[jira] [Commented] (IGNITE-12663) Let's limit all thread pools' sizes to 3 in tests
[
https://issues.apache.org/jira/browse/IGNITE-12663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17038201#comment-17038201
]
Ignite TC Bot commented on IGNITE-12663:
{panel:title=Branch: [pull/7429/head] Base: [master] : No blockers
found!|borderStyle=dashed|borderColor=#ccc|titleBGColor=#D6F7C1}{panel}
[TeamCity *-- Run :: All*
Results|https://ci.ignite.apache.org/viewLog.html?buildId=5050920buildTypeId=IgniteTests24Java8_RunAll]
> Let's limit all thread pools' sizes to 3 in tests
> -
>
> Key: IGNITE-12663
> URL: https://issues.apache.org/jira/browse/IGNITE-12663
> Project: Ignite
> Issue Type: Test
>Reporter: Ilya Kasnacheev
>Assignee: Ilya Kasnacheev
>Priority: Major
> Time Spent: 10m
> Remaining Estimate: 0h
>
> The rationale is, it is very painful shoving through thread dumps of test
> with N CPU threads in every pool.
> Why 3? It should allow all kinds of race conditions to still happen. Of
> course, we should allow specific tests to override this value.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
[jira] [Commented] (IGNITE-12663) Let's limit all thread pools' sizes to 3 in tests
[ https://issues.apache.org/jira/browse/IGNITE-12663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17037088#comment-17037088 ] Ilya Kasnacheev commented on IGNITE-12663: -- [~agura] Actually, Idea is very awful at presenting thread dumps. You can't search threads by name, and it keeps resetting your select to first thread. And the default JVM thread dump just dumps them all to console, without any filtering. In Idea you can't even save program output to file. It's unusable currently. I have faced the rebalance thread pool issue - it's 4 by default, and system pool has to be bigger. This makes the change almost useless, but I will still take a look where we can get it to. > Let's limit all thread pools' sizes to 3 in tests > - > > Key: IGNITE-12663 > URL: https://issues.apache.org/jira/browse/IGNITE-12663 > Project: Ignite > Issue Type: Test >Reporter: Ilya Kasnacheev >Assignee: Ilya Kasnacheev >Priority: Major > Time Spent: 10m > Remaining Estimate: 0h > > The rationale is, it is very painful shoving through thread dumps of test > with N CPU threads in every pool. > Why 3? It should allow all kinds of race conditions to still happen. Of > course, we should allow specific tests to override this value. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-12663) Let's limit all thread pools' sizes to 3 in tests
[ https://issues.apache.org/jira/browse/IGNITE-12663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17037080#comment-17037080 ] Andrey N. Gura commented on IGNITE-12663: - [~ilyak] 3 is pretty good number. But we don;t talk about numbers ) There are many tools that simplify thread dump analysis. E.g. Idea collapse similar threads into groups. If scrolling is problem for you just use quick navigation with patterns. I believe grep, sed and thread dump analysis tools will help you. There is no need to change code for it. > Let's limit all thread pools' sizes to 3 in tests > - > > Key: IGNITE-12663 > URL: https://issues.apache.org/jira/browse/IGNITE-12663 > Project: Ignite > Issue Type: Test >Reporter: Ilya Kasnacheev >Assignee: Ilya Kasnacheev >Priority: Major > Time Spent: 10m > Remaining Estimate: 0h > > The rationale is, it is very painful shoving through thread dumps of test > with N CPU threads in every pool. > Why 3? It should allow all kinds of race conditions to still happen. Of > course, we should allow specific tests to override this value. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-12663) Let's limit all thread pools' sizes to 3 in tests
[ https://issues.apache.org/jira/browse/IGNITE-12663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17037040#comment-17037040 ] Ilya Kasnacheev commented on IGNITE-12663: -- "You always can cut off thread dumps from the log (at least partially)." - this is very non-trivial to do since JVM will output e.g. epoll() as RUNNABLE [~agura] I'm not sure why our tests should run with different configuration depending on the machine it is running on. If 3 is bad value, let's discuss it. I still think it should be fixed. > Let's limit all thread pools' sizes to 3 in tests > - > > Key: IGNITE-12663 > URL: https://issues.apache.org/jira/browse/IGNITE-12663 > Project: Ignite > Issue Type: Test >Reporter: Ilya Kasnacheev >Assignee: Ilya Kasnacheev >Priority: Major > > The rationale is, it is very painful shoving through thread dumps of test > with N CPU threads in every pool. > Why 3? It should allow all kinds of race conditions to still happen. Of > course, we should allow specific tests to override this value. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (IGNITE-12663) Let's limit all thread pools' sizes to 3 in tests
[ https://issues.apache.org/jira/browse/IGNITE-12663?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17035556#comment-17035556 ] Andrey N. Gura commented on IGNITE-12663: - [~ilyak] What is your motivation? From my point this restriction will lead to: - many unexpected failed and frozen tests - starvations - impossibility to detect some races I really don't like this idea. > Let's limit all thread pools' sizes to 3 in tests > - > > Key: IGNITE-12663 > URL: https://issues.apache.org/jira/browse/IGNITE-12663 > Project: Ignite > Issue Type: Test >Reporter: Ilya Kasnacheev >Assignee: Ilya Kasnacheev >Priority: Major > > The rationale is, it is very painful shoving through thread dumps of test > with N CPU threads in every pool. > Why 3? It should allow all kinds of race conditions to still happen. Of > course, we should allow specific tests to override this value. -- This message was sent by Atlassian Jira (v8.3.4#803005)
