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
Scylla rejects deletions with empty-tuple restrictions #9311
Comments
The empty-range check causes more bugs than it fixes. Replace it with an explicit check for =NULL (see scylladb#7852). Fixes scylladb#9311. Fixes scylladb#9290. Tests: unit (dev), cql-pytest on Cassandra 4.0 Signed-off-by: Dejan Mircevski <dejan@scylladb.com>
@scylladb/scylla-maint we hit this issue in https://jenkins.scylladb.com/job/scylla-4.4/job/dtest-release/121/testReport/cql_additional_tests/TestCQL/empty_in_test/ Since branch-4.4 now includes 146f7b5 Please backport |
will do |
update: very non-trivial backport, it will take me a while, since it's no longer possible to ask the author to try and do so |
we could, but are empty tuple restrictions ever seen outside of our test environment? Perhaps it's better to keep the patch that rejects updates with null key values at the price of not allowing empty tuple restrictions. Unless some 4.4 users actually rely on empty tuple restrictions, in which case it's better to revert |
@psarna so you're suggesting that for 4.4, we need to disable the test, and leave the code as it is now? |
On Tue, 2021-12-07 at 18:29 +0700, @aleksbykov wrote:
|
@nyh yes, that's one of the ways to deal with the problem, but it depends on at least two important factors:
|
It's not an important bug. How did it start happening? Was a new test added? |
Now, rolling upgrade stay the same. Just for v4.4.8 upgrade starts failing once first node was upgrade and these query failed with error instead of return empty results, Also i think the upgrade tests from 4.4 -> 4.5 could fail before upgrade because same query executed before upgrade |
I reverted it from 4.4, we can revisit it but this gets us back to status quo ante. |
And this regression since https://jenkins.scylladb.com/view/scylla-4.5/job/scylla-4.5/job/dtest-release/111/testReport/cql_additional_tests/TestCQL/empty_in_test/ @avikivity so we should probably revert 44c784c from 4.5 |
Reverted from 4.5. |
EDIT: now all rolling upgrade jobs from 4.4 and 4.5 to 4.5.3 are failing with this error:
it means that, once these changes were reverted for 4.5 (and 4.4) the tests in SCT also had to be changed to not fail with this empty tuple... is it still supported for 4.6? |
The bug should be fixed in 4.6 (and not in 4.4 and 4.5, after the reverts). |
Already backported to all vulnerable branches, removing "Backport candidate" label. |
This is an unintended consequence of 1fdaeca.
Scylla:
Cassandra:
The text was updated successfully, but these errors were encountered: