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.

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

[BUG] segfault when using UcpListener in TCP-only setup #2346

Closed
abellina opened this issue May 5, 2021 · 2 comments
Closed

[BUG] segfault when using UcpListener in TCP-only setup #2346

abellina opened this issue May 5, 2021 · 2 comments
Assignees
Labels
bug Something isn't working P0 Must have for release shuffle things that impact the shuffle plugin

Comments

@abellina
Copy link
Collaborator

abellina commented May 5, 2021

This is to track: openucx/ucx#6755

In branch-0.6 if the spark.rapids.shuffle.ucx.listener.enabled=true internal configuration is turned on (it is off by default), we get a segfault in UCX 1.10.1 and in 1.11 so far (until openucx/ucx#6244 goes in).

In the mean time, the workaround described in openucx/ucx#6755 works:

spark.executorEnv.UCX_SOCKADDR_TLS_PRIORITY=tcp

We can close this issue after testing that we can remove this temporary workaround.

@abellina abellina added bug Something isn't working ? - Needs Triage Need team to review and classify labels May 5, 2021
@abellina abellina self-assigned this May 5, 2021
@abellina abellina added P0 Must have for release shuffle things that impact the shuffle plugin labels May 5, 2021
@abellina
Copy link
Collaborator Author

abellina commented May 7, 2021

This is getting traction for 1.10.1 and the expectation is that UCX 1.10.1 will ship with a fix.

@abellina
Copy link
Collaborator Author

This is fixed with v1.10.1 with the merge of: openucx/ucx#6765

@sameerz sameerz removed the ? - Needs Triage Need team to review and classify label May 11, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working P0 Must have for release shuffle things that impact the shuffle plugin
Projects
None yet
Development

No branches or pull requests

2 participants