update umbra and add umbra s3#951
Conversation
f18517e to
0c1d4b1
Compare
rschu1ze
left a comment
rschu1ze
left a comment
There was a problem hiding this comment.
@toschmidt I get a permission denied error when I try to push to this PR.
/data/ClickBench (schmidt/umbra26.06 >) $ git push umbra schmidt/umbra26.06
remote: Permission to umbra-db/ClickBench.git denied to rschu1ze.
fatal: unable to access 'https://github.com/umbra-db/ClickBench.git/': The requested URL returned error: 403
Could I ask you to merge from main and resolve the conflicts?
aaada50 to
f0fc7c9
Compare
|
Could you please allow pushes to your branch? It will be needed for the automation. |
f0fc7c9 to
29ed04f
Compare
GitHub Actions
Inactive
I rebased the branch, but I can't enable edits from maintainers. The fork is owned by an organization, and apparently GitHub does not support this option (https://github.com/orgs/community/discussions/5634). I can add you guys to our fork to allow edits, or do you know another way? |
|
Results for Logs:
|
|
Umbra works, but Umbra S3 doesn't. |
|
The problem is that the S3 variant requires a static AWS access key to reproduce results. We can use our own but it will leak somewhere in publicly available log files (like here). Would it be possible to omit -s3 for now? |
|
Yes, we can omit umbra-s3. It's possible to run it manually, the Readme contains details how to create and destroy the S3 bucket. For an automated run, I would not enable it as the bucket could be leaked and it requires permissions to create buckets in ec2 instance. Do you want to remove it entirely or just exclude from the automated benchmarks? |
Update the Umbra ClickBench definition: drop the primary key, ingest from the Athena hits.parquet via the umbra.parquetview table function (instead of a TSV COPY), and store the table with zstd compression (create.sql, the Docker variant, reads /data/hits.parquet from the bind mount). Switch the dataset download to hits.parquet (BENCH_DOWNLOAD_SCRIPT) to match, require the loaded row count to equal exactly 99,997,497 (a partial load otherwise sails through with implausibly fast timings on the surviving subset), and run the Docker container --privileged. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
29ed04f to
cc913e1
Compare
GitHub Actions
Waiting
|
I removed umbra-s3 and will do a follow up pull request. |
4 participants
Updates Umbra and improves runtimes by loading the data from parquet and using compression.
This also resolves the memory usage problem on smaller instances reported in #543 and the infinite loop during the concurrent measurements in #891.
Add a new Umbra variant that stores the data on S3 instead of the EBS device. Similar to ClickHouse (web), which also uses S3 as a storage backend.