancing across a group of KMS for HA. HADOOP-11620. Add support for load balancing across a group of KMS for HA.
(Arun Suresh via wang) (Arun Suresh via wang)


HADOOP-11183. Memory-based S3AOutputstream. (Thomas Demoor via stevel)

BUG FIXES BUG FIXES


HADOOP-11512. Use getTrimmedStrings when reading serialization keys HADOOP-11512. Use getTrimmedStrings when reading serialization keys
Expand Down
Expand Up @@ -763,13 +763,13 @@ for ldap providers in the same way as above does.
<property> <property>
<name>fs.s3a.connection.establish.timeout</name> <name>fs.s3a.connection.establish.timeout</name>
<value>5000</value> <value>5000</value>
<description>Socket connection setup timeout in seconds.</description> <description>Socket connection setup timeout in milliseconds.</description>
</property> </property>


<property> <property>
<name>fs.s3a.connection.timeout</name> <name>fs.s3a.connection.timeout</name>
<value>50000</value> <value>50000</value>
<description>Socket connection timeout in seconds.</description> <description>Socket connection timeout in milliseconds.</description>
</property> </property>


<property> <property>
Expand Down Expand Up @@ -845,6 +845,22 @@ for ldap providers in the same way as above does.
uploads to.</description> uploads to.</description>
</property> </property>


<property>
<name>fs.s3a.fast.upload</name>
<value>false</value>
<description>Upload directly from memory instead of buffering to
disk first. Memory usage and parallelism can be controlled as up to
fs.s3a.multipart.size memory is consumed for each (part)upload actively
uploading (fs.s3a.threads.max) or queueing (fs.s3a.max.total.tasks)</description>
</property>

<property>
<name>fs.s3a.fast.buffer.size</name>
<value>1048576</value>
<description>Size of initial memory buffer in bytes allocated for an
upload. No effect if fs.s3a.fast.upload is false.</description>
</property>

<property> <property>
<name>fs.s3a.impl</name> <name>fs.s3a.impl</name>
<value>org.apache.hadoop.fs.s3a.S3AFileSystem</value> <value>org.apache.hadoop.fs.s3a.S3AFileSystem</value>
Expand Down
Expand Up @@ -83,6 +83,14 @@ public class Constants {
// comma separated list of directories // comma separated list of directories
public static final String BUFFER_DIR = "fs.s3a.buffer.dir"; public static final String BUFFER_DIR = "fs.s3a.buffer.dir";


// should we upload directly from memory rather than using a file buffer
public static final String FAST_UPLOAD = "fs.s3a.fast.upload";
public static final boolean DEFAULT_FAST_UPLOAD = false;

//initial size of memory buffer for a fast upload
public static final String FAST_BUFFER_SIZE = "fs.s3a.fast.buffer.size";
public static final int DEFAULT_FAST_BUFFER_SIZE = 1048576; //1MB

// private | public-read | public-read-write | authenticated-read | // private | public-read | public-read-write | authenticated-read |
// log-delivery-write | bucket-owner-read | bucket-owner-full-control // log-delivery-write | bucket-owner-read | bucket-owner-full-control
public static final String CANNED_ACL = "fs.s3a.acl.default"; public static final String CANNED_ACL = "fs.s3a.acl.default";
Expand Down

0 comments on commit 15b7076

Please sign in to comment.