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
Updated instructions for restoring replication metadata in ZooKeeper by specifying the table name. Added a query to automate restoration for all replicated tables.
- Run `SYSTEM RESTORE REPLICA` to rebuild replication metadata in ZooKeeper again.
49
+
- Run `SYSTEM RESTORE REPLICA table_name` to rebuild replication metadata in ZooKeeper again.
50
50
- Start ClickHouse on the new cluster and switch production traffic.
51
51
- add more replicas as needed
52
52
53
53
54
54
NOTES:
55
55
56
-
1. If you are using a mount point that differs from /var/lib/clickhouse/data, adjust the rsync command accordingly to point to the correct location. For example, suppose you reconfigure the storage path as follows in /etc/clickhouse-server/config.d/config.xml.
56
+
1. You can build a script to run restore replica commands over all replicated tables by query:
2. If you are using a mount point that differs from /var/lib/clickhouse/data, adjust the rsync command accordingly to point to the correct location. For example, suppose you reconfigure the storage path as follows in /etc/clickhouse-server/config.d/config.xml.
57
64
```
58
65
<clickhouse>
59
66
<!-- Path to data directory, with trailing slash. -->
@@ -63,4 +70,10 @@ NOTES:
63
70
```
64
71
You'll need to use `/data1/clickhouse` instead of `/var/lib/clickhouse`in the rsync paths.
65
72
66
-
2. ClickHouse docker container image does not have rsync installed. Add it using apt-get or run sidecar in k8s.
73
+
3. ClickHouse Docker container image does not have rsync installed. Add it using apt-get or run sidecar in k8s or run a service pod with volumes attached.
74
+
For clickhouse-operator instances, you can to stop all pods by CHI definition.
0 commit comments