Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Investigate slow data loading #3111

Closed
ndeodhar opened this issue Dec 7, 2019 · 3 comments
Closed

Investigate slow data loading #3111

ndeodhar opened this issue Dec 7, 2019 · 3 comments
Assignees
@spolitov
Labels
area/docdb YugabyteDB core features area/ysql Yugabyte SQL (YSQL) priority/high High Priority
Milestone
v2.1

Comments

@ndeodhar
Copy link
Contributor

ndeodhar commented Dec 7, 2019

While loading data for TPCC benchmark, we see that the data loading starts off at ~1300 ops / sec and then drops to ~100 ops / sec within a few minutes and stays at that level. Total time to load data ends up being 2+ hours.

We should investigate what causes the throughput to drop 10x. I've attached a screenshot of data loading metrics.
Screen Shot 2019-12-06 at 3 29 03 PM

The first spike is when I tried loading data with 1 thread and the second spike is loading data with 4 threads. It's also interesting to note that there is no difference in speed while using 1 thread and 4 threads.
@ndeodhar ndeodhar added area/ysql Yugabyte SQL (YSQL) area/docdb YugabyteDB core features priority/high High Priority labels Dec 7, 2019
@ndeodhar ndeodhar added this to the v2.1 milestone Dec 7, 2019
@spolitov
Copy link
Contributor

spolitov commented Dec 9, 2019

During load, TPCC benchmark uses a fixed number of threads. It explains no difference between 1 or 4 threads.

spolitov added a commit that referenced this issue Dec 10, 2019
@spolitov
[#3111] Cleanup intents written for foreign key check
e0a1661 
Summary:
During foreign key check we write read intents even in case of snapshot isolation.
But it was not reflected in YBOperation::wrote_data, so those intents were not cleaned on transaction commit/abort.

This diff fixes the issue.

Test Plan: ybd --gtest_filter PgMiniTest.ForeignKey*

Reviewers: timur, jason, mikhail

Reviewed By: mikhail

Subscribers: ybase

Differential Revision: https://phabricator.dev.yugabyte.com/D7680
@rkarthik007 rkarthik007 mentioned this issue Dec 30, 2019
Open
@spolitov spolitov closed this as completed Jan 9, 2020
@rkarthik007
Copy link
Collaborator

Hi folks, this has been resolved correct?

@ndeodhar
Copy link
Contributor Author

Yes, it has been resolved.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@spolitov spolitov

Labels
area/docdb YugabyteDB core features area/ysql Yugabyte SQL (YSQL) priority/high High Priority
Projects
None yet
Milestone
v2.1
Development

No branches or pull requests
3 participants
@spolitov @rkarthik007 @ndeodhar