peer.js | A multipeer WebRTC client | Chat library

My suggestion is that you have to use two different channels for two different chaincodes. You can install as much chaincode as you want in every peer but the problems will begin when you try to instantiate chaincode as this process is directly involved with network and channel. Though they change the instantiate process from version >= 2.

For the records:

For running a business, some organizations have to come into a consensus and so they have to create a channel where only permissioned organizational properties have access.

Here within a channel every valid organizational peer hold a physical copy of business transactional records which is known as ledger where the ledger is identical which means every ledger that belongs to same channel must hold the same records.

A ledger consists of two distinct, though related, parts – a world state and a blockchain. The world state – a database that holds current values of a set of ledger states. The blockchain – a transaction log that records all the changes that have resulted in the current the world state.

A chaincode is the business logic that is used to read the records of ledger or to write new records into the ledger or to update existing records.

A chaincode is consist with multiple smart contracts.

So The convention is for each channel, you need an identical chaincode. If you have 2 channel, than it's the convention that you have two different chaincodes.

latest tag isn't available anymore. you have to use specific tag. Below is the snippet of dave comment infabric-maintainers group.

The function is peer.destroy().

Fyi, if you need to find undocumented API functions in Javascript libraries, you can do this by running your JS file in the browser and pressing F12 to bring up the debugging console. Set a breakpoint after the object that you want to investigate is instantiated:

I set a breakpoint AFTER I instantiated peer1 (line 51). Then I hovered over peer1 (line 50) to pop up a menu of all of the properties (methods and fields) belonging to this object. If you're looking for a particular function as in this case, sometimes you find it immediately in this menu, but often you have to expand the property called __proto__ :

You can see the destroy function located here. It's just a guess whether or not a function found in this manner actually does what you want it to do, but luckily in this case it does what we want.

Looking at the logs I can see why sometimes you get peer0 and occasionally you get peer1, and that is because the discovery results and processing result in it not being possible to get the list of peers in any particular order so with 2 peers in your org and the fact that this is not a long running app but a short running invocation sometimes peer0 will be first in the list of peers and sometimes peer1. Because you have only installed the chaincode on peer0, peer1 can't honour the evaluate request and return an error.

The node-sdk should detect this and then try peer0 but either you are using an older version of the node-sdk which definitely had an issue where it did not try a different peer or the node-sdk is thinking this is (incorrectly, but may not be able to distinguish) a chaincode response and passes that back to the caller. The solution to avoid the error is to install the chaincode on all the peers.

I found the culprit. There were two issues.

1. I didn't set environment variable DISCOVERY_AS_LOCALHOST to false in explorer config. That was the reason I saw: localhost.

2. In peer I had set variable CORE_PEER_GOSSIP_EXTERNALENDPOINT, but without port. That was the reason I saw undefined as port.

There have been some changes to "grpc" one of the underlying protocols used by Fabric, and this can cause timeout problems.

You can change the default grpc options for the Composer REST Server connection to the peers by modifying the connection.json of the card you are using to start the REST server.

For each peer in the connection.json add the following settings:

My question is what should I use in path in the following code

You can find this out yourself by peeking into the code of the repo.

There you can see, that the path option of PeerServer will be used as part of an express app, that itself creates a websocket connection:

If you remove and then add a new track to a PeerConnection you need to renegotiate the offer-answer to get it working. I will recommend you to use the replaceTrack API to avoid the re-negotiation problem while changing the camera input.

you are correct about the process. If all the operations are good, then I think the docker file is good. But you have to make sure, you can only instantiate the chain code from the peer in which the chain code is installed. Because the chain code installation will not change the channel config.