Changing /tmp file path?

[dhoconno]

Hi there,

I'm trying to run Nextstrain build on a cluster node where I don't have write access to /tmp. I can override this manually by setting TMPDIR=[path_I_can_access] Is there a more elegant way to specify the temp directory directly as part of the nextstrain build command?

Example output:

Creating conda environment workflow/envs/nextstrain.yaml...
Downloading and installing remote packages.
Environment for workflow/envs/nextstrain.yaml created (location: .snakemake/conda/5b992ba27014c417af06ad854a8e9177)
Using shell: /usr/bin/bash
Provided cores: 40
Rules claiming more threads will be scaled down.
Job stats:
job                                  count    min threads    max threads
---------------------------------  -------  -------------  -------------
add_branch_labels                        1              1              1
adjust_metadata_regions                  1              1              1
align                                    2              8              8
all                                      1              1              1
ancestral                                1              1              1
build_align                              1              8              8
calculate_epiweeks                       1              1              1
clade_files                              1              1              1
clades                                   1              1              1
colors                                   1              1              1
combine_input_metadata                   1              1              1
combine_samples                          1              1              1
combine_sequences_for_subsampling        1              1              1
diagnostic                               2              1              1
distances                                1              1              1
emerging_lineages                        1              1              1
export                                   1              1              1
filter                                   2              1              1
finalize                                 1              1              1
include_hcov19_prefix                    1              1              1
index_sequences                          1              1              1
logistic_growth                          1              1              1
mask                                     1              1              1
mutational_fitness                       1              1              1
recency                                  1              1              1
refine                                   1              1              1
rename_emerging_lineages                 1              1              1
sanitize_metadata                        2              1              1
subsample                                1              1              1
tip_frequencies                          1              1              1
traits                                   1              1              1
translate                                1              1              1
tree                                     1              8              8
total                                   37              1              8

Select jobs to execute...
Traceback (most recent call last):
  File "/var/lib/condor/execute/slot1/dir_27595/miniconda/lib/python3.9/site-packages/snakemake/__init__.py", line 699, in snakemake
    success = workflow.execute(
  File "/var/lib/condor/execute/slot1/dir_27595/miniconda/lib/python3.9/site-packages/snakemake/workflow.py", line 1073, in execute
    success = self.scheduler.schedule()
  File "/var/lib/condor/execute/slot1/dir_27595/miniconda/lib/python3.9/site-packages/snakemake/scheduler.py", line 491, in schedule
    run = self.job_selector(needrun)
  File "/var/lib/condor/execute/slot1/dir_27595/miniconda/lib/python3.9/site-packages/snakemake/scheduler.py", line 764, in job_selector_ilp
    self._solve_ilp(prob)
  File "/var/lib/condor/execute/slot1/dir_27595/miniconda/lib/python3.9/site-packages/snakemake/scheduler.py", line 813, in _solve_ilp
    prob.solve(solver)
  File "/var/lib/condor/execute/slot1/dir_27595/miniconda/lib/python3.9/site-packages/pulp/pulp.py", line 1913, in solve
    status = solver.actualSolve(self, **kwargs)
  File "/var/lib/condor/execute/slot1/dir_27595/miniconda/lib/python3.9/site-packages/pulp/apis/coin_api.py", line 137, in actualSolve
    return self.solve_CBC(lp, **kwargs)
  File "/var/lib/condor/execute/slot1/dir_27595/miniconda/lib/python3.9/site-packages/pulp/apis/coin_api.py", line 153, in solve_CBC
    vs, variablesNames, constraintsNames, objectiveName = lp.writeMPS(
  File "/var/lib/condor/execute/slot1/dir_27595/miniconda/lib/python3.9/site-packages/pulp/pulp.py", line 1782, in writeMPS
    return mpslp.writeMPS(self, filename, mpsSense=mpsSense, rename=rename, mip=mip)
  File "/var/lib/condor/execute/slot1/dir_27595/miniconda/lib/python3.9/site-packages/pulp/mps_lp.py", line 245, in writeMPS
    with open(filename, "w") as f:
FileNotFoundError: [Errno 2] No such file or directory: '/tmp/c0cf89aeb77c4e5cb38266e2421134f4-pulp.mps'

Not a high priority, but I thought I'd point it out in case others have the same issue.

Thanks,

dave

[tsibley]

Hi @dhoconno. TMPDIR is the only mechanism here for setting a different temp dir. The stack trace shows that the error you're getting arises deep from within Snakemake's scheduler, so any alternate mechanism we could add would have to ultimately set TMPDIR anyway in order to pass the value along. Glad you have a workaround, though!