HISTORY.txt 4.1 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110
  1. 2.2b1 (2007-03-31)
  2. ------------------
  3. - Individual program configuration sections can now specify an
  4. environment.
  5. - Added a 'version' command to supervisorctl. This returns the
  6. version of the supervisor2 package which the remote supervisord
  7. process is using.
  8. 2.1 (2007-03-17)
  9. ----------------
  10. - When supervisord was invoked more than once, and its configuration
  11. was set up to use a UNIX domain socket as the HTTP server, the
  12. socket file would be erased in error. The symptom of this was
  13. that a subsequent invocation of supervisorctl could not find the
  14. socket file, so the process could not be controlled (it and all of
  15. its subprocesses would need to be killed by hand).
  16. - Close subprocess file descriptors properly when a subprocess exits
  17. or otherwise dies. This should result in fewer "too many open
  18. files to spawn foo" messages when supervisor is left up for long
  19. periods of time.
  20. - When a process was not killable with a "normal" signal at shutdown
  21. time, too many "INFO: waiting for x to die" messages would be sent
  22. to the log until we ended up killing the process with a SIGKILL.
  23. Now a maximum of one every three seconds is sent up until SIGKILL
  24. time. Thanks to Ian Bicking.
  25. - Add an assertion: we never want to try to marshal None to XML-RPC
  26. callers. Issue 223 in the collector from vgatto indicates that
  27. somehow a supervisor XML-RPC method is returning None (which
  28. should never happen), but I cannot identify how. Maybe the
  29. assertion will give us more clues if it happens again.
  30. - Supervisor would crash when run under Python 2.5 because the
  31. xmlrpclib.Transport class in Python 2.5 changed in a
  32. backward-incompatible way. Thanks to Eric Westra for the bug
  33. report and a fix.
  34. - Tests now pass under Python 2.5.
  35. - Better supervisorctl reporting on stop requests that have a FAILED
  36. status.
  37. - Removed duplicated code (readLog/readMainLog), thanks to Mike
  38. Naberezny.
  39. - Added tailProcessLog command to the XML-RPC API. It provides a
  40. more efficient way to tail logs than readProcessLog(). Use
  41. readProcessLog() to read chunks and tailProcessLog() to tail.
  42. (thanks to Mike Naberezny).
  43. 2.1b1 (2006-08-30)
  44. ------------------
  45. - "supervisord -h" and "supervisorctl -h" did not work (traceback
  46. instead of showing help view (thanks to Damjan from Macedonia for
  47. the bug report).
  48. - Processes which started successfully after failing to start
  49. initially are no longer reported in BACKOFF state once they are
  50. started successfully (thanks to Damjan from Macdonia for the bug
  51. report).
  52. - Add new 'maintail' command to supervisorctl shell, which allows
  53. you to tail the 'main' supervisor log. This uses a new
  54. readMainLog xmlrpc API.
  55. - Various process-state-transition related changes, all internal.
  56. README.txt updated with new state transition map.
  57. - startProcess and startAllProcesses xmlrpc APIs changed: instead of
  58. accepting a timeout integer, these accept a wait boolean (timeout
  59. is implied by process' "startsecs" configuration). If wait is
  60. False, do not wait for startsecs.
  61. Known issues:
  62. - Code does not match state transition map. Processes which are
  63. configured as autorestarting which start "successfully" but
  64. subsequently die after 'startsecs' go through the transitions
  65. RUNNING -> BACKOFF -> STARTING instead of the correct transitions
  66. RUNNING -> EXITED -> STARTING. This has no real negative effect,
  67. but should be fixed for correctness.
  68. 2.0 (2006-08-30)
  69. ----------------
  70. - pidfile written in daemon mode had incorrect pid.
  71. - supervisorctl: tail (non -f) did not pass through proper error
  72. messages when supplied by the server.
  73. - Log signal name used to kill processes at debug level.
  74. - supervisorctl "tail -f" didn't work with supervisorctl sections
  75. configured with an absolute unix:// URL
  76. - New "environment" config file option allows you to add environment
  77. variable values to supervisord environment from config file.
  78. 2.0b1 (2006-07-12)
  79. ------------------
  80. - fundamental rewrite based on 1.0.6, use distutils (only) for
  81. installation, use ConfigParser rather than ZConfig, use HTTP for
  82. wire protocol, web interface, less lies in supervisorctl.