[
https://issues.apache.org/jira/browse/HAMA-521?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13261448#comment-13261448
]
Edward J. Yoon commented on HAMA-521:
-------------------------------------
{code}
<property>
<name>hama.tmp.dir</name>
<value>/tmp/hama-${user.name}</value>
<description>Temporary directory on the local filesystem.</description>
</property>
<property>
<name>bsp.disk.queue.dir</name>
<value>${hama.tmp.dir}/messages/</value>
<description>Temporary directory on the local message buffer on
disk.</description>
</property>
{code}
But, I can't find messages directory.
{code}
edward@slave:~/workspace/hama-trunk$ ls -al /tmp/hama-edward/
drwxr-xr-x 3 edward edward 4096 2012-04-25 19:36 .
drwxrwxrwt 25 root root 20480 2012-04-25 20:14 ..
drwxr-xr-x 3 edward edward 4096 2012-04-25 19:36 zookeeper
edward@slave:~/workspace/hama-trunk$ ls -al /tmp/message
messageQueue/ messageStorage/
edward@slave:~/workspace/hama-trunk$ ls -al
/tmp/messageQueue/diskqueue/job_1_0001/task_1_0001_000001/
drwxr-xr-x 2 edward edward 4096 2012-04-25 19:17 .
drwxr-xr-x 3 edward edward 4096 2012-04-25 19:17 ..
-rw-r--r-- 1 edward edward 0 2012-04-25 19:17 .0_messages.seq.crc
-rw-r--r-- 1 edward edward 8752 2012-04-25 19:17 .1_messages.seq.crc
-rw-r--r-- 1 edward edward 0 2012-04-25 19:17 .2_messages.seq.crc
-rw-r--r-- 1 edward edward 0 2012-04-25 19:17 .4_messages.seq.crc
-rwxrwxrwx 1 edward edward 0 2012-04-25 19:17 0_messages.seq
-rwxrwxrwx 1 edward edward 1118856 2012-04-25 19:17 1_messages.seq
-rwxrwxrwx 1 edward edward 0 2012-04-25 19:17 2_messages.seq
-rwxrwxrwx 1 edward edward 0 2012-04-25 19:17 4_messages.seq
{code}
> Improve message buffering to save memory
> ----------------------------------------
>
> Key: HAMA-521
> URL: https://issues.apache.org/jira/browse/HAMA-521
> Project: Hama
> Issue Type: Sub-task
> Reporter: Thomas Jungblut
> Assignee: Thomas Jungblut
> Attachments: HAMA-521.patch, HAMA-521_1.patch, HAMA-521_2.patch,
> HAMA-521_3.patch, mytest.patch
>
>
> Suraj and I had a bit of discussion about incoming and outgoing message
> buffering and scalability.
> Currently everything lies on the heap, causing huge amounts of GC and waste
> of memory. We can do better.
> Therefore we need to extract an abstract Messenger class which is directly
> under the interface but over the compressor class.
> It should abstract the use of the queues in the back (currently lot of
> duplicated code) and it should be backed by a sequencefile on local disk.
> Once sync() starts it should return a message iterator for combining and then
> gets put into a message bundle which is send over RPC.
> On the other side we get a bundle and looping over it putting everything into
> the heap making it much larger than it needs to be. Here we can also flush on
> disk because we are just using a queue-like method to the user-side.
> Plus points:
> In case we have enough heap (see our new metric system), we can also
> implement a buffering technology that is not flushing everything to disk.
> Open questions:
> I don't know how much slower the whole system gets, but it would save alot of
> memory. Maybe we should first evaluate if it is really needed.
> In any case, the refactoring of the duplicate code in the messengers is
> needed.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
