You are viewing a plain text version of this content. The canonical link for it is here.
Posted to github@arrow.apache.org by GitBox <gi...@apache.org> on 2021/06/23 21:19:06 UTC
[GitHub] [arrow-datafusion] alamb opened a new issue #614: physical_plan::repartition::tests::repartition_with_dropping_output_stream failing locally
alamb opened a new issue #614:
URL: https://github.com/apache/arrow-datafusion/issues/614
**Describe the bug**
Reported by @rdettai in https://github.com/apache/arrow-datafusion/issues/612#issuecomment-867008993
Locally, I have a test is failing, but it does not fail on CI:
- physical_plan::repartition::tests::repartition_with_dropping_output_stream
```
---- physical_plan::repartition::tests::repartition_with_dropping_output_stream stdout ----
Running case RoundRobinBatch(2)
BarrierExec::wait waiting on barrier
Partition 0 waiting on barrier
Partition 1 waiting on barrier
Partition 1 sending batch
Partition 1 sending batch
Partition 0 sending batch
Partition 0 sending batch
BarrierExec::wait done waiting
Running case Hash([Column { name: "my_awesome_field", index: 0 }], 2)
BarrierExec::wait waiting on barrier
Partition 0 waiting on barrier
Partition 1 waiting on barrier
Partition 1 sending batch
Partition 1 sending batch
Partition 0 sending batch
Partition 0 sending batch
BarrierExec::wait done waiting
thread 'physical_plan::repartition::tests::repartition_with_dropping_output_stream' panicked at 'assertion failed: `(left == right)`
left: `["+------------------+", "| my_awesome_field |", "+------------------+", "| frob |", "+------------------+"]`,
right: `["+------------------+", "| my_awesome_field |", "+------------------+", "| foo |", "| frob |", "| gar |", "| gaz |", "| goo |", "| grob |", "+------------------+"]`:
expected:
[
"+------------------+",
"| my_awesome_field |",
"+------------------+",
"| frob |",
"+------------------+",
]
actual:
[
"+------------------+",
"| my_awesome_field |",
"+------------------+",
"| foo |",
"| frob |",
"| gar |",
"| gaz |",
"| goo |",
"| grob |",
"+------------------+",
]
', datafusion/src/physical_plan/repartition.rs:799:13
```
**To Reproduce**
Not sure
**Expected behavior**
Tests to be deterministic
**Additional context**
Add any other context about the problem here.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] alamb closed issue #614: physical_plan::repartition::tests::repartition_with_dropping_output_stream failing locally
Posted by GitBox <gi...@apache.org>.
alamb closed issue #614:
URL: https://github.com/apache/arrow-datafusion/issues/614
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] alamb closed issue #614: physical_plan::repartition::tests::repartition_with_dropping_output_stream failing locally
Posted by GitBox <gi...@apache.org>.
alamb closed issue #614:
URL: https://github.com/apache/arrow-datafusion/issues/614
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] alamb commented on issue #614: physical_plan::repartition::tests::repartition_with_dropping_output_stream failing locally
Posted by GitBox <gi...@apache.org>.
alamb commented on issue #614:
URL: https://github.com/apache/arrow-datafusion/issues/614#issuecomment-867890412
Interestingly the test is repartitioning 8 rows of input data.
The expected result says "1 row" but the failure shown above has 6 rows -- so it still appears to be partitioned but perhaps with a different hash function 🤔
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] rdettai commented on issue #614: physical_plan::repartition::tests::repartition_with_dropping_output_stream failing locally
Posted by GitBox <gi...@apache.org>.
rdettai commented on issue #614:
URL: https://github.com/apache/arrow-datafusion/issues/614#issuecomment-954690732
It is worth noting that aHash, the hashing library that is used, [clearly states](https://github.com/tkaitchuck/aHash#goals-and-non-goals) that it's hash do not follow any clear standard, so the assumption made by the tests [here ](https://github.com/apache/arrow-datafusion/blob/2eb2daada0bf05ae4321d970ede68ef9b0850204/datafusion/src/physical_plan/repartition.rs#L791-L792) that the hash will always be the same is wrong.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] rdettai commented on issue #614: physical_plan::repartition::tests::repartition_with_dropping_output_stream failing locally
Posted by GitBox <gi...@apache.org>.
rdettai commented on issue #614:
URL: https://github.com/apache/arrow-datafusion/issues/614#issuecomment-954630477
Really sorry for not answering to your question for 4 months... it is pretty rude 😅 I didn't get this error for a while because I was working from a new laptop. But last week I used my old laptop again and this error popped up once more.
I just found the reason, we were using the nightly toolchain. It still happens after updating rust to allow builds with the 2021 edition so you can reproduce it by running `rustup default nightly`.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] rdettai commented on issue #614: physical_plan::repartition::tests::repartition_with_dropping_output_stream failing locally
Posted by GitBox <gi...@apache.org>.
rdettai commented on issue #614:
URL: https://github.com/apache/arrow-datafusion/issues/614#issuecomment-954702639
I am working on a PR that does exactly that. I'll submit it ASAP. Meanwhile you can assign this to me 😉
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] Dandandan commented on issue #614: physical_plan::repartition::tests::repartition_with_dropping_output_stream failing locally
Posted by GitBox <gi...@apache.org>.
Dandandan commented on issue #614:
URL: https://github.com/apache/arrow-datafusion/issues/614#issuecomment-954701695
That makes sense.
The nightly compiler might use a different code path (specialization) that is not available.
I think the correct way to proceed is to fix the test to not depend on the exact hash function used (e.g. check that we have something of the input, not exactly this input).
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] alamb commented on issue #614: physical_plan::repartition::tests::repartition_with_dropping_output_stream failing locally
Posted by GitBox <gi...@apache.org>.
alamb commented on issue #614:
URL: https://github.com/apache/arrow-datafusion/issues/614#issuecomment-867897817
I can not reproduce this locally (I tried under docker and it passed for me 125,000 times before I got bored of running it. I also tried a few previous versions of `hashbrown` and `ahash` .
@rdettai / @jorgecarleitao:
1. Does this happen all the time (every run) or only sometimes?
2. What is the system you see this on (i.e. what OS / core count?)
3. Have you run `cargo update` recently / can you upload your `Cargo.lock` file? (I am not sure that is related, I am just trying to reproduce the issue
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] alamb edited a comment on issue #614: physical_plan::repartition::tests::repartition_with_dropping_output_stream failing locally
Posted by GitBox <gi...@apache.org>.
alamb edited a comment on issue #614:
URL: https://github.com/apache/arrow-datafusion/issues/614#issuecomment-867890412
Interestingly the test is repartitioning 8 rows of input data.
The expected result says "1 row" but the failure shown above has 6 rows -- so it still appears to be partitioned but perhaps with a different hash values 🤔
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
[GitHub] [arrow-datafusion] Dandandan commented on issue #614: physical_plan::repartition::tests::repartition_with_dropping_output_stream failing locally
Posted by GitBox <gi...@apache.org>.
Dandandan commented on issue #614:
URL: https://github.com/apache/arrow-datafusion/issues/614#issuecomment-954702892
Cool - thanks a lot!
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscribe@arrow.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org