Support MinIO #4878
Merged
Support MinIO #4878
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
hard as fixing, so this PR does the latter. Most of the code already works, of course. But GC creates its own S3 client. It is _not_ the same as the S3A client and these differences were significant: * S3 endpoint needs to be configured. Otherwise, the client accesses the wrong "S3". * For most MinIO installations the client needs to be configured with path style. Otherwise it will attempt host style access by default and time out in some cases such as when running in Docker. Note that this cannot be configured on Hadoop versions before 2.8.0 so users will need to use a suitable build to GC on Spark 2.4.7. A mitigating factor is that this is _anyway_ needed to use Spark with MinIO -- regardless of GC or indeed lakeFS! With these in place I was able to GC using the Spark 3.1.2 / Hadoop 3 build. Fixes #4795. I will split testing on MinIO in Esti off of that issue for separate work.
arielshaqed
added
area/client/spark
area/integrations
bug
arielshaqed
force-pushed
the
chore/4795-test-committed-gc-on-minio
branch
from
18285f7 to
f5f7654
Compare
|
I am investigating test failures on non-Hadoop3. This is strange, I would not expect AWS S3 to stop working as a result of unrelated changes :-/ |
arielshaqed
force-pushed
the
chore/4795-test-committed-gc-on-minio
branch
from
bc82800 to
43f991b
Compare
This was referenced Dec 28, 2022
| @@ -29,9 +30,14 @@ object S3ClientBuilder extends io.treeverse.clients.S3ClientBuilder { | |||
| ) | |||
| } else None | |||
|
|
|||
| val builder = AmazonS3ClientBuilder | |||
| .standard() | |||
| .withPathStyleAccessEnabled(hc.getBoolean("fs.s3a.path.style.access", true)) | |||
There was a problem hiding this comment.
can you make "fs.s3a.path.style.access" a constant too? (to keep things unified)
| new AwsClientBuilder.EndpointConfiguration(endpoint, region) | ||
| ) | ||
| else | ||
| builder |
There was a problem hiding this comment.
Suggested change
| builder | |
| builder.withRegion(region) |
There was a problem hiding this comment.
Ouch! Good catch, thanks!
Jonathan-Rosenberg
requested changes
Dec 29, 2022
arielshaqed
commented
Jan 1, 2023
There was a problem hiding this comment.
Thanks!
I want to update documentation separately, under #4913.
| @@ -29,9 +30,14 @@ object S3ClientBuilder extends io.treeverse.clients.S3ClientBuilder { | |||
| ) | |||
| } else None | |||
|
|
|||
| val builder = AmazonS3ClientBuilder | |||
| .standard() | |||
| .withPathStyleAccessEnabled(hc.getBoolean("fs.s3a.path.style.access", true)) | |||
| new AwsClientBuilder.EndpointConfiguration(endpoint, region) | ||
| ) | ||
| else | ||
| builder |
There was a problem hiding this comment.
Ouch! Good catch, thanks!
Jonathan-Rosenberg
approved these changes
Jan 1, 2023
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
#4795 was either to discover how to make MinIO work or to make it work. Since the only way to discover all the roadblocks is to fix each one, discovery is as hard as fixing. So this PR does the latter.
Most of the code already works, of course. But GC creates its own S3
client. It is not the same as the S3A client and these differences
were significant:
S3 endpoint needs to be configured. Otherwise, the client accesses
the wrong "S3".
For most MinIO installations the client needs to be configured with
path style. Otherwise it will attempt host style access by default
and time out in some cases such as when running in Docker.
Note that this cannot be configured on Hadoop versions before 2.8.0
so users will need to use a suitable build to GC on Spark 2.4.7. A
mitigating factor is that this is anyway needed to use Spark with
MinIO -- regardless of GC or indeed lakeFS!
With these in place I was able to GC using the Spark 3.1.2 / Hadoop 3
build.
Fixes #4795. I will split testing on MinIO in Esti off of that issue
for separate work.