On Fri, Aug 08, 2014 at 11:45:53AM +0200, Marc Haber wrote: > I still can reproduce the behavior on a wheezy system. Here is the > output of sv stat: > > sudo sv stat /var/lib/cereal/sessions/sw01 > down: /var/lib/cereal/sessions/sw01: 82s; run: log: (pid 25393) 76s
Hi Marc, when runsv is told to exit either through "sv exit", or when receiving TERM, which also happens when the service directory symlink is removed from /etc/service/ (e.g. through update-service --remove), then from runsv(8): If the service is running, send it a TERM signal, and then a CONT signal. Do not restart the service. If the service is down, and no log service exists, runsv exits. If the service is down and a log service exists, runsv closes the standard input of the log service, and waits for it to terminate. If the log service is down, runsv exits. In your case the log service is still running, and so runsv is still waiting for it. The best solution is to make your log service detect that stdin is closed and exit as there's no more data to log. Does this fix your problem? Regards, Gerrit. PS: before runit 2.1.2 the sv(8) man page was wrong, stating that a TERM signal is sent to the log service, but actually only stdin is closed so that the log service has the chance to process everything that shall be logged. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
