I am surprised at the StateManagement functionality of Dapr, I'm unsure why this was included as part of the project? What is your opinion on interacting with the data outside of the Key Value mechanism Dapr supports? For instance when I want to make an Aggregate call to a MongoDB call, or build up a Search SQL query does this go through Dapr or I go direct to the data store?
It seems to me that the benefit of Dapr handing the state quickly goes when you need to make data calls outside of CRUD? I can see the retry policy but these are easily implemented using various libraries.
It seems to me that the benefit of Dapr handing the state quickly goes when you need to make data calls outside of CRUD? I can see the retry policy but these are easily implemented using various libraries.
If you had a collection of User states, and you wanted a screen to search the Users, for instance SELECT * FROM dapr1-Users WHERE Name LIKE '%James%', would you advise connecting directly to the Store or are you intending Dapr to support this concept and be the proxy to the store?
@amanbha dapr is installed on WSL but "dapr init" command is not properly working. After vizualizing the "dapr init" command execution in windows a new "daprd.exe" file has been placed in "c:\dapr" folder. is there a way to get the "daprd" file for linux manually so that I can try to see if dapr is working on WSL or not?
While doinging a
`
'head' is not recognized as an internal or external command,
operable program or batch file.
'head' is not recognized as an internal or external command,
operable program or batch file.
process_begin: CreateProcess(NULL, uname, ...) failed.
CGO_ENABLED=0 GOOS=windows GOARCH=amd64 go build -ldflags="-X github.com/dapr/dapr/pkg/version.commit=369ae9f -X github.com/dapr/dapr/pkg/version.version=edge -s -w" -o ./dist/windows_amd64/release/daprd.exe -mod=vendor ./cmd/daprd/main.go;
'CGO_ENABLED' is not recognized as an internal or external command,
operable program or batch file.
Makefile:132: recipe for target 'dist/windows_amd64/release/daprd.exe' failed
mingw32-make: * [dist/windows_amd64/release/daprd.exe] Error 1
'
Can anyone tell me what might be the fix for this?
mingw32-make.exe build to build the dapr repository it fails giving an error`
'head' is not recognized as an internal or external command,
operable program or batch file.
'head' is not recognized as an internal or external command,
operable program or batch file.
process_begin: CreateProcess(NULL, uname, ...) failed.
CGO_ENABLED=0 GOOS=windows GOARCH=amd64 go build -ldflags="-X github.com/dapr/dapr/pkg/version.commit=369ae9f -X github.com/dapr/dapr/pkg/version.version=edge -s -w" -o ./dist/windows_amd64/release/daprd.exe -mod=vendor ./cmd/daprd/main.go;
'CGO_ENABLED' is not recognized as an internal or external command,
operable program or batch file.
Makefile:132: recipe for target 'dist/windows_amd64/release/daprd.exe' failed
mingw32-make: * [dist/windows_amd64/release/daprd.exe] Error 1
'
Can anyone tell me what might be the fix for this?
While doinging a
mingw32-make.exe buildto build the dapr repository it fails giving an error
`'head' is not recognized as an internal or external command,
operable program or batch file.
'head' is not recognized as an internal or external command,
operable program or batch file.
process_begin: CreateProcess(NULL, uname, ...) failed.
CGO_ENABLED=0 GOOS=windows GOARCH=amd64 go build -ldflags="-X github.com/dapr/dapr/pkg/version.commit=369ae9f -X github.com/dapr/dapr/pkg/version.version=edge -s -w" -o ./dist/windows_amd64/release/daprd.exe -mod=vendor ./cmd/daprd/main.go;
'CGO_ENABLED' is not recognized as an internal or external command,
operable program or batch file.
Makefile:132: recipe for target 'dist/windows_amd64/release/daprd.exe' failed
mingw32-make: * [dist/windows_amd64/release/daprd.exe] Error 1
@anuragprafulla have you followed this instruction ? I think that git\bin is not in PATH
@aloneguid : This session from Ignite might help. All of what you mention is being touched upon: https://myignite.techcommunity.microsoft.com/sessions/82059?source=sessions
Trying a short version here: SF (Fully GA'ed service in Azure or available to deploy anywhere - integrated microservices stack - cluster and orchestration to programming models), SF Mesh (preview service in Azure to run distributed / microservices applications), Dapr (OSS project - experimental - microservices building blocks for developers - side cars to abstract common microservices concerns), Rudr (OSS project experimental - Kubernetes implementation of the Open Application Model - a model for describing how components work together as applications - targeting collaboration between developers and operations), Actions (original codename for Dapr), Arc (Private preview Azure service - Using Azure as the control plane for managing resources across cloud(s) and on-premises) - hope that somewhat helps. I still recommend you watch the video :-)
Saw an announcement post for Orleans 3.0 and got confused. Didn't know that project exists but I took a look and found it extremely similar to Dapr, except being written in C# and .NET-only. Is Dapr Microsoft's attempt to build basically the same thing but cross-language? Is Dapr gonna effectively replace Orleans? Thanks :)
@xJonathanLEI No. Dapr is not a replacement for Orleans. Dapr is broader than Orleans, providing pluggable building blocks that can be used from any language and dev framework. Also see https://github.com/dapr/dapr/wiki/FAQ#how-does-dapr-relate-to-orleans-and-service-fabric-reliable-actors
@anuragprafulla the best place to start is to add additional components to Dapr (bindings, state stores, secret stores, pub-sub) in components-contrib. There a few issues there with help-wanted labels on them, or you can add one of your choice.
Hi, where I can get one example of config file for command
dapr run --config ?
Hey @devictorsilva, usually you'd use things like Beats or Fluentd to forward logs to ElasticSearch and other sinks. That said, there are a few issues now raised by people that are exploring native ElasticSearch components for Dapr. Regarding config, what do you need it for? in the next release (0.2.0) planned for later this November, we aim at releasing all our tracing exporters currently configured in Config as external Dapr components. We are working on docs for these.
Greetings, maybe someone can make use of a small dapr grpc example, we've made, using spring boot and pub/sub - publisher, subscriber
And I'm trying to use NATS instead of Redis as message broker how I can implement this ?
I'm having hard time figuring out how to use pub/sub effectively. When I publish to a topic, and an app subscribes to that topic, does it always get all the messages from the very beginning? Is there some way to control that? With the number of messages flowing through the app, storing everything forever will quickly overwhelm the system
@krzysztofwos currently it's up to the configured pub/sub behind, e.g. Redis starts from 0 (https://github.com/dapr/components-contrib/blob/32f5f81b9b1856d9af08a7d89fae311b20134c76/pubsub/redis/redis.go#L155) with Azure Service Bus you can configure a message TTL (https://github.com/dapr/components-contrib/blob/32f5f81b9b1856d9af08a7d89fae311b20134c76/pubsub/azureservicebus/metadata.go#L17)
@krzysztofwos and @RicardoNiepel : There is a bug that was fixed that causes Dapr to not remove processed messages in pub/sub. The fix will be included in the version we are releasing this week (0.2.0). @RicardoNiepel for Redis Streams, this code is there to ensure that we read from the beginning of the stream only messages that were not acknowledged, meaning that with Redis you will not get all messages replayed.
@krzysztofwos In general, as long as you return 200OK from your app (in http) or a nil error (gRPC), the message will be marked as processed and will not repeat itself. Also keep in mind the following: When running with Dapr, you assign your app an ID. (if running with the CLI, a new ID will be automatically generated for you unless specified otherwise). Each app ID will get a copy of the message at least once. So if you run your app with the same ID, you should not see the message twice. If you run your app and Dapr with a new ID for every run, you will get all the messages.
@MedAnd Not much has happened on multi-tenant scenarios. It would be good to be clear on on your scenarios for multi-tenant, since I presume that you are not thinking of this in a K8s scenario, more for Service Fabric Mesh deployments? Is that correct? For Service Fabric support this is currently in discussion and I will know more at the community call on 26th Nov.
@yaron2 as far as i can see, there is the same problem with pub/sub event deserialisazion, as in this issue: dapr/dapr#717, the metadata is missing - not sure if it is only in PublishEventEnvelope or also in CloudEventEnvelope
@MedAnd I see. Currently an application (or service i.e. an EXE) cannot have multiple different instances of a single component type, for example multiple state stores or multiple type of pub/sub. This is something that has to be fixed and in on the roadmap, so that an application could, for example, use EventHub and Kafka together for two different pub/sub. The actor level is even more granular since these are distributed at an object instance level within application instance. Honest we have never really been asked at this level of granularity and it would be quite hard to configure at a per actor instance level. At the actor type level maybe. However this is really just a configuration issue and does not address a fully multi-tenant system, of security and network boundaries that need to come from the host platform (e.g. Service fabric or K8s) So the question is, are you expecting this for actors, or will multi-component configuation at an application level be enough?
How does dapr know that a specific dapr app has registered for an input binding? or sends dapr a POST to all apps using metadata.name as route name?
My understanding is that a bindings component is registered per namespace (if using k8s), but I cannot find any information about how this is connected to individual apps running in this namespace
My understanding is that a bindings component is registered per namespace (if using k8s), but I cannot find any information about how this is connected to individual apps running in this namespace
Ok, I think I found the relevant code - just for confirmation:
the runtime uses a OPTIONS HTTP request to the endpoint and if it does not return a 404, it assumes it has registered for this input binding?
(https://github.com/dapr/dapr/blob/master/pkg/runtime/runtime.go#L470)
If this is correct and it's currently not documented, I will add it.
the runtime uses a OPTIONS HTTP request to the endpoint and if it does not return a 404, it assumes it has registered for this input binding?
(https://github.com/dapr/dapr/blob/master/pkg/runtime/runtime.go#L470)
If this is correct and it's currently not documented, I will add it.
@RicardoNiepel you’re right, you pointed it out the correct code. If you think binding concept doc is not enough, please add that information.
Not sure of the protocols here, but I have a strange problem where when I trigger an Outbound binding named 'create-transcoder' my DAPR sidecar is triggering a different Inbound binding endpoint, seemingly randomly. I have 4 event names
create-transcoder start-transcoder stop-transcoder and delete-transcoder ... So far a request for create-transcoder will trigger either stop-transcoder or delete-transcoder ... redeploying the code to K8s appears to be the thing that changes which endpoint is being called, but it's never the right endpoint.
@jocko-wowza Thanks for reporting the issue. Could you please open the issue with the repro step in dapr repo ? using the issue is easier to investigate the problem instead of gitter.
@youngbupark -- Posted this issue: dapr/dapr#794 ... Thanks
Thanks for opening the issue! We will review it soon.