Hadoop HDFS
  1. Hadoop HDFS
  2. HDFS-3271

src/fuse_users.c: use re-entrant versions of getpwuid, getgid, etc

    Details

    • Type: Improvement Improvement
    • Status: Open
    • Priority: Minor Minor
    • Resolution: Unresolved
    • Affects Version/s: None
    • Fix Version/s: None
    • Component/s: None
    • Labels:
      None

      Description

      Use the re-entrant versions of these functions rather than using locking

        Issue Links

          Activity

          Hide
          Todd Lipcon added a comment -

          See HADOOP-7156. the reentrant functions are broken on some OSes so we switched to locking and non-reentrant calls.

          Show
          Todd Lipcon added a comment - See HADOOP-7156 . the reentrant functions are broken on some OSes so we switched to locking and non-reentrant calls.
          Hide
          Colin Patrick McCabe added a comment -

          I guess we could make FUSE respect that configuration option as well.

          Show
          Colin Patrick McCabe added a comment - I guess we could make FUSE respect that configuration option as well.
          Hide
          Allen Wittenauer added a comment -

          Or test for POSIX compliance at configure time... (FWIW, I'm still mostly convinced that this ugly hack was a waste of time for the majority of folks.)

          Show
          Allen Wittenauer added a comment - Or test for POSIX compliance at configure time... (FWIW, I'm still mostly convinced that this ugly hack was a waste of time for the majority of folks.)
          Hide
          Todd Lipcon added a comment -

          Or test for POSIX compliance at configure time...

          Testing for presence of a race is tricky.

          (FWIW, I'm still mostly convinced that this ugly hack was a waste of time for the majority of folks.)

          Sure, but for the folks who needed it, it saved their clusters from constant segfaults.

          Show
          Todd Lipcon added a comment - Or test for POSIX compliance at configure time... Testing for presence of a race is tricky. (FWIW, I'm still mostly convinced that this ugly hack was a waste of time for the majority of folks.) Sure, but for the folks who needed it, it saved their clusters from constant segfaults.
          Hide
          Colin Patrick McCabe added a comment -

          > > Or test for POSIX compliance at configure time...

          > Testing for presence of a race is tricky.

          We would have to check the library version in the configure script. This is actually pretty trivial in automake.

          Show
          Colin Patrick McCabe added a comment - > > Or test for POSIX compliance at configure time... > Testing for presence of a race is tricky. We would have to check the library version in the configure script. This is actually pretty trivial in automake.
          Hide
          Todd Lipcon added a comment -

          It's not a bug in the library, but rather a bug in one of the nss backends (sssd). Plus, that would require re-building the native libs on every different version of EL6, whereas right now a single binary works against any EL6 release.

          Show
          Todd Lipcon added a comment - It's not a bug in the library, but rather a bug in one of the nss backends (sssd). Plus, that would require re-building the native libs on every different version of EL6, whereas right now a single binary works against any EL6 release.
          Hide
          Colin Patrick McCabe added a comment -

          Since we already have a configuration knob for using the re-entrant functions in the .xml files, that seems like the logical thing to use. Yes, it will be a bit of a pain to get it over to the FUSE side, but not impossible.

          For now, I'll wait for some before/after performance numbers before doing anything on this...

          Show
          Colin Patrick McCabe added a comment - Since we already have a configuration knob for using the re-entrant functions in the .xml files, that seems like the logical thing to use. Yes, it will be a bit of a pain to get it over to the FUSE side, but not impossible. For now, I'll wait for some before/after performance numbers before doing anything on this...

            People

            • Assignee:
              Colin Patrick McCabe
              Reporter:
              Colin Patrick McCabe
            • Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

              Dates

              • Created:
                Updated:

                Development