-
Notifications
You must be signed in to change notification settings - Fork 466
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.
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
CREATE CLUSTER REPLICA
takes several seconds to return
#13854
Comments
This is reproducible locally -- each After creating a bunch of replicas, even after destroying them all, Materialize still seems to be in "slow mode" (new replicas take several seconds), so maybe there is some kind of resource leak ... I am also seeing significant CPU usage in the Postgres binary locally (~20% of one core). Not sure if that's related. |
I can reproduce the problem locally. Start Materialize: for i in {0..100}; do
psql -p 6875 -h localhost -U materialize -q -c "create cluster replica default.r$i size '1';"
done Timing output:
The compute processes hover at around 1-2% CPU utilization, postgres and |
It looks much better after 550670b, and I'm now able to create 100 replicas. My terminal now gets spammed with:
EDIT: The error message is because by default postgres only allows 100 concurrent connections. Increasing the limit ( |
@antiguru what's the timing like when creating the 100 replicas? Is it low? (0-500ms per replica, no increase over time) I think investigating whether we should be worried about this fatal postgres error is a separate issue. |
It still slows down over time:
|
My SSD is back at 100% utilization around the 100 replica mark. |
I would describe this as still being the bad behavior (I would have typed |
We can probably close this and reopen if observed again. |
Closing this as it's outdated. We can open a new issue if acute slowness is observed again. |
What version of Materialize are you using?
unstable-d5cddec91fdd0088783aa2231bb53cd5c78b2efa
How did you install Materialize?
Materialize Cloud
What is the issue?
CREATE CLUSTER REPLICA
takes several seconds to return on latest main. This didn't use to be the case (in fact, I think it works fine on alpha.3)It seems to get slower the more replicas there are.
Relevant log output
No response
The text was updated successfully, but these errors were encountered: