Package Installation Failing May 2020 - DataFrames ,CSV

Package Announcements
, , ,
1

Today (5/12/2020, 5th May 2020 getting this on Julia TLS (1.0.5) version. It’s our production environment. Help please!!

Code:
using Pkg
Pkg.add(Pkg.PackageSpec(;name=“LearnBase”, version=“0.3”))
specific version was set for another issue: Package Installation Failing April 2020 - JuliaDB, OnlineStats
Pkg.add(“CSV”)
Pkg.add(“DataFrames”)
Pkg.add(“HTTP”)
Pkg.add(“JSON”)
Pkg.add(“JuliaDB”)
Pkg.add(“MbedTLS”)
using CSV
using JuliaDB
using HTTP
using JSON
using Dates
@info VERSION
@info Pkg.installed()

Errors:

ERROR: LoadError: Unsatisfiable requirements detected for package DataFrames [a93c6f00]:
DataFrames [a93c6f00] log:
├─possible versions are: [0.11.7, 0.12.0, 0.13.0-0.13.1, 0.14.0-0.14.1, 0.15.0-0.15.2, 0.16.0, 0.17.0-0.17.1, 0.18.0-0.18.4, 0.19.0-0.19.4, 0.20.0-0.20.2, 0.21.0] or uninstalled
├─restricted to versions 0.19.4 by an explicit requirement, leaving only versions 0.19.4
└─restricted by compatibility requirements with CSV [336ed68f] to versions: 0.21.0 — no versions left
└─CSV [336ed68f] log:
├─possible versions are: [0.3.0-0.3.1, 0.4.0-0.4.3, 0.5.0-0.5.26, 0.6.0-0.6.2] or uninstalled
└─restricted to versions 0.6.2 by an explicit requirement, leaving only versions 0.6.2

We should ensure that LTS version and packages are not broken. Should I pin to any specific versions? Or should I install everything and keep the “good” installed directory somewhere safe, and
just copy the package directory to the deployed environment? Any guidelines?

Thanks all!

2

It looks like your requirement explicitly asks for DataFrames@0.19.4 and CSV@0.6.2 which don’t go together - have you explicitly asked for those versions at some point?

What happens if you ]add DataFrames@0.21?

And yes I think if you’ve got a production system on the LTS version somewhere it might make sense to pin the packages you’re using to prevent things from breaking.

3 Likes
3

Thanks @nilshg, it make sense. I remember for some reason, I had to set DataFrames to that specific version. I guess this fixed my issue, but let me validate this properly.

1 Like
4

You should probably use a Julia environment with fixed package versions if this is your production system.

1 Like
5

@dpsanders - Can you suggest or direct me to some documentation/sample? Here is the steps on production environment (which is very standard)

  1. Deployment process [one time task]: Copy source code and install all packages → this is done by sudo user, lets say “sshuser”
  2. Run time our packages are executed by another service account user → let say “service_user”

In my case #2 user never find the packages installed by #1 user

I know we are deviating from original issue, if not allowed, I can open another new ticket

Thanks everyone!

6

There is a discussion somewhere about how to do this, but I am not an expert. I suggest you open a different thread about that.

7

fair enough, thanks anyway.