You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
However, would also be interesting to investigate, perhaps in a test, what happens if we send the master a TruncateTable RPC, for a colocated table? The code looks as if it warns, but then continues?
4586 // Truncate on a colocated table should not hit master because it should be handled by a write
4587 // DML that creates a table-level tombstone.
4588 LOG_IF(WARNING, table->IsColocatedUserTable()) << "cannot truncate a colocated table on master";
...
4597 // Send a Truncate() request to each tablet in the table.
4598 SendTruncateTableRequest(table);
Summary:
Truncate Table will be disallowed on tables with a snapshot schedule as part of this change unless the flag to enable it is explicitly turned ON.
Note that this restriction currently does not yet apply to colocated tables which will be handled as part of a separate issue.
Test Plan: ybd --cxx_test yb-admin-snapshot-schedule-test --gtest-filter YbAdminSnapshotScheduleTest.TestTruncateDisallowedWithPitr
Reviewers: bogdan, skedia
Reviewed By: skedia
Subscribers: asrivastava, ybase
Differential Revision: https://phabricator.dev.yugabyte.com/D16271
Description
Since TRUNCATE is not transactional and we do not yet have #7130, we should just disable TRUNCATE being executed against tables involved in PITR.
The work is likely to be similar to #8415.
The text was updated successfully, but these errors were encountered: