# lsst_ci failing

XMLWordPrintable

## Details

• Type: Bug
• Status: Done
• Resolution: Done
• Fix Version/s: None
• Component/s:
• Labels:
None
• Story Points:
2
• Team:
SQuaRE

## Description

The weekly, nightly, and lsst_distrib clean build are failing due to the same error message from lsst_ci.

  lsst_ci: 15.0+15 .......................................................................................................................................................................::::: [2018-04-22T06:02:20.924442Z] ValueError: Instrument name and input dataset URL must be set in config file ::::: [2018-04-22T06:02:20.924450Z] ===================== 1 failed, 2 passed in 327.27 seconds ===================== ::::: [2018-04-22T06:02:20.955850Z] Global pytest run: failed ::::: [2018-04-22T06:02:20.976613Z] Failed test output: ::::: [2018-04-22T06:02:20.979598Z] Global pytest output is in /home/jenkins-slave/workspace/science-pipelines/lsst_distrib/centos-6.py3/lsstsw/build/lsst_ci/tests/.tests/pytest-lsst_ci.xml.failed ::::: [2018-04-22T06:02:20.979622Z] The following tests failed: ::::: [2018-04-22T06:02:20.982067Z] /home/jenkins-slave/workspace/science-pipelines/lsst_distrib/centos-6.py3/lsstsw/build/lsst_ci/tests/.tests/pytest-lsst_ci.xml.failed ::::: [2018-04-22T06:02:20.982180Z] 1 tests failed ::::: [2018-04-22T06:02:20.982695Z] scons: *** [checkTestStatus] Error 1 ::::: [2018-04-22T06:02:20.983341Z] scons: building terminated because of errors. ERROR (349 sec). 

## Activity

Hide
Simon Krughoff added a comment -

Joshua Hoblitt I believe the PR on validate_drp addresses this issue.

Show
Simon Krughoff added a comment - Joshua Hoblitt I believe the PR on validate_drp addresses this issue.
Hide
Joshua Hoblitt added a comment - - edited

Simon Krughoff The PR looks reasonable to me.

I'm duplicating my comment on the PR here in hopes that someone will explain why verify is requiring a url.

I presume this will resolve the error message ValueError: Instrument name and input dataset URL must be set in config file but I'm not sure why lsst.verify.Job is consuming a url to the data set when it is being processed as an eups product. Does that raise problems if the repo is every renamed/relocated? It also doesn't tell us what about the state of the dataset - wouldn't it be more useful to pass the eups product version or HEAD sha?

Show
Joshua Hoblitt added a comment - - edited Simon Krughoff The PR looks reasonable to me. I'm duplicating my comment on the PR here in hopes that someone will explain why verify is requiring a url. I presume this will resolve the error message ValueError: Instrument name and input dataset URL must be set in config file but I'm not sure why lsst.verify.Job is consuming a url to the data set when it is being processed as an eups product. Does that raise problems if the repo is every renamed/relocated? It also doesn't tell us what about the state of the dataset - wouldn't it be more useful to pass the eups product version or HEAD sha?
Hide
Simon Krughoff added a comment -

I agree with all that you say. We have not really pinned down how we specify exactly what data are processed for a particular validate_drp run. The first cut was to specify the URL to the gitlfs repository. This with the config file give a good notion of what was processed, but, as you point out, the repository itself could change.

I think this commit fixes the breakage of the nightly and we should give more thought to how we should do metrics provenance properly.

Show
Simon Krughoff added a comment - I agree with all that you say. We have not really pinned down how we specify exactly what data are processed for a particular validate_drp run. The first cut was to specify the URL to the gitlfs repository. This with the config file give a good notion of what was processed, but, as you point out, the repository itself could change. I think this commit fixes the breakage of the nightly and we should give more thought to how we should do metrics provenance properly.
Hide
Joshua Hoblitt added a comment -

master is once again building and DM-14178 is now resolved.

Simon Krughoff Do you want to open associated ticket(s) for verify or leave this issue open?

Show
Joshua Hoblitt added a comment - master is once again building and DM-14178 is now resolved. Simon Krughoff Do you want to open associated ticket(s) for verify or leave this issue open?
Hide
Simon Krughoff added a comment -

I want to open a new ticket. I'll just close this when I do that, if that's ok with you.

Show
Simon Krughoff added a comment - I want to open a new ticket. I'll just close this when I do that, if that's ok with you.
Hide
Joshua Hoblitt added a comment -

Works for me.

Show
Joshua Hoblitt added a comment - Works for me.
Hide
Simon Krughoff added a comment -

Triggering ticket has been created.

Show
Simon Krughoff added a comment - Triggering ticket has been created.

## People

• Assignee:
Simon Krughoff
Reporter:
Joshua Hoblitt
Watchers:
Angelo Fausti, Jonathan Sick, Joshua Hoblitt, Kian-Tat Lim, Michael Wood-Vasey, Simon Krughoff, Tim Jenness