sorry for the multiple emails. Something went wrong on my email provider
side.
Instead of splitting into two separate threads as its hard keep track of
the discussion. We can continue this discussion on users list as it will
keep the storm users in the discussion as well.
Thanks,
Harsha
On Fri,
I concur; staying 1 major GA version behind is a nice idea. There's a limit
to what needs to be sacrificed for legacy support...
Cheers,
A.
On Fri, Aug 12, 2016 at 5:31 PM, Bobby Evans wrote:
> +1 for me too.
>
> Java 9 is already in early access and is supposed to be GA
+1 for me too.
Java 9 is already in early access and is supposed to be GA fairly soon. I
think the last time we talked about this we decided that staying 1 release of
java behind the latest was decent practice, and I still agree. - Bobby
On Friday, August 12, 2016 4:56 AM, Satish Duggana
+1 on moving to Java 8 for 2.x release.
As Jungtaek has already mentioned Oracle stopped supporting JDK/JRE 7 since
April 2015. Java 8 brings better APIs and it will make our code look better
and future user APIs can be designed lot better.
Thanks,
Satish.
On Fri, Aug 12, 2016 at 3:11 PM,
I was mistaken the address of user mailing list. Forwarding this to user@.
-- Forwarded message -
From: Jungtaek Lim
Date: 2016년 8월 12일 (금) 오후 6:38
Subject: Re: [Discussion] Dropping Java 7 support on master
To: ,
test mail ignore.
Hi All,
Dropping java 7 support on master will allow us to use the new api
in Java 8 and since the master is being used for java migration
its good to make the decision now. Let me know your thoughts.
Thanks,
Harsha
Hi.
Thank you for replies.
Now, it worked.
2. second, you should make sure your topology has been assigned more than
10 task. because the rebalacing is based on the task to do more workers. If
you did not assigne 10 tasks(at least) to the topology explicitly, the
rebalancing can not work out as
I have one more question.
Based on some tests, some workers totally shut down and re-run, which
means each task lost all the data.
Some workers do not shut down and are resumed after rebalance, which means
each task keeps the data.
So, I assume Storm has some capabilities to distinguish them
Anybody? How do you configure log levels in local mode, please point me to the
right place to look, thanks.
Jonas
On Aug 9, 2016, at 5:37 PM, Jonas Krauß wrote:
> Looking further in the source code of storm I found that actually this
> logging is hard-coded as
10 matches
Mail list logo