CHANGES.txt 12 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286
  1. 3.0a2
  2. - Fixed the README.txt example for defining the supervisor RPC
  3. interface in the configuration file. Thanks to Drew Perttula.
  4. - Fixed a bug where process communication events would not have the
  5. proper payload if the payload data was very short.
  6. - when supervisord attempted to kill a process with SIGKILL after
  7. the process was not killed within "stopwaitsecs" using a "normal"
  8. kill signal, supervisord would crash with an improper
  9. AssertionError. Thanks to Calvin Hendryx-Parker.
  10. - On Linux, Supervisor would consume too much CPU in an effective
  11. "busywait" between the time a subprocess exited and the time at
  12. which supervisor was notified of its exit status. Thanks to Drew
  13. Perttula.
  14. - RPC interface behavior change: if the RPC method
  15. "sendProcessStdin" is called against a process that has closed its
  16. stdin file descriptor (e.g. it has done the equivalent of
  17. "sys.stdin.close(); os.close(0)"), we return a NO_FILE fault
  18. instead of accepting the data.
  19. - Changed the semantics of the process configuration 'autorestart'
  20. parameter with respect to processes which move between the RUNNING
  21. and EXITED state. 'autorestart' was previously a boolean. Now
  22. it's a trinary, accepting one of 'false', 'unexpected', or 'true'.
  23. If it's 'false', a process will never be automatically restarted
  24. from the EXITED state. If it's 'unexpected', a process that
  25. enters the EXITED state will be automatically restarted if it
  26. exited with an exit code that was not named in the process
  27. config's 'exitcodes' list. If it's 'true', a process that enters
  28. the EXITED state will be automatically restarted unconditionally.
  29. The default is now 'unexpected' (it was previously 'true'). The
  30. readdition of this feature is a reversion of the behavior change
  31. note in the changelog notes for 3.0a1 that asserted we never cared
  32. about the process' exit status when determining whether to restart
  33. it or not.
  34. - setup.py develop (and presumably setup.py install) would fail
  35. under Python 2.3.3, because setuptools attempted to import
  36. 'splituser' from urllib2, and it didn't exist.
  37. - It's now possible to use 'setup.py install' and 'setup.py develop'
  38. on systems which do not have a C compiler if you set the environment
  39. variable "NO_MELD3_EXTENSION_MODULES=1" in the shell in which
  40. you invoke these commands (a new version of meld3, 0.6.1 was released
  41. which respects this envvar and does not try to compile its optional
  42. C extensions when it's set).
  43. - The test suite would fail on Python versions <= 2.3.3 because
  44. the "assertTrue" and "assertFalse" methods of unittest.TestCase
  45. didn't exist in those versions.
  46. 3.0a1
  47. - Default config file comment documented 10 secs as default for
  48. 'startsecs' value in process config, in reality it was 1 sec.
  49. Thanks to Christoph Zwerschke.
  50. - Make note of subprocess environment behavior in README.txt.
  51. Thanks to Christoph Zwerschke.
  52. - New "strip_ansi" config file option attempts to strip ANSI escape
  53. sequences from logs for smaller/more readable logs (submitted by
  54. Mike Naberezny).
  55. - The XML-RPC method supervisor.getVersion() has been renamed for
  56. clarity to supervisor.getAPIVersion(). The old name is aliased
  57. for compatibility but is deprecated and will be removed in a
  58. future version (Mike Naberezny).
  59. - Improved web interface styling (Mike Naberezny, Derek DeVries)
  60. - The XML-RPC method supervisor.startProcess() now checks that
  61. the file exists and is executable (Mike Naberezny).
  62. - Two environment variables, "SUPERVISOR_PROCESS_NAME" and
  63. "SUPERVISOR_PROCESS_GROUP" are set in the environment of child
  64. processes, representing the name of the process and group in
  65. supervisor's configuration.
  66. - Process state map change: a process may now move directly from the
  67. STARTING state to the STOPPING state (as a result of a stop
  68. request).
  69. - Behavior change: if 'autorestart' is true, even if a process exits
  70. with an "expected" exit code, it will still be restarted. In the
  71. immediately prior release of supervisor, this was true anyway, and
  72. no one complained, so we're going to consider that the "officially
  73. correct" behavior from now on.
  74. - Supervisor now logs subprocess stdout and stderr independently.
  75. The old program config keys "logfile", "logfile_backups" and
  76. "logfile_maxbytes" are superseded by "stdout_logfile",
  77. "stdout_logfile_backups", and "stdout_logfile_maxbytes". Added
  78. keys include "stderr_logfile", "stderr_logfile_backups", and
  79. "stderr_logfile_maxbytes". An additional "redirect_stderr" key is
  80. used to cause program stderr output to be sent to its stdin
  81. channel. The keys "log_stderr" and "log_stdout" have been
  82. removed.
  83. - '[program:x]' config file sections now represent "homgeneous
  84. process groups" instead of single processes. A "numprocs" key in
  85. the section represents the number of processes that are in the
  86. group. A "process_name" key in the section allows composition of
  87. the each process' name within the homogeneous group.
  88. - A new kind of config file section, '[group:x]' now exists,
  89. allowing users to group heterogeneous processes together into a
  90. process group that can be controlled as a unit from a client.
  91. - Supervisord now emits "events" at certain points in its normal
  92. operation. These events include supervisor state change events,
  93. process state change events, and "process communication events".
  94. - A new kind of config file section '[eventlistener:x]' now exists.
  95. Each section represents an "event listener pool", which is a
  96. special kind of homogeneous process group. Each process in the
  97. pool is meant to receive supervisor "events" via its stdin and
  98. perform some notification (e.g. send a mail, log, make an http
  99. request, etc.)
  100. - Supervisord can now capture data between special tokens in
  101. subprocess stdout/stderr output and emit a "process communications
  102. event" as a result.
  103. - Supervisor's XML-RPC interface may be extended arbitrarily by
  104. programmers. Additional top-level namespace XML-RPC interfaces
  105. can be added using the '[rpcinterface:foo]' declaration in the
  106. configuration file.
  107. - New 'supervisor'-namespace XML-RPC methods have been added:
  108. getAPIVersion (returns the XML-RPC API version, the older
  109. "getVersion" is now deprecated), "startProcessGroup" (starts all
  110. processes in a supervisor process group), "stopProcessGroup"
  111. (stops all processes in a supervisor process group), and
  112. "sendProcessStdin" (sends data to a process' stdin file
  113. descriptor).
  114. - 'supervisor'-namespace XML-RPC methods which previously accepted
  115. ony a process name as "name" (startProcess, stopProcess,
  116. getProcessInfo, readProcessLog, tailProcessLog, and
  117. clearProcessLog) now accept a "name" which may contain both the
  118. process name and the process group name in the form
  119. 'groupname:procname'. For backwards compatibility purposes,
  120. "simple" names will also be accepted but will be expanded
  121. internally (e.g. if "foo" is sent as a name, it will be expanded
  122. to "foo:foo", representing the foo process within the foo process
  123. group).
  124. - 2.X versions of supervisorctl will work against supervisor 3.0
  125. servers in a degraded fashion, but 3.X versions of supervisorctl
  126. will not work at all against supervisor 2.X servers.
  127. Known issues:
  128. - supervisorctl and the web interface do not yet allow you to stop
  129. / start / restart a process group as a unit.
  130. - supervisorctl and the web interface do not allow you to tail or
  131. otherwise examine stderr log files of processes.
  132. - buffered event notifications may be lost at supervisor shutdown
  133. or restart time.
  134. Acknowledgements:
  135. Maintainable Software (http://www.maintainable.com) contracted
  136. Agendless Consulting to add the event notification features and
  137. extensible XML-RPC namespaces feature to supervisor.
  138. 2.2b1
  139. - Individual program configuration sections can now specify an
  140. environment.
  141. - Added a 'version' command to supervisorctl. This returns the
  142. version of the supervisor2 package which the remote supervisord
  143. process is using.
  144. 2.1
  145. - When supervisord was invoked more than once, and its configuration
  146. was set up to use a UNIX domain socket as the HTTP server, the
  147. socket file would be erased in error. The symptom of this was
  148. that a subsequent invocation of supervisorctl could not find the
  149. socket file, so the process could not be controlled (it and all of
  150. its subprocesses would need to be killed by hand).
  151. - Close subprocess file descriptors properly when a subprocess exits
  152. or otherwise dies. This should result in fewer "too many open
  153. files to spawn foo" messages when supervisor is left up for long
  154. periods of time.
  155. - When a process was not killable with a "normal" signal at shutdown
  156. time, too many "INFO: waiting for x to die" messages would be sent
  157. to the log until we ended up killing the process with a SIGKILL.
  158. Now a maximum of one every three seconds is sent up until SIGKILL
  159. time. Thanks to Ian Bicking.
  160. - Add an assertion: we never want to try to marshal None to XML-RPC
  161. callers. Issue 223 in the collector from vgatto indicates that
  162. somehow a supervisor XML-RPC method is returning None (which
  163. should never happen), but I cannot identify how. Maybe the
  164. assertion will give us more clues if it happens again.
  165. - Supervisor would crash when run under Python 2.5 because the
  166. xmlrpclib.Transport class in Python 2.5 changed in a
  167. backward-incompatible way. Thanks to Eric Westra for the bug
  168. report and a fix.
  169. - Tests now pass under Python 2.5.
  170. - Better supervisorctl reporting on stop requests that have a FAILED
  171. status.
  172. - Removed duplicated code (readLog/readMainLog), thanks to Mike
  173. Naberezny.
  174. - Added tailProcessLog command to the XML-RPC API. It provides a
  175. more efficient way to tail logs than readProcessLog(). Use
  176. readProcessLog() to read chunks and tailProcessLog() to tail.
  177. (thanks to Mike Naberezny).
  178. 2.1b2
  179. - Added new tailProcessLog() command to the XML-RPC API that
  180. is more efficient for just tailing than the existing
  181. readProcessLog() command (Mike Naberezny).
  182. 2.1b1
  183. - "supervisord -h" and "supervisorctl -h" did not work (traceback
  184. instead of showing help view (thanks to Damjan from Macedonia for
  185. the bug report).
  186. - Processes which started successfully after failing to start
  187. initially are no longer reported in BACKOFF state once they are
  188. started successfully (thanks to Damjan from Macdonia for the bug
  189. report).
  190. - Add new 'maintail' command to supervisorctl shell, which allows
  191. you to tail the 'main' supervisor log. This uses a new
  192. readMainLog xmlrpc API.
  193. - Various process-state-transition related changes, all internal.
  194. README.txt updated with new state transition map.
  195. - startProcess and startAllProcesses xmlrpc APIs changed: instead of
  196. accepting a timeout integer, these accept a wait boolean (timeout
  197. is implied by process' "startsecs" configuration). If wait is
  198. False, do not wait for startsecs.
  199. Known issues:
  200. Code does not match state transition map. Processes which are
  201. configured as autorestarting which start "successfully" but
  202. subsequently die after 'startsecs' go through the transitions
  203. RUNNING -> BACKOFF -> STARTING instead of the correct transitions
  204. RUNNING -> EXITED -> STARTING. This has no real negative effect,
  205. but should be fixed for correctness.
  206. 2.0
  207. - pidfile written in daemon mode had incorrect pid.
  208. - supervisorctl: tail (non -f) did not pass through proper error
  209. messages when supplied by the server.
  210. - Log signal name used to kill processes at debug level.
  211. - supervisorctl "tail -f" didn't work with supervisorctl sections
  212. configured with an absolute unix:// URL
  213. - New "environment" config file option allows you to add environment
  214. variable values to supervisord environment from config file.
  215. 2.0b1
  216. - fundamental rewrite based on 1.0.6, use distutils (only) for
  217. installation, use ConfigParser rather than ZConfig, use HTTP for
  218. wire protocol, web interface, less lies in supervisorctl.