I use save and load cmd transfer eis images to a new computer, how to run pcb demo?
should i transfer EIS folder to the new computer and use "docker-compose build " cmd to run PCB demo
the logs show disk memory is not enough.
can you verify is it now you are trying to setup the whole EIS on the new system ?
If yes, you might need to start over again starting from docker installation until the EIS build step by following to the README.md file.
Technically, you need to follow what Murni suggested you to do, to first add the new node to the ETCD cluster (the step that runs etcd_cluster_add.sh) and then generate an EIS bundle and deploy that bundle on the new node.
However, to add the new node to the ETCD cluster, this node needs to have network access, to build the ia_etcd container image and to sync the configurations from the primary node. Is your environment okay to do this? If this is not a problem, then it should work for you.
Back to the questions you had with the errors, if you examine the provision_eis.sh file, you can find it not only builds the ia_etcd container image and starts it, but also adds the EIS user, group, and creates the necessary directories. This missing /opt/intel/eis/sockets in your log is created when provision_eis.sh is run (mkdir -p $EIS_INSTALL_PATH/sockets/).
I checked back your posts, and it did appear you have your own private registry server, implying new nodes could connect to it, so this process should work for you. Please feel free to let us know if I misunderstood you or you need more assistance.
Do your additional nodes have no network access at all? My biggest concern is that EIS uses ETCD to store and sync configurations for various container services, which requires network connectivity for the synchronization. If a node is disconnected completely, we will need to save a copy of the ETCD data store, copy it over to the node, and start a standalone ETCD instance with that copy of configurations. It's not impossible but it's a hassle. Synchronization will also need to be done manually, defeating many of the EIS benefits.
If you don't mind all these, after you set up your ETCD data store, you can follow IEdgeInsights/docker_setup/deploy/README.md on your primary node to generate a bundle of the services you want to run on your additional nodes, and run this newly generated docker-compose.yml file on the new node, after you have loaded back the transferred images. There might be other problems that need to be overcome. You might as well want to build the whole set of EIS services from scratch..
i use my friend's HuaWeiWIFI connect this computer to internet. the build process failed. that computer have EIS ver1.6 , every EIS images it has three register version . the var/lib/docker/tmp file size 73G so i delete this file and /var/lib/docker/layer2 file , the free disk is enough but i can't use "docker-compose build" i have change the docker-compose file ,delete unused service.
I just do as you say , make a bundle file , and use "docker-compose up" encounter an error "stat /var/lib/docker/overlay2/****** : no such file or directory
You shouldn’t have manually removed Docker system files... Instead, use “docker rmi “ to removed Docker images no longer needed. I also find “docker system prune” useful, to removed temporary and unused files.
Can you tell us more about your environment and how you want to use it?
already install docker and docker compose
it a demo machine, i want run the PCB demo on it
i just use "docker system prune -a" cmd and re-load images into that docker(this step load the image layers into that file) and use "docker-compose up" cmd
encounter some error like this
ERROR for ia_video_ingestion cannot start service ia_video_ingestion : error while mounting volume 'var/lib/docker/edgeinsightssoftware_vol_eis_socket/_data' : failed to mount local volume: mount /opt/intel/eis/sockets:/var/lib/docker/volumes/edgeinsightssoftware_vol_eis_socket/_data, flags : 0x1000:no such file or directory