[Swan-dev] Pluto memory consumption

[Andrew Cagney]

On 21 March 2017 at 11:02, Paul Wouters <paul at nohats.ca> wrote:
>
> That's because since we converted to libevent, we have no way of
> listing these events to find them and delete them.

Yea, they are global in the sense that they don't seem to be tied to a
state, and reschedule themselves when they trigger:

$ awk -f ../../../../testing/utils/count-pointers.awk west.pluto.log
west.pluto.log:2052: pe at 0x7f5fda66c628 1:: | event_schedule_tv: new
EVENT_LOG_DAILY-pe at 0x7f5fda66c628
west.pluto.log:2059: pe at 0x7f5fda66c938 1:: | event_schedule_tv: new
EVENT_SD_WATCHDOG-pe at 0x7f5fda66c938
west.pluto.log:29: pe at 0x7f5fda667758 1:: | event_schedule_tv: new
EVENT_PENDING_DDNS-pe at 0x7f5fda667758
west.pluto.log:6244: pe at 0x7f5fda6750f8 1:: | event_schedule_tv: new
EVENT_SHUNT_SCAN-pe at 0x7f5fda6750f8
west.pluto.log:33: pe at 0x7f5fda667838 1:: | event_schedule_tv: new
EVENT_PENDING_PHASE2-pe at 0x7f5fda667838
west.pluto.log:25: pe at 0x7f5fda667628 1:: | event_schedule_tv: new
EVENT_REINIT_SECRET-pe at 0x7f5fda667628

(I did notice some other global events that do clean up).

I'll push the updated script so that the curious can play with it (it
might help us figure out what is going on with the certificates - a
leak or just lots of them).