[
https://issues.apache.org/jira/browse/HAMA-436?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Thomas Jungblut updated HAMA-436:
---------------------------------
Attachment: HAMA-436_1.patch
I added some more information to the ClusterStatus.
The Webinterface now contains (like Hadoop) the number of current running task
per host, the last seen and the maximum number of tasks.
Actually these datas are not working properly, I don't know why the data in the
BSPMaster is not right.
For example while running bench example, every task should be used, I see the
messages from them, but the web interface just shows a single task.
{noformat}
ubuntu.ubuntu-domain:61002 to ubuntu.ubuntu-domain:61000 : 512
ubuntu.ubuntu-domain:61002 to ubuntu.ubuntu-domain:61001 : 512
{noformat}
As you can see on the ports there are 3 tasks, but the interface just showed
one.
It seems that the ''groomServers'' map in the master isn't updated correctly.
The heartbeat (last_seen) seem to be not used as well.
Additionally I observed that there is a problem with the PiEstimator example in
local mode.
{noformat}
conf.set(MASTER_TASK, hostName + ":" + Constants.DEFAULT_PEER_PORT);
{noformat}
The default port does not exist in local mode.
> Web Interface does not update Superstep Count
> ---------------------------------------------
>
> Key: HAMA-436
> URL: https://issues.apache.org/jira/browse/HAMA-436
> Project: Hama
> Issue Type: Bug
> Components: bsp
> Affects Versions: 0.3.0
> Reporter: Thomas Jungblut
> Assignee: Thomas Jungblut
> Fix For: 0.4.0
>
> Attachments: HAMA-436.patch, HAMA-436_1.patch, hostbug.PNG,
> superstepbug.png
>
>
> Web Interface does not update Superstep Count during executiong of a job. See
> the Screenshot.
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
