Details

    • Type: Improvement Improvement
    • Status: Closed
    • Priority: Major Major
    • Resolution: Fixed
    • Affects Version/s: None
    • Fix Version/s: 0.2.0
    • Component/s: debian, rpm
    • Labels:
      None

      Description

      Adding machine learning capabilities from http://mahout.apache.org/ would enhance the Bigtop platform.

      For post 0.1.0.

      1. BIGTOP-11.diff
        14 kB
        Andrew Bayer
      2. BIGTOP-11-mk2.diff
        16 kB
        Andrew Bayer
      3. BIGTOP-11-mk3.diff
        16 kB
        Andrew Bayer

        Activity

        Hide
        Peter Linnell added a comment -

        One question I have at the moment is which of the https://cwiki.apache.org/MAHOUT/mahout-collections.html need to be added.

        Show
        Peter Linnell added a comment - One question I have at the moment is which of the https://cwiki.apache.org/MAHOUT/mahout-collections.html need to be added.
        Hide
        Andrew Bayer added a comment -

        Initial packaging of Mahout in Bigtop.

        Show
        Andrew Bayer added a comment - Initial packaging of Mahout in Bigtop.
        Hide
        Bruno Mahé added a comment -

        +1 LGTM

        I guess packages build?

        Also I was wondering if the /usr/bin/mahout script generated by the install script should also contain a Apache license header?

        Show
        Bruno Mahé added a comment - +1 LGTM I guess packages build? Also I was wondering if the /usr/bin/mahout script generated by the install script should also contain a Apache license header?
        Hide
        Andrew Bayer added a comment -

        Yup, it builds and installs on both CentOS and Lucid, though I don't know if it, y'know, works. =) As to the possibility of a license header - I was aping the existing packages, so that'd probably be something to do across all packages. You should open a bug.

        Show
        Andrew Bayer added a comment - Yup, it builds and installs on both CentOS and Lucid, though I don't know if it, y'know, works. =) As to the possibility of a license header - I was aping the existing packages, so that'd probably be something to do across all packages. You should open a bug.
        Hide
        Andrew Bayer added a comment -

        Updated patch with alternatives logic added.

        Show
        Andrew Bayer added a comment - Updated patch with alternatives logic added.
        Hide
        Bruno Mahé added a comment -

        Comments for the spec file:

        • Add an empty after the license header
        • Add an empty line before the if suse_version
        • The requires need to be updated for the alternative command
        Show
        Bruno Mahé added a comment - Comments for the spec file: Add an empty after the license header Add an empty line before the if suse_version The requires need to be updated for the alternative command
        Hide
        Bruno Mahé added a comment -

        The same comments apply to the deb side

        Show
        Bruno Mahé added a comment - The same comments apply to the deb side
        Hide
        Andrew Bayer added a comment -

        What's the requires we need?

        Show
        Andrew Bayer added a comment - What's the requires we need?
        Hide
        Roman Shaposhnik added a comment -

        On debian – nothing extra is required. On RPM it seems that it depends on the OS chkconfig for RHEL/CentOS, and update-alternatives for SLES

        Show
        Roman Shaposhnik added a comment - On debian – nothing extra is required. On RPM it seems that it depends on the OS chkconfig for RHEL/CentOS, and update-alternatives for SLES
        Hide
        Andrew Bayer added a comment -

        Added /sbin/chkconfig dependency on the spec, as we do with the hadoop package. Not worrying about SUSE yet. =)

        Show
        Andrew Bayer added a comment - Added /sbin/chkconfig dependency on the spec, as we do with the hadoop package. Not worrying about SUSE yet. =)
        Hide
        Bruno Mahé added a comment -

        I do worry about it and would like to avoid having a variable set of supported os depending on the day of the week
        Furthermore your patch already contain some SUSE specifc macros, so it's confusing. And adding a require for suse is not that hard.

        Show
        Bruno Mahé added a comment - I do worry about it and would like to avoid having a variable set of supported os depending on the day of the week Furthermore your patch already contain some SUSE specifc macros, so it's confusing. And adding a require for suse is not that hard.
        Hide
        Andrew Bayer added a comment -

        Would you mind doing that as a patch after I push this one, rather than relaying directions to me? =)

        Show
        Andrew Bayer added a comment - Would you mind doing that as a patch after I push this one, rather than relaying directions to me? =)
        Hide
        Bruno Mahé added a comment -

        +1 for now since you are in a hurry. Let's fix it asap

        Show
        Bruno Mahé added a comment - +1 for now since you are in a hurry. Let's fix it asap

          People

          • Assignee:
            Andrew Bayer
            Reporter:
            Jolly Chen
          • Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

            Dates

            • Created:
              Updated:
              Resolved:

              Development