My goal is to edit the schema of my Alpha cluster using Ratel, but without turning on ACL.
I’m running a Ratel docker image and a single Alpha / Zero (installed myself) in another docker image on the same system. Without setting the whitelist option, I’m able to connect to the Alpha cluster through Ratel. So far so good.
However, when trying to change the schema through Ratel, I’m getting an “unauthorized ip address at 172.17.0.1”, which is the docker gateway IP.
Reading the threads led me to whitelisting the Ratel container IP address 172.17.0.3 as below:
Unfortunately, it seems that using the --security flag turns on ACL. I am now prompted for ACL login for the Alpha cluster in Ratel. The default groot/password does not work, giving the message “The server is not ready to accept requests”.
Could anyone please advise what I could change to get on track? I don’t require Ratel or ACL for my production cluster but would like to use Ratel during development.
The --security flag does not enable ACL. Check that an alpha node is actually up and running and the connection details in Ratel are correct.
Click “Continue” in the Ratel UI, no need to login if you did not setup ACL.
You’re correct that alpha is not running correctly.
Even with a clean install curl https://get.dgraph.io -sSf | bash the following error occurs:
E0724 04:22:54.282077 55 groups.go:1229] Error during SubscribeForUpdates for prefix
"\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x15dgraph.graphql.schema\x00": unable to find
any servers for group: 1. closer err: <nil>
Could you advise on how I can create a new schema file (what file name, folder)? This seems strange after a clean install, but perhaps something is missing.
It seems, that the the new install ‘somehow’ is still reading the RAFT config state from the old w directory from the previous install, as per the error in the Alpha logs: unable to find any servers for group 1.
Could you once again ensure that you clean-up all the directories from the previous installation, viz., the zw for the Zeroes and the w and the p dir for the Alphas before re-trying the installation.
Also, can you let us know if you use the GraphQL API or just DQL ? In case of the former (GraphQL), once you install a fresh cluster after another clean-up, you can first try uploading your GraphQL schema via the /admin/schema endpoint and then attempt to load your data via the Live Loader. Another option is to provide both the DQL and GraphQL schemas via the -s and -g options of the Live Loader as well.
Of course, in case of the latter (just DQL), you can Live Load the data directly by supplying the DQL schema (via the -s flag) and the RDF/JSON data files (via the -f flag).
Thanks @rarvikar for those suggestions. The installation was completely clean, i.e., rolled a fresh docker image. This makes me wonder if there is a problem with the source installation files. Probably not, but it’s curious behavior.
I’ll try the other options you suggested to get past this barrier.
Thanks @Jeff_Kanel. Can you let us know which image are you using with respect to the Dgraph version ? I’ll be happy to test the same just to ensure everything works OK.