importccl: use direct datum workload generator by default#36250
Merged
craig[bot] merged 1 commit intocockroachdb:masterfrom May 7, 2019
Merged
importccl: use direct datum workload generator by default#36250craig[bot] merged 1 commit intocockroachdb:masterfrom
craig[bot] merged 1 commit intocockroachdb:masterfrom
Conversation
Member
|
This change is |
danhhz
reviewed
Mar 28, 2019
Contributor
danhhz
left a comment
There was a problem hiding this comment.
As discussed in #36106, I'm comfortable with this. Now that I'm sitting here with the Approve button in front of me, I'm getting cold feet about merging this before 19.1.0 goes out the door. This only affects roachtests run on master, but even with our due diligence, there's always a chance this destabilizes fixtures import and we really need to get every roachtest datapoint we can right now. Are you okay with waiting on this for a couple weeks?
Contributor
Author
|
SGTM |
Contributor
|
19.1.0 is out, should we merge this? |
This removes the COCKROACH_IMPORT_WORKLOAD_FASTER flag now that it has been tested for awhile, and, critically, now that it is actually faster than writing/reading CSVs. The difference is most pronouced when using distsql sorted IMPORT as it reads twice -- for sampling and then ingest -- but is also measurable in single-pass 'direct' ingest. All checked passed when running 'fixtures import tpcc --warehouses=1000', and ingest time using distsql import went from 55m via CSV to 41m using the worklaod reader directly. Release note: none.
8a10324 to
bd7e022
Compare
Contributor
Author
|
bors r+ |
craig bot
pushed a commit
that referenced
this pull request
May 7, 2019
36250: importccl: use direct datum workload generator by default r=dt a=dt This removes the COCKROACH_IMPORT_WORKLOAD_FASTER flag now that it has been tested for awhile, and, critically, now that it is actually faster than writing/reading CSVs. The difference is most pronouced when using distsql sorted IMPORT as it reads twice -- for sampling and then ingest -- but is also measurable in single-pass 'direct' ingest. All checked passed when running 'fixtures import tpcc --warehouses=1000', and ingest time using distsql import went from 55m via CSV to 41m using the worklaod reader directly. Release note: none. 37136: telemetry: make telemetry endpoint read-only r=dt a=dt Previously it was reusing the code used in periodic reporting exactly, which had the side-effect of clearing the report. This changes the endpoint to pass a paramter to the report generation to prevent that. It also makes explicit what the various callers of GetFeatureCounts want -- raw vs quantized, read-only vs resetting. Release note: none. Co-authored-by: David Taylor <tinystatemachine@gmail.com>
Contributor
Build succeeded |
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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.
This removes the COCKROACH_IMPORT_WORKLOAD_FASTER flag now that it has been tested for awhile,
and, critically, now that it is actually faster than writing/reading CSVs.
The difference is most pronouced when using distsql sorted IMPORT as it reads twice --
for sampling and then ingest -- but is also measurable in single-pass 'direct' ingest.
All checked passed when running 'fixtures import tpcc --warehouses=1000', and ingest time
using distsql import went from 55m via CSV to 41m using the worklaod reader directly.
Release note: none.