Hi.
I’m using zeebe-node 0.23.2 for my worker
and single broker v0.23.1 (docker compose yaml from zeebe-docker-compose)
whenever broker restarts, the worker does not reconnect
I’ve setup retries and onReady and other hooks - they don’t seem to function.
I saw this: https://github.com/creditsenseau/zeebe-client-node-js/issues/164
My questions -
When is the 0.23.3 expected (assuming it resolves this issue)?
Is there a reliable workaround (e.g. way to detect the disconnect, then create new ZBClient() object, or simply restart the docker container)? e.g. the client has a connected property - can I sample it periodically? I’ve tried onReady() and onConnectionError() hooks - they don’t work
Tx, @jwulf
It works! node client reconnects after restarting broker + the ‘onReady’ handler is called !
I found 2 non-critical issues:
(1) the ‘onConnectionError’ was not called on disconnect - note I did a “docker-compose restart zeebe” so the restart was fast. I did get the GRPC errors in the client’s log though, but no call to ‘onConnectionError’
(2) My worker microservice is built “FROM node:12”; I had to revert back to “FROM node:10” due to failure to install zeebe-node, originating from failure to install grpc module, due to failure while installing dependency “segfault-handler@1.3.0” - apparently node 12 is not shipped with required header files and node-gyp fails
To sum up, zeebe-node is not installing on that particular docker image (node:12). Obviously this is not a zeebe-node direct issue, just thought you’d like to know
I am getting a new release ready, and I will remove the segfault-handler. It emits a crash log if you opt-in for the pure JS implementation and it seg faults. But I’ve had trouble with it too - i had to explicitly add npm rebuild to my docker build script.