HISTORY.txt 4.1 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107
  1. 2.2b1 (2007-03-31)
  2. - Individual program configuration sections can now specify an
  3. environment.
  4. - Added a 'version' command to supervisorctl. This returns the
  5. version of the supervisor2 package which the remote supervisord
  6. process is using.
  7. 2.1 (2007-03-17)
  8. - When supervisord was invoked more than once, and its configuration
  9. was set up to use a UNIX domain socket as the HTTP server, the
  10. socket file would be erased in error. The symptom of this was
  11. that a subsequent invocation of supervisorctl could not find the
  12. socket file, so the process could not be controlled (it and all of
  13. its subprocesses would need to be killed by hand).
  14. - Close subprocess file descriptors properly when a subprocess exits
  15. or otherwise dies. This should result in fewer "too many open
  16. files to spawn foo" messages when supervisor is left up for long
  17. periods of time.
  18. - When a process was not killable with a "normal" signal at shutdown
  19. time, too many "INFO: waiting for x to die" messages would be sent
  20. to the log until we ended up killing the process with a SIGKILL.
  21. Now a maximum of one every three seconds is sent up until SIGKILL
  22. time. Thanks to Ian Bicking.
  23. - Add an assertion: we never want to try to marshal None to XML-RPC
  24. callers. Issue 223 in the collector from vgatto indicates that
  25. somehow a supervisor XML-RPC method is returning None (which
  26. should never happen), but I cannot identify how. Maybe the
  27. assertion will give us more clues if it happens again.
  28. - Supervisor would crash when run under Python 2.5 because the
  29. xmlrpclib.Transport class in Python 2.5 changed in a
  30. backward-incompatible way. Thanks to Eric Westra for the bug
  31. report and a fix.
  32. - Tests now pass under Python 2.5.
  33. - Better supervisorctl reporting on stop requests that have a FAILED
  34. status.
  35. - Removed duplicated code (readLog/readMainLog), thanks to Mike
  36. Naberezny.
  37. - Added tailProcessLog command to the XML-RPC API. It provides a
  38. more efficient way to tail logs than readProcessLog(). Use
  39. readProcessLog() to read chunks and tailProcessLog() to tail.
  40. (thanks to Mike Naberezny).
  41. 2.1b1 (2006-08-30)
  42. - "supervisord -h" and "supervisorctl -h" did not work (traceback
  43. instead of showing help view (thanks to Damjan from Macedonia for
  44. the bug report).
  45. - Processes which started successfully after failing to start
  46. initially are no longer reported in BACKOFF state once they are
  47. started successfully (thanks to Damjan from Macdonia for the bug
  48. report).
  49. - Add new 'maintail' command to supervisorctl shell, which allows
  50. you to tail the 'main' supervisor log. This uses a new
  51. readMainLog xmlrpc API.
  52. - Various process-state-transition related changes, all internal.
  53. README.txt updated with new state transition map.
  54. - startProcess and startAllProcesses xmlrpc APIs changed: instead of
  55. accepting a timeout integer, these accept a wait boolean (timeout
  56. is implied by process' "startsecs" configuration). If wait is
  57. False, do not wait for startsecs.
  58. Known issues:
  59. Code does not match state transition map. Processes which are
  60. configured as autorestarting which start "successfully" but
  61. subsequently die after 'startsecs' go through the transitions
  62. RUNNING -> BACKOFF -> STARTING instead of the correct transitions
  63. RUNNING -> EXITED -> STARTING. This has no real negative effect,
  64. but should be fixed for correctness.
  65. 2.0 (2006-08-30)
  66. - pidfile written in daemon mode had incorrect pid.
  67. - supervisorctl: tail (non -f) did not pass through proper error
  68. messages when supplied by the server.
  69. - Log signal name used to kill processes at debug level.
  70. - supervisorctl "tail -f" didn't work with supervisorctl sections
  71. configured with an absolute unix:// URL
  72. - New "environment" config file option allows you to add environment
  73. variable values to supervisord environment from config file.
  74. 2.0b1 (2006-07-12)
  75. - fundamental rewrite based on 1.0.6, use distutils (only) for
  76. installation, use ConfigParser rather than ZConfig, use HTTP for
  77. wire protocol, web interface, less lies in supervisorctl.