Uh oh!
There was an error while loading. Please reload this page.
[SPARK-9516][UI] Improvement of Thread Dump Page - #7910
Conversation
CodingCat
commented
Aug 3, 2015
CodingCat
commented
Aug 3, 2015
SparkQA
commented
Aug 3, 2015
Test build #39606 has finished for PR 7910 at commit
|
SparkQA
commented
Aug 4, 2015
Test build #39641 has finished for PR 7910 at commit
|
squito
commented
Aug 4, 2015
Hi @CodingCat just a quick question from looking at the screenshot -- can you still click on one thread to expand the stack trace for just that? nothing looks like a hyperlink anymore. Again I'm not a UI expert by any means, so I'll defer to others opinions, but I wouldn't think they were clickable from the look of it. |
SparkQA
commented
Aug 4, 2015
Test build #39721 has finished for PR 7910 at commit
|
SparkQA
commented
Aug 4, 2015
Test build #39719 has finished for PR 7910 at commit
|
CodingCat
commented
Aug 5, 2015
Hi, @squito , sorry for the late reply, here is the screenshot the rows are still clickable...I will add |
SparkQA
commented
Aug 5, 2015
Test build #39799 has finished for PR 7910 at commit
|
CodingCat
commented
Aug 5, 2015
SparkQA
commented
Aug 5, 2015
Test build #39874 timed out for PR 7910 at commit |
SparkQA
commented
Aug 5, 2015
Test build #39914 has finished for PR 7910 at commit
|
SparkQA
commented
Aug 6, 2015
Test build #39944 has finished for PR 7910 at commit
|
squito
commented
Aug 6, 2015
Jenkins, retest this please |
There was a problem hiding this comment.
The filter seems strange to me. Won't stripping out newlines etc. lead to some weird matches?
Also, I think String.matches may not be what we want -- that matches against the entire string, not against any contained substring. It would most likely mean that everyone would have to make their pattern ".*<pattern>.*". I think we'd rather use pattern.find(), since that is more like "grep". Should probably also be in multiline mode, so:
valgrepExp=Option(request.getParamter("grepexp")).map(Pattern.compile(_, Pattern.MULTILINE))
...
grepExp.map { grep => grep.matcher(thread.stackTrace).find()}.getOrElse(true)SparkQA
commented
Aug 6, 2015
Test build #40043 has finished for PR 7910 at commit
|
CodingCat
commented
Aug 6, 2015
CodingCat
commented
Aug 6, 2015
Thanks, @squito , I addressed your comments and uploaded the latest version of code |
SparkQA
commented
Aug 6, 2015
Test build #255 has finished for PR 7910 at commit
|
SparkQA
commented
Aug 6, 2015
Test build #40065 has finished for PR 7910 at commit
|
CodingCat
commented
Aug 6, 2015
-_-||| |
SparkQA
commented
Aug 6, 2015
Test build #40073 has finished for PR 7910 at commit
|
CodingCat
commented
Aug 7, 2015
it seems that Jenkins is very very unstable in these days |
SparkQA
commented
Aug 7, 2015
Test build #40112 has finished for PR 7910 at commit
|
JoshRosen
commented
Aug 7, 2015
Yeah, we've been combatting a really bad test flakiness issue over the past 48 hours :( |
JoshRosen
commented
Oct 16, 2015
I just tested this out and noticed a few problems:
Given all of this, I don't think that this PR is ready to merge in its current form. @CodingCat, if you don't anticipate having time to make these changes, would you mind closing this in the meantime to help de-clutter the PR review queue? Thanks! |
CodingCat
commented
Oct 17, 2015
I might work on this during the week and next week, so, I prefer to keeping this open |
SparkQA
commented
Oct 22, 2015
Test build #44152 has finished for PR 7910 at commit
|
CodingCat
commented
Oct 22, 2015
CodingCat
commented
Oct 22, 2015
I adjusted the patch according to Josh's suggestions, now the search is covering all columns (I thought it only cares about stacktrace column) the stacktrace is shown as a new row For the StackTrace row, the current implementation is to add a new row following the row showing Thread ID, Name and State, instead of toggling on "hidden" class the reason is that the hidden stacktrace row is a big trouble when you sort and then expand, or grep some lines..... so....Welcome to review it..... |
SparkQA
commented
Oct 22, 2015
Test build #44175 has finished for PR 7910 at commit
|
CodingCat
commented
Oct 30, 2015
@squito@JoshRosen just ping... |
There was a problem hiding this comment.
add a toLowerCase here -- I tried searching for "RUNNABLE" first and it didn't find anything.
squito
commented
Oct 30, 2015
Hi @CodingCat, thanks for the reminder. This is looking great! I really like it. Just found a few minor things, and one design question about highlighting the task threads. |
CodingCat
commented
Nov 1, 2015
@squito Thanks for the comments, just addressed the issues |
SparkQA
commented
Nov 2, 2015
Test build #44780 has finished for PR 7910 at commit
|
There was a problem hiding this comment.
this only takes 1 arg now (crazy that this works in js ...)
squito
commented
Nov 2, 2015
thanks for the update @CodingCat , just one tiny nit, otherwise lgtm. @JoshRosen would you like to take another look? |
SparkQA
commented
Nov 3, 2015
Test build #44861 has finished for PR 7910 at commit
|
CodingCat
commented
Nov 9, 2015
Hi, @JoshRosen , since 1.6 has been cut, will we push this feature in the coming version? |
andrewor14
commented
Dec 15, 2015
retest this please |
andrewor14
commented
Dec 15, 2015
Looks great! I'm going to merge this in master if no one objects. |
SparkQA
commented
Dec 15, 2015
Test build #47699 has finished for PR 7910 at commit
|
andrewor14
commented
Dec 16, 2015
Merged into master. Thanks @CodingCat and everyone who reviewed this! |
CodingCat
commented
Dec 16, 2015
thanks @JoshRosen and @squito for reviewing it thanks for @andrewor14 for merging |








https://issues.apache.org/jira/browse/SPARK-9516
@squito@JoshRosen It's ready for the review now