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
Not sure if it's a bug or not, but I prefer to be sure.
Describe the bug
If we had a disk failure, we cannot recreate the table without removing the node on zookeeper.
How to reproduce
clickhouse-server 20.4
The error can be reproduced easily using docker containers. Two containers are present, clickhouse1a and clickhouse1b, both are them are part of the same shard, both of them have a volume, and both of them have an hostname (clickhouse1a and clickhouse1b).
We create a table using the ON CLUSTER clause.
CREATE TABLE test ON CLUSTER 'cluster' (foo UInt32) Engine = ReplicatedMergeTree('/clickhouse/test', '{replica-name}') PRIMARY KEY foo ORDER BY foo
We insert datas on the table.
INSERT INTO test VALUES (1)
then, we stop and remove the container clickhouse1b, and after this we remove the volume affiliated to clickhouse1b.
After running the container clickhouse1b again, and try to recreate the table again, we got this error.
Received exception from server (version 20.4.3):
Code: 57. DB::Exception: Received from 127.0.0.1:10001. DB::Exception: Directory for table data data/default/test/ already exists.
Expected behavior
Different way to fix it. ClickHouse could possibly get his metadatas back from zookeeper, or during creation of the table, with a setting or by default, to be less strict in a case that's the same instance with the same schemas create the table again.
Error message and/or stacktrace
If applicable, add screenshots to help explain your problem.
Additional context
Add any other context about the problem here.
The text was updated successfully, but these errors were encountered:
YiuRULE
added
the
bug
Confirmed user-visible misbehaviour in official release
label
Jun 9, 2020
Hum slighty different but maybe related. Ensure as I don't know too much the workflow of ClickHouse with zookeeper during a table creation. the difference between both are on this one, we were able to create the table successfully but lost the metadatas in a way or an another, while on the first issue, we wasn't able to create the table at all.
Not sure if it's a bug or not, but I prefer to be sure.
Describe the bug
If we had a disk failure, we cannot recreate the table without removing the node on zookeeper.
How to reproduce
clickhouse-server 20.4
The error can be reproduced easily using docker containers. Two containers are present, clickhouse1a and clickhouse1b, both are them are part of the same shard, both of them have a volume, and both of them have an hostname (
clickhouse1a
andclickhouse1b
).We create a table using the
ON CLUSTER
clause.CREATE TABLE test ON CLUSTER 'cluster' (foo UInt32) Engine = ReplicatedMergeTree('/clickhouse/test', '{replica-name}') PRIMARY KEY foo ORDER BY foo
We insert datas on the table.
INSERT INTO test VALUES (1)
then, we stop and remove the container
clickhouse1b
, and after this we remove the volume affiliated toclickhouse1b
.After running the container
clickhouse1b
again, and try to recreate the table again, we got this error.Expected behavior
Different way to fix it. ClickHouse could possibly get his metadatas back from zookeeper, or during creation of the table, with a setting or by default, to be less strict in a case that's the same instance with the same schemas create the table again.
Error message and/or stacktrace
If applicable, add screenshots to help explain your problem.
Additional context
Add any other context about the problem here.
The text was updated successfully, but these errors were encountered: