I want to make a container that is able to transfer files between itself and other containers on the cluster. I have multiple containers that are responsible for executing a task, and they are waiting to get an input file to do so. I want a separate container to be responsible for handling files before and after the task is executed by the other containers. As an example:
have all files on the file manager container.
let the file manager container automatically copy a file to a task executing container.
let task executing container run the task.
transfer the output of the task executing container to the file manager container.
And i want to do this automatically, so that for example 400 input files can be processed to output files in this way. What would be the best way to realise such a process with kubernetes? Where should I start?
A simple approach would be to set up the NFS or use the File system like AWS EFS or so.
You can mount the File system or NFS directly to POD which will be in ReadWriteMany access method.
ReadWriteMany - Multiple POD can access the single file system.
If you don't want to use the Managed service like EFS or so you can also set up the file system on K8s checkout the MinIO : https://min.io/
All files will be saved in the File system and as per POD requirement, it can simply access it from the file system.
You can create different directories to separate the outputs.
If you want only read operation, meaning all PODs can read the files only you can also set up the ReadOnlyMany access mode.
If you are GCP you can checkout this nice document : https://cloud.google.com/filestore/docs/accessing-fileshares
Related
How can i retrieve any directory/file from the completed pod of a particular job in kubernetes?
I am trying to store the output file locally from the container.
Kubernetes pods don't maintain state by default. You will need to use persistent volumes for data persistence (PVC/PV).
In order to retrieve files/directories from a completed job, you will have to mount the same persistent volume to another pod after your job is completed in order to retrieve the "output file"
Thanks for the suggestion!! I finally manage to do it by using init-container along with main container... one to run the real functioning i want the job for and another one to let the container sleep and then mounted an empty directory (PVC too can be used ofcourse) and shared with both main container and init-container.
I'd like to run perf record and perf script on a process running in a container in Kubernetes (actually on Openshift). Following the approach from this blogpost I was able to get perf record working (in the sidecar). However the perf script cannot read symbols (in sidecar) because these are present only in the main container.
Therefore I'd like to mount the complete filesystem of the main container into the sidecar, e.g. under /main and then run perf script --symfs=/main. I don't want to copy the complete filesystem into an emptyDir. I've found another nice blogpost about using overlay filesystem; however IIUC I would need to create the overlay in the main container and I don't want to run that as a privileged container and require commands (like mount) to be present.
Is there any way to create sort of reverse mount, exposing a part of container to be mounted by other containers within the same pod?
Is there a way to specify a tar file of a docker image in a deployment manifest file for kubernetes? The nodes have access to a mounted network drive that will have the tar file. There's a post where the image is loaded by docker on each node, but I was wondering if there's a way just to specify the tar file and have Kubernetes do the loading and running.
--edit--
To be more exact, say I have a mounted network drive on each node, is there a way with just the manifest file to instruct kubernetes to load that image directly from tar file and not have to put it into a docker registry.
In general, no, Kubernetes can only access container images from a registry, not from a network drive, see documentation.
However, you could have a private registry inside your cluster (see docs). You could also have the images locally on the nodes (pre-pulled images) and have Kubernetes access them from there by setting imagePullPolicy to Never (see docs).
You have provided quite limited information about your environment and how it would looks like.
Two things comes to my mind.
Use initContainer to download this file using wget or similar.
Init containers are exactly like regular containers, except:
Init containers always run to completion.
Each init container must complete successfully before the next one starts.
That way you can be sure that tar file will be downloaded before your application will start. Example can be found here
Use Mount Volume
In your deployment, statefulset, pod (not sure what you are using), you can Mount Volume into pod. After that you will be able to inside pod specified path from volume. Please keep in mind that you have to use proper access modes.
To run .tar file you can use some bash commands like in this documentation.
I have a simple use case. I am trying to deploy two pods on two different nodes in Kubernetes. Pod A is a server which creates a file abc.txt after receiving an API request. I want to mount this abc.txt file onto Pod B.
If the file jhsdiak.conf (the name of this file is randomly generated) is not present on pod B before it starts, pod B will create its own default file. Hence to avoid this, the file has to be mounted onto Pod B before it starts.
Here are the things I have tried
Shared Volume using dynamically provisioned PVC -> This approach works fine if both the pods are created on the same node. Not otherwise as GCP doesn't support ReadWriteMany.
Using Kubectl CP to copy the files from Pod A to host path and then creating configmaps/secrets to mount it onto Pod B -> This approach fails as the name of file jhsdiak.conf is randomly generated.
InitContainers -> I am not sure how I can use an init container to move files from one pod to another.
Using NFS Persisted storage -> I haven't tried it yet, but seems like a lot of overhead to just move one file between pods.
Is there a better or more efficient way to solve this problem?
A similar solution is to use Cloud Storage for storing your files.
But, I have another solution than "file". Create a PubSub topic and push your files in it with Pod A.
Create a pull subscription and poll it with Pod B.
You can achieve what you want, I mean sending data from A to B, and you don't have to worry about file system.
If your cluster is compliant with Knative, the eventing solution can help you for staying inside the cluster (if it's a requirement)
I have an API that describes itself through an openapi3 file. This app is contained in a pod that also has a sidecar app that is supposed to read this file at startup time.
My probleme is how my sidecar app can read the openapi file from the other container ?
I know I could do it using a volume (emptyDir) and modify the command so my api copies the file at startup time. I'd rather not go this route. I have been looking for a feature, where I define a volume which is mapped to an existing folder in my app, but without being empty. Is there such a thing ?
One of the simplest approaches is to use emptyDir: https://kubernetes.io/docs/concepts/storage/volumes/#emptydir
In your container that generates the file that needs to be shared, mount emptyDir volume with write access and copy the file there. In your sidecar that needs to read the file, mount the same volume as read only and read the file.
With this pattern, all containers in the pod can have access to the same file system with read / write as needed.