Skip to content

Avoid checking permission of Babelfish temp tables on parallel worker #560

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

Conversation

Deepesh125
Copy link
Contributor

@Deepesh125 Deepesh125 commented Apr 1, 2025

Description

Consider following facts,

  1. Babelfish temp tables are implemented using ENR which is not shared between different backends. So if Parallel worker
    tries to check permissions on Babelfish then it will fail.

  2. Any user should be able to access Babelfish temp tables under given session.

  3. Postgres by default does not allow parallel operations on temp tables. Attempt to do so will result in run time error.

Due to above facts, we should avoid permission check on temp tables within parallel workers while ensuring that leader
does required permission check on other tables. This commits achieves this behaviour by introducing following three
hooks,

  1. ParallelQueryMain_hook -- Hook that allows other extensions to pass on additional details from Leader node to
    parallel worker. For example, Babelfish extension can pass details of Babelfish temp table defined under current
    session with Parallel workers.

  2. ExecInitParallelPlan_hook -- Hook that allows Parallel worker to gather additional details passed by Leader node.
    For example, Babelfish extension can collect the details of Babelfish temp table shared by Leader node so that it
    can avoid permission checks.

  3. ExecCheckOneRelPerms_hook -- Hook that allows extension control permission checking on given relation/table.
    For example, Babelfish can use it to avoid permission check on temp tables under parallel worker.

Issues Resolved

BABEL-5703

Check List

  • Commits are signed per the DCO using --signoff

By submitting this pull request, I confirm that my contribution is under the terms of the PostgreSQL license, and grant any person obtaining a copy of the contribution permission to relicense all or a portion of my contribution to the PostgreSQL License solely to contribute all or a portion of my contribution to the PostgreSQL open source project.

For more information on following Developer Certificate of Origin and signing off your commits, please check here.

@tanscorpio7
Copy link
Contributor

Just curious how parallel workers fetches other details for temp tables relations after permission check ?

It must need more information about the relation while executing the node with temp table.

@Deepesh125
Copy link
Contributor Author

Just curious how parallel workers fetches other details for temp tables relations after permission check ?

It must need more information about the relation while executing the node with temp table.

Postgres code is written in a way that it will throw an error if any kind of access made to temp table under parallel operations. If we are concerned with operation metadata then it would be already calculated by leader node. Parallel worker just init the plan and pass it to executor run. Do you have any specific example that we should try?

@Deepesh125 Deepesh125 changed the title Dummy PR to test fix for BABEL-5703 Avoid checking permission of Babelfish temp tables on parallel worker Apr 4, 2025
* chance to skip permission check for such use cases.
*/
if (IsBabelfishParallelWorker() &&
ExecCheckOneRelPerms_hook &&
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can we put extension hook check first to be more safe

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

what difference will it make?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Shouldn't the parallel worker check be in the extension?

* chance to skip permission check for such use cases.
*/
if (IsBabelfishParallelWorker() &&
ExecCheckOneRelPerms_hook &&
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Shouldn't the parallel worker check be in the extension?

@Deepesh125 Deepesh125 requested a review from kuntalghosh April 10, 2025 03:24
@Deepesh125 Deepesh125 merged commit 5e62ffa into babelfish-for-postgresql:BABEL_5_X_DEV__PG_17_X Apr 10, 2025
2 checks passed
@Deepesh125 Deepesh125 deleted the jira-babel-5703-3 branch April 10, 2025 03:40
Deepesh125 added a commit to amazon-aurora/postgresql_modified_for_babelfish that referenced this pull request Apr 10, 2025
…babelfish-for-postgresql#560)

Consider following facts,

1. Babelfish temp tables are implemented using ENR which is not shared between different backends. So if Parallel
worker tries to check permissions on Babelfish then it will fail.
2. Any user should be able to access Babelfish temp tables under given session.
3. Postgres by default does not allow parallel operations on temp tables. Attempt to do so will result in run time error.

Due to above facts, we should avoid permission check on temp tables within parallel workers while ensuring that leader
does required permission check on other tables. This commits achieves this behaviour by introducing following three
hooks,

ParallelQueryMain_hook -- Hook that allows other extensions to pass on additional details from Leader node to
parallel worker. For example, Babelfish extension can pass details of Babelfish temp table defined under current
session with Parallel workers.

ExecInitParallelPlan_hook -- Hook that allows Parallel worker to gather additional details passed by Leader node.
For example, Babelfish extension can collect the details of Babelfish temp table shared by Leader node so that it
can avoid permission checks.

ExecCheckOneRelPerms_hook -- Hook that allows extension control permission checking on given relation/table.
For example, Babelfish can use it to avoid permission check on temp tables under parallel worker.

Task: BABEL-5703
Signed-off-by: Dipesh Dhameliya <[email protected]>
(cherry picked from commit 5e62ffa)
Deepesh125 added a commit to amazon-aurora/postgresql_modified_for_babelfish that referenced this pull request Apr 10, 2025
…babelfish-for-postgresql#560)

Consider following facts,

1. Babelfish temp tables are implemented using ENR which is not shared between different backends. So if Parallel
worker tries to check permissions on Babelfish then it will fail.
2. Any user should be able to access Babelfish temp tables under given session.
3. Postgres by default does not allow parallel operations on temp tables. Attempt to do so will result in run time error.

Due to above facts, we should avoid permission check on temp tables within parallel workers while ensuring that leader
does required permission check on other tables. This commits achieves this behaviour by introducing following three
hooks,

ParallelQueryMain_hook -- Hook that allows other extensions to pass on additional details from Leader node to
parallel worker. For example, Babelfish extension can pass details of Babelfish temp table defined under current
session with Parallel workers.

ExecInitParallelPlan_hook -- Hook that allows Parallel worker to gather additional details passed by Leader node.
For example, Babelfish extension can collect the details of Babelfish temp table shared by Leader node so that it
can avoid permission checks.

ExecCheckOneRelPerms_hook -- Hook that allows extension control permission checking on given relation/table.
For example, Babelfish can use it to avoid permission check on temp tables under parallel worker.

Task: BABEL-5703
Signed-off-by: Dipesh Dhameliya <[email protected]>
(cherry picked from commit 5e62ffa)
Deepesh125 added a commit to amazon-aurora/postgresql_modified_for_babelfish that referenced this pull request Apr 10, 2025
…babelfish-for-postgresql#560)

Consider following facts,

1. Babelfish temp tables are implemented using ENR which is not shared between different backends. So if Parallel
worker tries to check permissions on Babelfish then it will fail.
2. Any user should be able to access Babelfish temp tables under given session.
3. Postgres by default does not allow parallel operations on temp tables. Attempt to do so will result in run time error.

Due to above facts, we should avoid permission check on temp tables within parallel workers while ensuring that leader
does required permission check on other tables. This commits achieves this behaviour by introducing following three
hooks,

ParallelQueryMain_hook -- Hook that allows other extensions to pass on additional details from Leader node to
parallel worker. For example, Babelfish extension can pass details of Babelfish temp table defined under current
session with Parallel workers.

ExecInitParallelPlan_hook -- Hook that allows Parallel worker to gather additional details passed by Leader node.
For example, Babelfish extension can collect the details of Babelfish temp table shared by Leader node so that it
can avoid permission checks.

ExecCheckOneRelPerms_hook -- Hook that allows extension control permission checking on given relation/table.
For example, Babelfish can use it to avoid permission check on temp tables under parallel worker.

Task: BABEL-5703
Signed-off-by: Dipesh Dhameliya <[email protected]>
(cherry picked from commit 5e62ffa)
Deepesh125 added a commit to amazon-aurora/postgresql_modified_for_babelfish that referenced this pull request Apr 10, 2025
…babelfish-for-postgresql#560)

Consider following facts,

1. Babelfish temp tables are implemented using ENR which is not shared between different backends. So if Parallel
worker tries to check permissions on Babelfish then it will fail.
2. Any user should be able to access Babelfish temp tables under given session.
3. Postgres by default does not allow parallel operations on temp tables. Attempt to do so will result in run time error.

Due to above facts, we should avoid permission check on temp tables within parallel workers while ensuring that leader
does required permission check on other tables. This commits achieves this behaviour by introducing following three
hooks,

ParallelQueryMain_hook -- Hook that allows other extensions to pass on additional details from Leader node to
parallel worker. For example, Babelfish extension can pass details of Babelfish temp table defined under current
session with Parallel workers.

ExecInitParallelPlan_hook -- Hook that allows Parallel worker to gather additional details passed by Leader node.
For example, Babelfish extension can collect the details of Babelfish temp table shared by Leader node so that it
can avoid permission checks.

ExecCheckOneRelPerms_hook -- Hook that allows extension control permission checking on given relation/table.
For example, Babelfish can use it to avoid permission check on temp tables under parallel worker.

Task: BABEL-5703
Signed-off-by: Dipesh Dhameliya <[email protected]>
(cherry picked from commit 5e62ffa)
Deepesh125 added a commit to amazon-aurora/postgresql_modified_for_babelfish that referenced this pull request Apr 10, 2025
…babelfish-for-postgresql#560)

Consider following facts,

1. Babelfish temp tables are implemented using ENR which is not shared between different backends. So if Parallel
worker tries to check permissions on Babelfish then it will fail.
2. Any user should be able to access Babelfish temp tables under given session.
3. Postgres by default does not allow parallel operations on temp tables. Attempt to do so will result in run time error.

Due to above facts, we should avoid permission check on temp tables within parallel workers while ensuring that leader
does required permission check on other tables. This commits achieves this behaviour by introducing following three
hooks,

ParallelQueryMain_hook -- Hook that allows other extensions to pass on additional details from Leader node to
parallel worker. For example, Babelfish extension can pass details of Babelfish temp table defined under current
session with Parallel workers.

ExecInitParallelPlan_hook -- Hook that allows Parallel worker to gather additional details passed by Leader node.
For example, Babelfish extension can collect the details of Babelfish temp table shared by Leader node so that it
can avoid permission checks.

ExecCheckOneRelPerms_hook -- Hook that allows extension control permission checking on given relation/table.
For example, Babelfish can use it to avoid permission check on temp tables under parallel worker.

Task: BABEL-5703
Signed-off-by: Dipesh Dhameliya <[email protected]>
(cherry picked from commit 5e62ffa)
Deepesh125 pushed a commit to amazon-aurora/postgresql_modified_for_babelfish that referenced this pull request Apr 13, 2025
Deepesh125 added a commit that referenced this pull request Apr 14, 2025
…#560) (#567)

Consider following facts,

1. Babelfish temp tables are implemented using ENR which is not shared between different backends. So if Parallel
worker tries to check permissions on Babelfish then it will fail.
2. Any user should be able to access Babelfish temp tables under given session.
3. Postgres by default does not allow parallel operations on temp tables. Attempt to do so will result in run time error.

Due to above facts, we should avoid permission check on temp tables within parallel workers while ensuring that leader
does required permission check on other tables. This commits achieves this behaviour by introducing following three
hooks,

ParallelQueryMain_hook -- Hook that allows other extensions to pass on additional details from Leader node to
parallel worker. For example, Babelfish extension can pass details of Babelfish temp table defined under current
session with Parallel workers.

ExecInitParallelPlan_hook -- Hook that allows Parallel worker to gather additional details passed by Leader node.
For example, Babelfish extension can collect the details of Babelfish temp table shared by Leader node so that it
can avoid permission checks.

ExecCheckOneRelPerms_hook -- Hook that allows extension control permission checking on given relation/table.
For example, Babelfish can use it to avoid permission check on temp tables under parallel worker.

Task: BABEL-5703
Signed-off-by: Dipesh Dhameliya <[email protected]>
(cherry picked from commit 5e62ffa)
Deepesh125 added a commit that referenced this pull request Apr 14, 2025
…#560) (#566)

Consider following facts,

1. Babelfish temp tables are implemented using ENR which is not shared between different backends. So if Parallel
worker tries to check permissions on Babelfish then it will fail.

2. Any user should be able to access Babelfish temp tables under given session.

3. Postgres by default does not allow parallel operations on temp tables. Attempt to do so will result in run time error.

Due to above facts, we should avoid permission check on temp tables within parallel workers while ensuring that
leader does required permission check on other tables. This commits achieves this behaviour by introducing following
three hooks,

ParallelQueryMain_hook -- Hook that allows other extensions to pass on additional details from Leader node to
parallel worker. For example, Babelfish extension can pass details of Babelfish temp table defined under current
session with Parallel workers.

ExecInitParallelPlan_hook -- Hook that allows Parallel worker to gather additional details passed by Leader node.
For example, Babelfish extension can collect the details of Babelfish temp table shared by Leader node so that it
can avoid permission checks.

ExecCheckOneRelPerms_hook -- Hook that allows extension control permission checking on given relation/table.
For example, Babelfish can use it to avoid permission check on temp tables under parallel worker.

Task: BABEL-5703
Signed-off-by: Dipesh Dhameliya <[email protected]>
(cherry picked from commit 5e62ffa)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants