Skip to content

Latest commit

 

History

History
149 lines (109 loc) · 4.72 KB

FAQ.md

File metadata and controls

149 lines (109 loc) · 4.72 KB
Raw

FAQ for BlueBrain Spack Usage

Building Software

Q: How do I build and load a piece of software?

Expand answer

We'll install MG(1), a standard editor.

Make sure you're setup as per the instructions linked on the README, then install the editor via

  $ spack install mg

One can then run mg with

  $ spack load mg
  $ mg foo.txt

Q: Why do I have to rebuild the entire world?

Expand answer

If you are on the BlueBrain5, you shouldn't need to.

As described here, one can use the system packages available with appropriate configuration options. If those instructions don't help, please use the #spack channel on Slack.

Q: Why is it so slow to interact with the Spack repository on GPFS

Expand answer

Make sure the spack repo is checked out in a subdirectory of $HOME. The spack repository is quite large, and when it is checked out under a /gpfs/bbp.cscs.ch/project/* directory, performance can be 10x slower than on the SSD provided storage of $HOME.

Q: Is there a binary cache?

Expand answer

No. Possibilities are being explored to provide a binary cache configuration for our desktops.

Modules

Q: Why are the module files not being rebuilt?

Expand answer

The spack module tcl refresh command respects a exclude list that can be found via:

  $ spack config blame modules

This exclude list is overruled by a corresponding include list. If your software is not listed in the latter, no modules will be generated for it.

Use

  $ spack config add "modules:default:tcl:include:[my_package]"
  $ spack module tcl refresh my_package

To produce an up-to-date module for my_package (adjust as needed).

Concretization Issues

Q: Why does Spack fail with cryptic error messages

Expand answer

When determining what to build (to concretize a spec in "Spack lingo"), sometimes Spack will not be able to satisfy the requirements of all software needing to be built. It may then display a somewhat cryptic error message:

  ❯ spack spec -I py-morphology-repair-workflow\^[email protected]:
  Input spec
  --------------------------------
   -   py-morphology-repair-workflow
   -       ^[email protected]:

  Concretized
  --------------------------------
  ==> Error: py-morphology-repair-workflow ^[email protected]: is unsatisfiable, errors are:
    no version satisfies the given constraints

      To see full clingo unsat cores, re-run with `spack --show-cores=full`
      For full, subset-minimal unsat cores, re-run with `spack --show-cores=minimized
      Warning: This may take (up to) hours for some specs

By running the recommendation, one may produce more cryptic output:

  ❯ spack --show-cores=full spec -I py-morphology-repair-workflow\^[email protected]:
  Input spec
  --------------------------------
   -   py-morphology-repair-workflow
   -       ^[email protected]:

  Concretized
  --------------------------------
  ==> Error: py-morphology-repair-workflow ^[email protected]: is unsatisfiable, conflicts are:
    condition(5258)
    condition(5275)
    condition(5281)
    dependency_condition(5258,"py-morph-validator","py-pandas")
    dependency_condition(5275,"py-morphology-repair-workflow","py-morph-validator")
    dependency_condition(5281,"py-morphology-repair-workflow","py-pandas")
    dependency_type(5258,"build")
    dependency_type(5275,"run")
    dependency_type(5281,"run")
    imposed_constraint(5258,"version_satisfies","py-pandas","0.25:1.2.99")
    no version satisfies the given constraints
    root("py-morphology-repair-workflow")
    version_satisfies("py-pandas","1.3:")

      For full, subset-minimal unsat cores, re-run with `spack --show-cores=minimized
      Warning: This may take (up to) hours for some specs

By analyzing the output, it can be seen that py-morphology-repair-workflow depends on py-morph-validator, which in turn depends on py-pandas between versions 0.25 and 1.2.99 (using the numerical references together with the package names / versions). This in turn conflicts with the user requirement of a py-pandas newer than 1.3. Removing said user requirement will make the software install. In other instances, loosening dependency requirements in packages may be the appropriate solution.