Distributed alter query(delete/update/drop partition) on cross replication clusters #11508
Conversation
src/Interpreters/DDLWorker.cpp
Outdated
| "There are two exactly the same Snowball instances " + address.readableString() + " in cluster " | ||
| + task.cluster_name, ErrorCodes::INCONSISTENT_CLUSTER_DEFINITION); | ||
| } | ||
| else ///circular replication is used. |
There was a problem hiding this comment.
What does mean circular replication?
There was a problem hiding this comment.
|
Related #1287? |
|
For my test case in integration test, it reported "RequestsDependencyWarning: urllib3 (1.23) or chardet (3.0.4) doesn't match a supported version!", this is related with "requests" package in python. For other failed tests, I can see no relation with the changes I made, so what shall I do next? |
|
Hmm, you checkout the current branch from 20.4 and try to merge into 20.4. I will fix. |
|
Your commits now live here #11703. I will probably close this PR. |
| { | ||
| throw Exception("Cannot execute DDL query, because leader was suddenly changed or logical error.", ErrorCodes::LEADERSHIP_CHANGED); | ||
| } | ||
| // if (query_context.getClientInfo().query_kind == ClientInfo::QueryKind::SECONDARY_QUERY) |
|
@nikitamikhaylov looks like you pushed to the |
|
@azat Yes, I pushed empty commit here and I don’t know how to return it back :) Maybe we can simply close this PR and corresponding branch will be deleted? |
This will not helps, since you did not push the ref of PR but create new one:
To remove create ref by your you can run the following: |
|
Or better using qualified path: |
Done |
I hereby agree to the terms of the CLA available at: https://yandex.ru/legal/cla/?lang=en
Changelog category (leave one):
Short description
Add support for distributed ddl(update/delete/drop partition) on crosss replication clusters.
Detailed description (optional):
For distributed ddl on crosss replication clusters, the syntax should be similar to the following "drop" statement:
alter table db1.tab1_local on cluster cluster1 drop partition '2018-01-01';On the other types of clusters, the DDLs remain the same as before.