feat(corda-connector): scp jars to nodes #621
Assignees
Labels
Comments
petermetz
added a commit
to petermetz/cacti
that referenced
this issue
Mar 17, 2021
Implements transferring the cordapp jar files onto Corda nodes via SSH+SCP. It supports deploying to multiple nodes in a single request via an array of deployment configuration parameters that the caller can specify. Credentials are travelling on the wire in plain text right now which is of course unacceptable and a follow-up issue has been created to rectify this by adding keychain support or some other type of solution that does not make it necessary for the caller to send up SSH and RPC credentials with their deployment request. Another thing that we'll have to fix prior to GA is that right now there is no SSH host key verification. For development mode Corda nodes: There's possibility to declare which cordapp jar contains database migrations for H2 and the deployment endpoint can run these as well. The graceful shutdown method is implemented on our RPC connection class because this is not yet supported on the version of Corda that we are targeting with the connector (v4.5). The new test that verifies that all is working well is called deploy-cordapp-jars-to-nodes.test.ts and what it does is very similar to the older test called jvm-kotlin-spring-server.test.ts but this one does its invocations with a contract that has been built and deployed from scratch not like the old test which relied on the jars already being present in the AIO Corda container by default. The current commit is also tagged in the container registry as: hyperledger/cactus-connector-corda-server:2021-03-16-feat-621 Fixes hyperledger#621 Signed-off-by: Peter Somogyvari <peter.somogyvari@accenture.com>
petermetz
added a commit
to petermetz/cacti
that referenced
this issue
Mar 19, 2021
Implements transferring the cordapp jar files onto Corda nodes via SSH+SCP. It supports deploying to multiple nodes in a single request via an array of deployment configuration parameters that the caller can specify. Credentials are travelling on the wire in plain text right now which is of course unacceptable and a follow-up issue has been created to rectify this by adding keychain support or some other type of solution that does not make it necessary for the caller to send up SSH and RPC credentials with their deployment request. Another thing that we'll have to fix prior to GA is that right now there is no SSH host key verification. For development mode Corda nodes: There's possibility to declare which cordapp jar contains database migrations for H2 and the deployment endpoint can run these as well. The graceful shutdown method is implemented on our RPC connection class because this is not yet supported on the version of Corda that we are targeting with the connector (v4.5). The new test that verifies that all is working well is called deploy-cordapp-jars-to-nodes.test.ts and what it does is very similar to the older test called jvm-kotlin-spring-server.test.ts but this one does its invocations with a contract that has been built and deployed from scratch not like the old test which relied on the jars already being present in the AIO Corda container by default. The current commit is also tagged in the container registry as: hyperledger/cactus-connector-corda-server:2021-03-16-feat-621 Fixes hyperledger#621 Signed-off-by: Peter Somogyvari <peter.somogyvari@accenture.com>
petermetz
added a commit
to petermetz/cacti
that referenced
this issue
Mar 19, 2021
Implements transferring the cordapp jar files onto Corda nodes via SSH+SCP. It supports deploying to multiple nodes in a single request via an array of deployment configuration parameters that the caller can specify. Credentials are travelling on the wire in plain text right now which is of course unacceptable and a follow-up issue has been created to rectify this by adding keychain support or some other type of solution that does not make it necessary for the caller to send up SSH and RPC credentials with their deployment request. Another thing that we'll have to fix prior to GA is that right now there is no SSH host key verification. For development mode Corda nodes: There's possibility to declare which cordapp jar contains database migrations for H2 and the deployment endpoint can run these as well. The graceful shutdown method is implemented on our RPC connection class because this is not yet supported on the version of Corda that we are targeting with the connector (v4.5). The new test that verifies that all is working well is called deploy-cordapp-jars-to-nodes.test.ts and what it does is very similar to the older test called jvm-kotlin-spring-server.test.ts but this one does its invocations with a contract that has been built and deployed from scratch not like the old test which relied on the jars already being present in the AIO Corda container by default. The current commit is also tagged in the container registry as: hyperledger/cactus-connector-corda-server:2021-03-16-feat-621 Fixes hyperledger#621 Signed-off-by: Peter Somogyvari <peter.somogyvari@accenture.com>
petermetz
added a commit
to petermetz/cacti
that referenced
this issue
Mar 23, 2021
Implements transferring the cordapp jar files onto Corda nodes via SSH+SCP. It supports deploying to multiple nodes in a single request via an array of deployment configuration parameters that the caller can specify. Credentials are travelling on the wire in plain text right now which is of course unacceptable and a follow-up issue has been created to rectify this by adding keychain support or some other type of solution that does not make it necessary for the caller to send up SSH and RPC credentials with their deployment request. Another thing that we'll have to fix prior to GA is that right now there is no SSH host key verification. For development mode Corda nodes: There's possibility to declare which cordapp jar contains database migrations for H2 and the deployment endpoint can run these as well. The graceful shutdown method is implemented on our RPC connection class because this is not yet supported on the version of Corda that we are targeting with the connector (v4.5). The new test that verifies that all is working well is called deploy-cordapp-jars-to-nodes.test.ts and what it does is very similar to the older test called jvm-kotlin-spring-server.test.ts but this one does its invocations with a contract that has been built and deployed from scratch not like the old test which relied on the jars already being present in the AIO Corda container by default. The current commit is also tagged in the container registry as: hyperledger/cactus-connector-corda-server:2021-03-16-feat-621 Fixes hyperledger#621 Signed-off-by: Peter Somogyvari <peter.somogyvari@accenture.com>
petermetz
added a commit
to petermetz/cacti
that referenced
this issue
Mar 24, 2021
Implements transferring the cordapp jar files onto Corda nodes via SSH+SCP. It supports deploying to multiple nodes in a single request via an array of deployment configuration parameters that the caller can specify. Credentials are travelling on the wire in plain text right now which is of course unacceptable and a follow-up issue has been created to rectify this by adding keychain support or some other type of solution that does not make it necessary for the caller to send up SSH and RPC credentials with their deployment request. Another thing that we'll have to fix prior to GA is that right now there is no SSH host key verification. For development mode Corda nodes: There's possibility to declare which cordapp jar contains database migrations for H2 and the deployment endpoint can run these as well. The graceful shutdown method is implemented on our RPC connection class because this is not yet supported on the version of Corda that we are targeting with the connector (v4.5). The new test that verifies that all is working well is called deploy-cordapp-jars-to-nodes.test.ts and what it does is very similar to the older test called jvm-kotlin-spring-server.test.ts but this one does its invocations with a contract that has been built and deployed from scratch not like the old test which relied on the jars already being present in the AIO Corda container by default. The current commit is also tagged in the container registry as: hyperledger/cactus-connector-corda-server:2021-03-16-feat-621 Fixes hyperledger#621 Signed-off-by: Peter Somogyvari <peter.somogyvari@accenture.com>
petermetz
added a commit
to petermetz/cacti
that referenced
this issue
Mar 24, 2021
Implements transferring the cordapp jar files onto Corda nodes via SSH+SCP. It supports deploying to multiple nodes in a single request via an array of deployment configuration parameters that the caller can specify. Credentials are travelling on the wire in plain text right now which is of course unacceptable and a follow-up issue has been created to rectify this by adding keychain support or some other type of solution that does not make it necessary for the caller to send up SSH and RPC credentials with their deployment request. Another thing that we'll have to fix prior to GA is that right now there is no SSH host key verification. For development mode Corda nodes: There's possibility to declare which cordapp jar contains database migrations for H2 and the deployment endpoint can run these as well. The graceful shutdown method is implemented on our RPC connection class because this is not yet supported on the version of Corda that we are targeting with the connector (v4.5). The new test that verifies that all is working well is called deploy-cordapp-jars-to-nodes.test.ts and what it does is very similar to the older test called jvm-kotlin-spring-server.test.ts but this one does its invocations with a contract that has been built and deployed from scratch not like the old test which relied on the jars already being present in the AIO Corda container by default. The current commit is also tagged in the container registry as: hyperledger/cactus-connector-corda-server:2021-03-16-feat-621 Fixes hyperledger#621 Signed-off-by: Peter Somogyvari <peter.somogyvari@accenture.com>
petermetz
added a commit
to petermetz/cacti
that referenced
this issue
Mar 25, 2021
Implements transferring the cordapp jar files onto Corda nodes via SSH+SCP. It supports deploying to multiple nodes in a single request via an array of deployment configuration parameters that the caller can specify. Credentials are travelling on the wire in plain text right now which is of course unacceptable and a follow-up issue has been created to rectify this by adding keychain support or some other type of solution that does not make it necessary for the caller to send up SSH and RPC credentials with their deployment request. Another thing that we'll have to fix prior to GA is that right now there is no SSH host key verification. For development mode Corda nodes: There's possibility to declare which cordapp jar contains database migrations for H2 and the deployment endpoint can run these as well. The graceful shutdown method is implemented on our RPC connection class because this is not yet supported on the version of Corda that we are targeting with the connector (v4.5). The new test that verifies that all is working well is called deploy-cordapp-jars-to-nodes.test.ts and what it does is very similar to the older test called jvm-kotlin-spring-server.test.ts but this one does its invocations with a contract that has been built and deployed from scratch not like the old test which relied on the jars already being present in the AIO Corda container by default. The current commit is also tagged in the container registry as: hyperledger/cactus-connector-corda-server:2021-03-16-feat-621 Fixes hyperledger#621 Signed-off-by: Peter Somogyvari <peter.somogyvari@accenture.com>
petermetz
added a commit
to petermetz/cacti
that referenced
this issue
Mar 25, 2021
Implements transferring the cordapp jar files onto Corda nodes via SSH+SCP. It supports deploying to multiple nodes in a single request via an array of deployment configuration parameters that the caller can specify. Credentials are travelling on the wire in plain text right now which is of course unacceptable and a follow-up issue has been created to rectify this by adding keychain support or some other type of solution that does not make it necessary for the caller to send up SSH and RPC credentials with their deployment request. Another thing that we'll have to fix prior to GA is that right now there is no SSH host key verification. For development mode Corda nodes: There's possibility to declare which cordapp jar contains database migrations for H2 and the deployment endpoint can run these as well. The graceful shutdown method is implemented on our RPC connection class because this is not yet supported on the version of Corda that we are targeting with the connector (v4.5). The new test that verifies that all is working well is called deploy-cordapp-jars-to-nodes.test.ts and what it does is very similar to the older test called jvm-kotlin-spring-server.test.ts but this one does its invocations with a contract that has been built and deployed from scratch not like the old test which relied on the jars already being present in the AIO Corda container by default. The current commit is also tagged in the container registry as: hyperledger/cactus-connector-corda-server:2021-03-16-feat-621 Fixes hyperledger#621 Signed-off-by: Peter Somogyvari <peter.somogyvari@accenture.com>
petermetz
added a commit
to petermetz/cacti
that referenced
this issue
Mar 26, 2021
Implements transferring the cordapp jar files onto Corda nodes via SSH+SCP. It supports deploying to multiple nodes in a single request via an array of deployment configuration parameters that the caller can specify. Credentials are travelling on the wire in plain text right now which is of course unacceptable and a follow-up issue has been created to rectify this by adding keychain support or some other type of solution that does not make it necessary for the caller to send up SSH and RPC credentials with their deployment request. Another thing that we'll have to fix prior to GA is that right now there is no SSH host key verification. For development mode Corda nodes: There's possibility to declare which cordapp jar contains database migrations for H2 and the deployment endpoint can run these as well. The graceful shutdown method is implemented on our RPC connection class because this is not yet supported on the version of Corda that we are targeting with the connector (v4.5). The new test that verifies that all is working well is called deploy-cordapp-jars-to-nodes.test.ts and what it does is very similar to the older test called jvm-kotlin-spring-server.test.ts but this one does its invocations with a contract that has been built and deployed from scratch not like the old test which relied on the jars already being present in the AIO Corda container by default. The current commit is also tagged in the container registry as: hyperledger/cactus-connector-corda-server:2021-03-16-feat-621 Fixes hyperledger#621 Signed-off-by: Peter Somogyvari <peter.somogyvari@accenture.com>
petermetz
added a commit
that referenced
this issue
Mar 26, 2021
Implements transferring the cordapp jar files onto Corda nodes via SSH+SCP. It supports deploying to multiple nodes in a single request via an array of deployment configuration parameters that the caller can specify. Credentials are travelling on the wire in plain text right now which is of course unacceptable and a follow-up issue has been created to rectify this by adding keychain support or some other type of solution that does not make it necessary for the caller to send up SSH and RPC credentials with their deployment request. Another thing that we'll have to fix prior to GA is that right now there is no SSH host key verification. For development mode Corda nodes: There's possibility to declare which cordapp jar contains database migrations for H2 and the deployment endpoint can run these as well. The graceful shutdown method is implemented on our RPC connection class because this is not yet supported on the version of Corda that we are targeting with the connector (v4.5). The new test that verifies that all is working well is called deploy-cordapp-jars-to-nodes.test.ts and what it does is very similar to the older test called jvm-kotlin-spring-server.test.ts but this one does its invocations with a contract that has been built and deployed from scratch not like the old test which relied on the jars already being present in the AIO Corda container by default. The current commit is also tagged in the container registry as: hyperledger/cactus-connector-corda-server:2021-03-16-feat-621 Fixes #621 Signed-off-by: Peter Somogyvari <peter.somogyvari@accenture.com>
jordigiam
pushed a commit
to kikoncuo/cactus
that referenced
this issue
Apr 8, 2021
Implements transferring the cordapp jar files onto Corda nodes via SSH+SCP. It supports deploying to multiple nodes in a single request via an array of deployment configuration parameters that the caller can specify. Credentials are travelling on the wire in plain text right now which is of course unacceptable and a follow-up issue has been created to rectify this by adding keychain support or some other type of solution that does not make it necessary for the caller to send up SSH and RPC credentials with their deployment request. Another thing that we'll have to fix prior to GA is that right now there is no SSH host key verification. For development mode Corda nodes: There's possibility to declare which cordapp jar contains database migrations for H2 and the deployment endpoint can run these as well. The graceful shutdown method is implemented on our RPC connection class because this is not yet supported on the version of Corda that we are targeting with the connector (v4.5). The new test that verifies that all is working well is called deploy-cordapp-jars-to-nodes.test.ts and what it does is very similar to the older test called jvm-kotlin-spring-server.test.ts but this one does its invocations with a contract that has been built and deployed from scratch not like the old test which relied on the jars already being present in the AIO Corda container by default. The current commit is also tagged in the container registry as: hyperledger/cactus-connector-corda-server:2021-03-16-feat-621 Fixes hyperledger#621 Signed-off-by: Peter Somogyvari <peter.somogyvari@accenture.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Is your feature request related to a problem? Please describe.
Right now the deploy contract endpoint only adds the jar to the class path of the connector back end, but not to the cordapp dir of the corda node itself (which must be done via SSH unfortunately, same way as with Fabric).
So this task is about adding support for that SSH/SCP mechanism that will make sure that the jars that are being sent up will also be present on the cordapp dir of the node when the request returns with a successful status code of 2xx.
Describe the solution you'd like
targetNodeHostsarray where each item of the array is an object with keychain references in it (and possibly other necessary metdata such as the cordapp dir file path).This way the deploy endpoint can loop through the target node hosts, SSH into all of them and dump the jars in the right location for the corda nodes to pick it up.
Describe alternatives you've considered
A bunch! This has been going on for months in terms of design/research...
Looked into braid, looked into corda attachments, etc. Nothing seems to provide the necessary flexibility that we want so it's just SSH for now, gets the job done and with the keychain mechanism in place we don't have to sacrifice security either.
Additional context
Right now the deploy contract endpoint only works if the jar is already present in the cordapp dir of the node so this is why we need this enhancement.
cc: @takeutak @sfuji822 @hartm @jonathan-m-hamilton @AzaharaC @jordigiam @kikoncuo
The text was updated successfully, but these errors were encountered: