3 * dinitctl: Warn when stopping a service would cause a dependent to stop. Require a force flag to
5 * dinitctl: implement "restart" command [done]
6 * Get rid of "floating" service state [done]
10 * Easy way to reload service description (including if service is running, where possible).
11 * provide a way to check configuration for errors (maybe a dinitctl command)
12 - eg. check a service and all dependencies.
13 * report process launch failure reason (stage & errno) via dinitctl.
14 * "chain-to" can result in an unbreakable loop if the chain is circular. Chained services should not be
15 started during shutdown to prevent this (also avoids a race condition where the chained service is left
16 running when everything else has shutdown).
17 * Show "activated" state in "dinitctl list" output
21 * Limit memory use by control connections. Currently clients can queue commands without limit.
22 * Consider using mlockall (if system process).
23 * Dinitctl command to get full status of a service.
24 * "triggered" service type: external process notifies Dinit when the service
26 - key thing is we want some way to eg mount filesystem once the disk comes up,
27 configure network when device comes up, etc, potentially relying an an external
29 * on shutdown, after repeated intervals with no activity, display information
30 about services we are waiting on (or, do this when prompted via ^C or C-A-D).
31 * Documentation must be complete (see section below).
32 * Proper support for socket activation?
33 * Chaining of service process input/output?
34 * Be able to boot and shutdown Linux and FreeBSD (or OpenBSD).
38 * On linux when running with PID != 1, write PID to /proc/sys/kernel/cad_pid so
39 that we still receive SIGINT from ctrl+alt+del (must be done after /proc is
40 mounted, possibly could be left to a service script)
41 * Perhaps need a way to prevent script services from re-starting.
42 (eg there's no need to mount filesystems twice; there might be various other
43 system initialisations that can't or shouldn't really be "undone" and so do
44 not need to be re-done).
45 * Internationalisation
46 * A service can prevent shutdown/reboot by failing to stop. Maybe make
47 multiple CTRL-ALT-DEL presses (or ^C since that's more portable) commence
48 immediate shutdown (or launch a simple control interface).
49 * When we take down a service or tty session, it would be ideal if we could kill
50 the whole process tree, not just the leader process (need cgroups or pid
51 namespace or other mechanism).
52 * Allow logging tasks to memory (growing or circular buffer) and later
53 switching to disk logging (allows for filesystem mounted readonly on boot).
54 But perhaps this really the responsibility of another daemon.
55 * Allow running services with different resource limits, chroot, cgroups,
56 namespaces (pid/fs/uid), etc
57 * Support chaining service output to another process (logger) input; if the
58 service dies the file descriptor of its stdout isn't closed and is reassigned
59 when the service is restarted, so that minimal output is lost.
60 - even more, it would be nice if a single logger process could be responsible
61 for receiving output from multiple services. This would require some kind of
62 protocol for passing new output descriptors to the logger (for when a
65 Even later / Maybe never:
66 -------------------------
67 * Support recognising /etc/init.d services automatically (as script services, with
68 no dependency management - or upstart compatible dependency management)
69 Also BSD's rc.d style scripts (PROVIDE, REQUIRE).
70 * Place some reasonable, soft limit on the number of services to be started
71 simultaneously, to prevent thrashing. Services that are taking a long time
72 to start don't count to the limit. Maybe use CPU/IO usage as a controlling
74 * Cron-like tasks (if started, they run a sub-task periodically. Stopping the
75 task will wait until the sub-task is complete).
76 * Allow to run services attached to virtual tty, allow connection to that tty (ala "screen").
77 * SystemD-like handling of filesystem mounts (see autofs documentation in kernel)
78 i.e. a mount point gets an autofs attached, and lazily gets mounted when accessed
79 (or is mounted in parallel). Probably put the functionality in a separate daemon.
84 * Design philosophy/rationale document
85 * More system integration documentation?