Skip to content

Conversation

@arukuka
Copy link
Member

@arukuka arukuka commented Nov 29, 2022

Summary

  • write /Db/result/*/status_ERROR when CC exited with code non-zero

Purpose

  • If CC was aborted, client could not know correct job status

Contents

  • Change wrapper script
    • Use infinite loop
    • Write /Db/result/*/status_ERROR before launching CC if status_COMPLETED or status_ERROR were not written

Testing Methods Performed

  • Run demo
    • Add code which occurs SIGSEGV in specific job

    • patch
      diff --git a/src/ComputationContainer/Job/Jobs/CorrelJob.hpp b/src/ComputationContainer/Job/Jobs/CorrelJob.hpp
      index 565b0e43..d486d4d3 100644
      --- a/src/ComputationContainer/Job/Jobs/CorrelJob.hpp
      +++ b/src/ComputationContainer/Job/Jobs/CorrelJob.hpp
      @@ -18,6 +18,8 @@ public:
               const std::vector<std::list<int>> &arg
           )
           {
      +        int *ptr = nullptr;
      +        *ptr = 1;
               std::list<std::vector<Share>> cols;
               for (const auto &w : arg[0])
               {
  • Use ShellCheck for static analysis

@arukuka arukuka force-pushed the feature/otsuka/cc_unexpected_error_handling branch from 1514b3c to 2913323 Compare November 29, 2022 11:19
If partial party exited, other parties run job forever
and cannot sync `job_id`
@arukuka arukuka merged commit 68a0c90 into main Nov 30, 2022
@arukuka arukuka deleted the feature/otsuka/cc_unexpected_error_handling branch November 30, 2022 05:09
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants