[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17259801#comment-17259801
]
David Wyles commented on SPARK-33635:
-
Thanks a lot for getting down to the route cause and
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17259446#comment-17259446
]
Jungtaek Lim commented on SPARK-33635:
--
One more point, though the root cause is actually the
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17259407#comment-17259407
]
Apache Spark commented on SPARK-33635:
--
User 'HeartSaVioR' has created a pull request for this
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17259399#comment-17259399
]
Jungtaek Lim commented on SPARK-33635:
--
I've spent some time to trace the issue, and noticed
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17257962#comment-17257962
]
Yukihito X commented on SPARK-33635:
[~david.wyles], I tried your sample code in my local dev
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17256046#comment-17256046
]
David Wyles commented on SPARK-33635:
-
[~gsomogyi] I now have my results.
I was so unhappy about
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17252153#comment-17252153
]
Gabor Somogyi commented on SPARK-33635:
---
{quote}Remember, based on all my testing, and raw kafka
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17251911#comment-17251911
]
David Wyles commented on SPARK-33635:
-
[~gsomogyi] Just so you know, I'm still doing this.
I've
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17248550#comment-17248550
]
Gabor Somogyi commented on SPARK-33635:
---
{quote}The collect in this test case is only 13 items of
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17248549#comment-17248549
]
Gabor Somogyi commented on SPARK-33635:
---
Mixed up with DStreams, in Strutured Streaming and SQL
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17247879#comment-17247879
]
David Wyles commented on SPARK-33635:
-
[~gsomogyi]
"try to turn off Kafka consumer caching. Apart
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17247878#comment-17247878
]
David Wyles commented on SPARK-33635:
-
"Since you're measuring speed I've ported the Kafka source
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17247875#comment-17247875
]
David Wyles commented on SPARK-33635:
-
I'll give all those a go and get back to you.
> Performance
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17246643#comment-17246643
]
Gabor Somogyi commented on SPARK-33635:
---
{quote}I no longer believe this is a true regression in
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17246640#comment-17246640
]
Gabor Somogyi commented on SPARK-33635:
---
I've changed to SQL because you're not executing a
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17246633#comment-17246633
]
Gabor Somogyi commented on SPARK-33635:
---
Since you're measuring speed I've ported the Kafka source
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17246631#comment-17246631
]
Gabor Somogyi commented on SPARK-33635:
---
BTW, I'm sure you know but using collect gathers all the
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17246625#comment-17246625
]
Gabor Somogyi commented on SPARK-33635:
---
[~david.wyles] try to turn off Kafka consumer caching.
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17246430#comment-17246430
]
David Wyles commented on SPARK-33635:
-
Having performed my tests I can conclude that the kafka
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17245329#comment-17245329
]
David Wyles commented on SPARK-33635:
-
The diffs on that library between 2.4.5 and 3.0.0 are very
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17245309#comment-17245309
]
David Wyles commented on SPARK-33635:
-
Also note, in my example - the only thing that changes is the
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17245306#comment-17245306
]
David Wyles commented on SPARK-33635:
-
Is anyone even able to confirm my results are not just unique
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17245304#comment-17245304
]
David Wyles commented on SPARK-33635:
-
Fair point, the library I suspect is spark-sql-kafka-0-10, is
[
https://issues.apache.org/jira/browse/SPARK-33635?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17245277#comment-17245277
]
Sean R. Owen commented on SPARK-33635:
--
I don't think this is actionable by others unless you can
24 matches
Mail list logo