Page MenuHomePhabricator

bin/phd status lists long dead daemons
Closed, InvalidPublic

Description

After a reboot bin/phd status lists currently running daemons as well as ones that have died a while ago:

➜  phabricator git:(master) ✗ bin/phd status
Log  Daemon       Host      Overseer Started                  Class                                Arguments
1429 1687:zwxufis localhost 1687     Jun 24 2016, 10:59:24 PM PhabricatorRepositoryPullLocalDaemon
1430 1687:5mz3yqr localhost 1687     Jun 24 2016, 10:59:24 PM PhabricatorTriggerDaemon
1431 1687:7e3z7oo localhost 1687     Jun 24 2016, 10:59:24 PM PhabricatorTaskmasterDaemon
1428 2769:rgxdnog colors    2769     Jun 24 2016, 10:47:55 PM PhabricatorTaskmasterDaemon
1427 2769:xxfjvha colors    2769     Jun 24 2016, 10:47:55 PM PhabricatorTriggerDaemon
1426 2769:j72ubcn colors    2769     Jun 24 2016, 10:47:55 PM PhabricatorRepositoryPullLocalDaemon
1425 3458:zystifg colors    3458     Jun 24 2016, 9:55:39 PM  PhabricatorTaskmasterDaemon
1424 3458:ppuwtnz colors    3458     Jun 24 2016, 9:55:39 PM  PhabricatorTriggerDaemon
1423 3458:d7wu6n4 colors    3458     Jun 24 2016, 9:55:39 PM  PhabricatorRepositoryPullLocalDaemon
➜  phabricator git:(master) ✗ ps aux | grep phd
phduser   1687  0.0  0.3 240896 26416 pts/0    S    15:59   0:00 php ./phd-daemon
phduser   1692  0.1  0.4 232448 33912 ?        Ss   15:59   0:00 php ./exec_daemon.php PhabricatorRepositoryPullLocalDaemon
phduser   1696  0.1  0.4 234628 35712 ?        Ss   15:59   0:00 php ./exec_daemon.php PhabricatorTriggerDaemon
phduser   1702  0.1  0.4 232448 33668 ?        Ss   15:59   0:00 php ./exec_daemon.php PhabricatorTaskmasterDaemon
adminis+  4260  0.0  0.0  14224   976 pts/0    S+   16:05   0:00 grep --color=auto --exclude-dir=.bzr --exclude-dir=CVS --exclude-dir=.git --exclude-dir=.hg --exclude-dir=.svn phd

The dead ones are listed with the host "colors", which is the hostname of the machine they're running on.

If I stop the daemons the top 3 disappear, but the rest remain, which causes the front end to never report that the daemons are not running.

Event Timeline

epriestley added a subscriber: epriestley.

Bug reports MUST include reproduction steps that allow us to reproduce the issue locally. Without this information, we can not move forward.

See Contributing Bug Reports and Providing Reproduction Steps for help writing an actionable bug report.

Ok, I just installed a new instance of the server in another VM.
Steps to reproduce, as far as I can tell, are:

  1. create a new phd-scpecific user (call it phduser)
  2. set up phd to run under that user
  3. bin/phd start
  4. verify all is good
  5. sudo reboot
  6. once booted, bin/phd start
  7. now bin/phd status lists 6 daemons instead of 3 and the later 3 are not actually alive.

eventually the dead ones appear to fall off the edge, but in the meantime UI doesn't give me a warning that my daemons are not actually running.