Skip to content

Further comments from rereading - 03-cvmfs_containers #7

@katilp

Description

@katilp

Some points of confusion, likely to be solved all together with some brief additional clarifications.

  • The statement "This is the recommended way!" leaves the reader asking what are the other ways then. It is not clear compared to what.

  • From the last point of the previous episode (for the large images: "Do not use for large-scale job submission nor on GitLab!"), I expect that the light-weight images will then be good for both of those usages. It is not explicitly mentioned, unless the recommended way above refers to this (and if so, then these two points go together).

  • "instead we use Docker manually as if it was installed on a local machine" gives (me) the impression of having to open the docker environment locally, but it certainly refers to the yaml script part. Maybe for the sake of clarity, use the same wording (not necessarily "manually") in the text and in the yaml file comments so that a slightly confused or distracted reader can easily find the point.

  • for a curious reader, maybe a link to the docker docs for mount options rslave and rshared? The previous lectures on docker only cover -v <path on host>:<path in container> here

  • "The caveat of using these light-weight images is that they cannot be run autonomously, but always need to have CVMFS access to do anything useful." -> but what has been done so far in this episode through the yaml script is precisely to make cvmfs available, isn't it? And demonstrated by the gitlab ci job with ZPeakAnalysis?

  • after this episode, I have a container available in my gitlab container registry and I can see from the ci output that the job has successfully on a container which was able to read the CMSSW area from cvmfs. It is not entirely clear to me what I can do with it:

  • Is it only usable in the context of gitlab ci as?

  • Could I use it in batch job submission (referring to the last point in the previous episode mentioned above)?

  • and I could use it in docker locally if I had cvmfs mounted?

  • if it makes sense, maybe add a point related to the questions above (how should I use this thing I just did) in the key points at the end

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions