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

backup: TableDescriptor's being seen in offline mode when they should be public #40951

Closed
pbardea opened this issue Sep 20, 2019 · 0 comments · Fixed by #40996
Closed

backup: TableDescriptor's being seen in offline mode when they should be public #40951

pbardea opened this issue Sep 20, 2019 · 0 comments · Fixed by #40996
Assignees
@pbardea

Comments

@pbardea
Copy link
Contributor

pbardea commented Sep 20, 2019 •
edited
Loading

Describe the problem

Please describe the issue you observed, and any steps we can take to reproduce it:
There have been flakes in several tests that require the test to wait up to 5 seconds for a RESTORed/IMPORTed table to be no longer seen as offline.

To Reproduce

Run the TestBackupRestoreWithConcurrentWrites test under stress (be sure to remove the code the adds the 5 second delay).

Expected behavior
The table descriptor should no longer be visible in the offline state as soon as the RESTORE or IMPORT completes.

Additional data / screenshots
According to the table descriptor lease RFC, table leases may be held for up to 2 version changes (which may explain why we're seeing table descriptors one version behind).

Preliminary experiments have been done to add an intermediary state and OnSuccess of the job mark the table as PUBLIC_STAGING then PUBLIC. However, when the test is run under stress we still see failures.

Additional context
What was the impact?

  • Test flakiness
@pbardea pbardea self-assigned this Sep 20, 2019
pbardea added a commit to pbardea/cockroach that referenced this issue Sep 20, 2019
@pbardea
backup: add wait to flaky test
891b407 
This test has been flaky because the TableDescriptor being read is being
seen as OFFLINE. This issue seems to disappear after waiting a bit. This
also happens during IMPORT and this issue was avoided in a similar way
(see cockroachdb#36832).

This should be removed and cockroachdb#40951 is tracking this issue.

Release justification: Work-around for a flaky test.

Release note: None
@pbardea pbardea mentioned this issue Sep 20, 2019
Closed
pbardea added a commit to pbardea/cockroach that referenced this issue Sep 20, 2019
@pbardea
backup: add wait to flaky test
a5e6f9e 
This test has been flaky because the TableDescriptor being read is being
seen as OFFLINE. This issue seems to disappear after waiting a bit. This
also happens during IMPORT and this issue was avoided in a similar way
(see cockroachdb#36832).

This should be removed and cockroachdb#40951 is tracking this issue.

Release justification: Work-around for a flaky test.

Release note: None
pbardea added a commit to pbardea/cockroach that referenced this issue Sep 23, 2019
@pbardea
backup: add wait to flaky test
e083fbf 
This test has been flaky because the TableDescriptor being read is being
seen as OFFLINE. This issue seems to disappear after waiting a bit. This
also happens during IMPORT and this issue was avoided in a similar way
(see cockroachdb#36832).

This should be removed and cockroachdb#40951 is tracking this issue.

Release justification: Work-around for a flaky test.

Release note: None
@pbardea pbardea mentioned this issue Sep 23, 2019
Merged
@craig craig bot closed this as completed in bf42553 Sep 23, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@pbardea pbardea

Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging a pull request may close this issue.

sql: fix lease interaction with offline tables pbardea/cockroach
1 participant
@pbardea