Hey there. I’m doing some testing with V1.0.4 and finding some strange issues with exports.
I have a 5 node cluster and all 5 have the same configuration for the export config option:
“export”: “/mnt/dgraph_exports”
When I run “curl http://localhost:8081/admin/export” on dgraph1 (or any other node for that matter)
dgraph1-4 will not get any exports in their exports directory, however dgraph5 will have exports for dgraph2 in
it’s export directory:
dgraph5:/mnt/dgraph_exports/dgraph-2-2018-04-02-14-47.rdf.gz
We are currently running V0.8.3 and not running into this issue.
Any advice would be greatly appreciated.
What groups are the nodes serving? What is the value of the replicas flag that was used while starting Zero? Can you share the output of the /state endpoint? The export file is only created on the leader of a group.
dgraph-2-2018-04-02-14-47.rdf.gz - The 2 after dgraph- signifies the group for which the export is done. So dgraph5 must be the leader for group 2 while export was performed.
while i am doing the data export, i got an issue in both 0.8.3 and 1.0.4(latest)
ISSUE: {“errors”:[{“code”:“ErrorUnauthorized”,“message”:“Request from IP: 172.17.0.1”}
please tell me the solution.
Export can only be initiated from the instance that is running Dgraph server for now. From the next release, you will be able to specify a IP whitelist for the servers that can initiate the export request.