batching would go right here if you want to take a stab at it: https://github.com/authzed/spicedb/blob/main/pkg/validationfile/loader.go#L126

I'll have a look at batching using Zed and the loader.go file. We are using a persistent cockroachdb installation in our DEV and QA lanes, and a spun up Kubernetes pod on our local dev boxes.

opened this issue to keep track of the request: https://github.com/authzed/spicedb/issues/752

for the persistent cases you could just load it using

zed

and call it a day

If that will handle things in a nice manner, that will definitely be the way to go for today, though I'd love to see it added in the bootstrap. I had tried the Zed tool before, but didn't see the batch option. Loading 2 million records at startup wasn't feasible at 10 transactions a second.

Thanks

@Jake - Is there a particular version that you recommend. I thought I had the latest installed, but trying to use the batch import results in

Error: failed to acquire semaphore for batch number 2: context canceled

whether the batch size is 500, 1000, or 10.

@joey thank you for answering my earlier question.

is there a way to list all the objects per user where

checkPermission

would pass - eg a list of objects that user has been granted access to?

you want

LookupResources

Perfect, thank you so much!

Hello! I'm trying to run spicedb using Kubernetes. I spun up an RDS Aurora cluster to use as the datastore. Then spun up the server... but I'm getting an error saying:

in-memory datastore is not persistent and not feasible to run in a high availability fashion

Which means that it isn't actually connecting to the datastore. I was attempting to follow this example: https://github.com/authzed/examples/blob/main/kubernetes/example.yaml#L113-L117, so I added the value as being the URI for the RDS Aurora Cluster. Is there something else needed to get the server to talk to the DB?

did you specify

SPICEDB_DATASTORE_ENGINE

as well?

I did not! I missed that that was a required. Thanks! 🙏

anytime 🙂

Is there a way to see all the created relationships within the hosted "UI"? I'm seeing 100 created relationships.

currently, no

the intention was to provide an overview during development

the recommendation is to use

zed

or the API with filters to find the relationships you want

we'll likely change that over time

thank you.

Hello! I'm back 😅 . I'm trying to setup the backend to use rds aurora (mysql) I'm getting this error: > {"level":"error","error":"failed to create datastore: NewMySQLDatastore: failed to create connector: invalid DSN: missing the slash separating the database name","time":"2022-08-18T19:50:06Z","message":"terminated with errors"} The only port I have setup is the GRPC port: > - name: "grpc" > containerPort: 50051 > protocol: "TCP" Wondering if I need to setup another port from the example? https://github.com/authzed/examples/blob/main/kubernetes/example.yaml#L133-L146... they seemed to all be for other things, so I left them out.

that means the connection string you're using for MySQL is missing the logical database name at the end

Is the connection string different than the URI?

its a URI + the logical database name

I'm running

client.LookupResources

and getting

recv error rpc error: code = Unavailable desc = upstream request timeout

is it due to the number of resources?

are you doing this against authzed.com?