Crate db information schema is missing the table we created

CrateDB
1

We had a table and we tried to add more nodes to the cluster as part of this exercise now we have the main node and the table schema is missing. We are able to query but we are not able to see the table schema at all. The information schema table also doesn’t show us the table name.

Any help will be greatly appreciated.

2

Hi @Naveen_Vijay

Could you share a bit more information about what the setup looks like?
What CrateDB version are you using?
Which deployment method?
Could you share the contents of the crate.yml / config used?
Are you connecting via Admin UI?
Do you see all the nodes in the Admin UI?
Could you share the logs of a node at startup?

3

Thank you for reaching out - See response below.

What CrateDB version are you using?

4.6.7

Which deployment method?

RPM installer single node instance ported to multi node and during the node synchronization phase there was some problem and we had to drop one table on one of the nodes and we are starting to see this issue.

Could you share the contents of the crate.yml / config used?

cluster.name: gra
node.name: "10.100.1.142"
path.data: /data/crate/
path.repo: /data/backup/crate/
network.bind_host: 10.100.1.142
network.publish_host: 10.100.1.142
network.host: 10.100.1.142
udc.enabled: false
bootstrap.memory_lock: true
gateway.expected_nodes: 3
gateway.recover_after_nodes: 3
cluster.initial_master_nodes 
cluster.routing.allocation.disk.watermark.low: 97%
cluster.routing.allocation.disk.watermark.high: 98%
cluster.routing.allocation.disk.watermark.flood_stage: 99%

Are you connecting via Admin UI?

Admin UI and also crash. Our application is able to query but we are not able to insert any new records to the table because of this problem.

Do you see all the nodes in the Admin UI?

Yes just that tables are not showing up.

Could you share the logs of a node at startup?

Yes We will try to get that in some time. As far as I know the logs are regular logs.

4

Did you change the path.data directory after you were running a single node?

The path.data directory holds also the state information of the cluster. When you start CrateDB as a single node with some of the Linux distro packages, it could be that you end up with a configuration with discovery.type set to single-node which would prevent that node to be part of a multi-node cluster. When you change path.data after starting a cluster and not moving the contents to the new location, you essentially start a new cluster with a new state.

1 Like