XComp merged PR #23836:
URL: https://github.com/apache/flink/pull/23836
--
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: issues-unsubscr...@flink.apache.
patricklucas commented on PR #23836:
URL: https://github.com/apache/flink/pull/23836#issuecomment-1896272486
@XComp sure, coming up shortly.
Thanks for the review, @afedulov.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to Git
XComp commented on PR #23836:
URL: https://github.com/apache/flink/pull/23836#issuecomment-1896250037
Sure. Could you provide backport PRs for 1.18 and 1.17? Then I can merge all
the necessary PRs at once.
--
This is an automated message from the Apache Git Service.
To respond to the mess
afedulov commented on PR #23836:
URL: https://github.com/apache/flink/pull/23836#issuecomment-1894316627
Unfortunately there are no definitive guidelines for such edge cases in
Flink.
That said, I believe the current behavior can be justifiably considered a
bug that your PR fixes. The ri
patricklucas commented on PR #23836:
URL: https://github.com/apache/flink/pull/23836#issuecomment-1887371839
My suspicion is that there's an extremely small chance of that but I have no
way to quantify it.
In our case it was actually the other way around: we set this option and
were
afedulov commented on PR #23836:
URL: https://github.com/apache/flink/pull/23836#issuecomment-1883003220
> But yes, the main issue is the inconsistent behavior
The purpose is clear and I think it is a good idea to improve it. My only
concern is a scenario like this:
- A user has a pipe
patricklucas commented on PR #23836:
URL: https://github.com/apache/flink/pull/23836#issuecomment-1874157629
I do not think there is much room for unexpected behavior here. The surface
area of the change is quite small and follows the existing model for
translating a Hadoop config option sp
patricklucas commented on PR #23836:
URL: https://github.com/apache/flink/pull/23836#issuecomment-1873963500
@afedulov it is quite popular to use alternate endpoint URLs for blob
storage services with S3, less so for GCS. I would actually expect it to be
more popular, as testing jobs that i
afedulov commented on PR #23836:
URL: https://github.com/apache/flink/pull/23836#issuecomment-1852263941
@patricklucas My understanding is that one would typically want to override
`fs.gs.storage.root.url` only in very specific scenarios, such as the one you
mentioned with `fake-gcs-server`
XComp commented on PR #23836:
URL: https://github.com/apache/flink/pull/23836#issuecomment-1851864223
@afedulov do you have capacity to review this PR?
--
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 g
patricklucas commented on PR #23836:
URL: https://github.com/apache/flink/pull/23836#issuecomment-1851710390
@XComp before this becomes a dead PR, do you know whom might be a good
candidate to ask for a review?
--
This is an automated message from the Apache Git Service.
To respond to the
patricklucas commented on PR #23836:
URL: https://github.com/apache/flink/pull/23836#issuecomment-1837516482
I considered a slightly different approach, delegating the lookup of the
config option to the GCS Hadoop connector itself:
https://github.com/apache/flink/compare/master...patrickluc
patricklucas commented on PR #23836:
URL: https://github.com/apache/flink/pull/23836#issuecomment-1835801919
@flinkbot run azure
--
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 comme
patricklucas commented on PR #23836:
URL: https://github.com/apache/flink/pull/23836#issuecomment-1833953627
@flinkbot run azure
--
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 comme
flinkbot commented on PR #23836:
URL: https://github.com/apache/flink/pull/23836#issuecomment-1833417452
## CI report:
* da5b66cd0bfca529b8f28869721534b2c375309a UNKNOWN
Bot commands
The @flinkbot bot supports the following commands:
- `@flinkbot run azure`
patricklucas commented on PR #23836:
URL: https://github.com/apache/flink/pull/23836#issuecomment-1833404691
Once this is accepted, I am ready with backport branches for 1.18 and 1.17.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to
patricklucas opened a new pull request, #23836:
URL: https://github.com/apache/flink/pull/23836
## What is the purpose of the change
This change configures the the GCS client used by the GSFileSystem's
RecoverableWriter codepath to use the same GCS root URL as used by Hadoop.
W
17 matches
Mail list logo
