[
https://issues.apache.org/jira/browse/SAMZA-90?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13836996#comment-13836996
]
Jakob Homan commented on SAMZA-90:
----------------------------------
bq. (inputStreams.head.getStream,
Won't this just pull in the first stream this TI is responsible for, rather
than all of them? Although thinking about this more, all of them will be quite
a big string for TIs that are consuming lots of streams (say when
co-partitioning).
I was thinking of this more along the lines of a quick way of identifying which
stream task had emitted the log line. All the extra information above might be
useful in a toDetailedString method? Or, alternatively, provide the streams and
partition string in a getIdentifier method.
> Provide meaningful toString for task instances
> ----------------------------------------------
>
> Key: SAMZA-90
> URL: https://issues.apache.org/jira/browse/SAMZA-90
> Project: Samza
> Issue Type: Improvement
> Components: container
> Affects Versions: 0.6.0
> Reporter: Jakob Homan
> Assignee: Rekha Joshi
> Labels: newbie
> Fix For: 0.7.0
>
> Attachments: SAMZA-90.0.patch
>
>
> Right now if one logs from within a StreamTask, the toString looks a little
> something like this:
> {noformat}
> 2013-11-25 23:34:33 AssembleCallGraph [INFO] In window:
> com.linkedin.samsa.jobs.AssembleCallGraph@5782b9b5{noformat}
> This is less than useful when a container is responsible for more than one
> task instance.
> When creating the task instance we should override toString with the topics
> and partition it is reponsible for, to make it easier to grok during logging.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
