Re: 3.18.1: duplicate cf-exexd processes

On Mon, 2022-03-14 at 01:26 -0700, Xander Cage wrote:
> before 3.18 cf-execd runs the agent in an thread, now cf-execd replicates itself in a new process
> (hence duplicate processes) and starts the agent from this process, at least in my understanding.
And it is correct. cf-execd forks (duplicates) itself and the forked (child) process spawns cf-agent
and terminates. However, terminated processes need to be reaped by their parent processes and the
original (parent) cf-execd process only does that once per minute. So every time cf-execd spawns
cf-agent, there's a dead/inactive/zombie/defunct cf-execd for a minute.

Hope the bit of extra explanation helps.

--
Vratislav

> 
> On Sunday, March 13, 2022 at 9:57:00 PM UTC+1 brontolinux wrote:
> > Hello there
> > I am hitting a related problem in Debian 11.
> > I have just ported by policies from 3.15.5 to 3.18.1. I got both CFEngine and systemd
> > complaining about the duplicated process so I applied Nick's patch, and now I have a different
> > problem.
> > I have not completely understood what's going on. Is it cf-agent that starts a new cf-execd? Or
> > is it cf-execd that clones itself and then spawns cf-agent, but then doesn't terminate? Whatever
> > it is, when the new agent run starts, one cf-execd is killed (by what?), then the agent starts,
> > and then systemd complains again about the duplicated cf-execd... I am trying to figure out, but
> > my brain is already in bed tonight.
> > Anyone who knows better about the changes around cf-execd in 3.18 and understands what the heck
> > is going on here? And, especially, how to fix this?
> > 
> > > Mar 13 21:43:54 minardi systemd[1]: cf-execd.service: Main process exited, code=killed,
> > > status=9/KILL
> > > Mar 13 21:43:54 minardi systemd[1]: cf-execd.service: Failed with result 'signal'.
> > > Mar 13 21:43:54 minardi systemd[1]: cf-execd.service: Unit process 60482 (cf-execd) remains
> > > running after unit stopped.
> > > Mar 13 21:43:54 minardi systemd[1]: cf-execd.service: Unit process 60483 (sh) remains running
> > > after unit stopped.
> > > Mar 13 21:43:54 minardi systemd[1]: cf-execd.service: Unit process 60503 (cf-agent) remains
> > > running after unit stopped.
> > > Mar 13 21:43:54 minardi systemd[1]: cf-execd.service: Consumed 1.083s CPU time.
> > > Mar 13 21:43:54 minardi cf-agent[60503]: CFEngine(agent)  R: [INFO
> > > abort_agent_if_disabled/methods] The agent is starting. To disable, create flag file
> > > /etc/cfengine/disable
> > > Mar 13 21:43:54 minardi cf-agent[60503]: CFEngine(agent)  R: [INFO main/methods] Promises
> > > version 5.28.2 from branch master
> > > Mar 13 21:43:54 minardi cf-agent[60503]: CFEngine(agent)  R: [INFO main/methods] Our policy
> > > server is 52.211.82.124
> > > Mar 13 21:43:55 minardi cf-agent[60503]: CFEngine(agent)  R: [INFO apply_roles/methods]
> > > Applying roles: standard_node, cfdevnode, home_pc.
> > > Mar 13 21:44:00 minardi cf-agent[60503]: CFEngine(agent)  R: [INFO main/methods] Custom
> > > configuration completed
> > > Mar 13 21:44:04 minardi systemd[1]: cf-execd.service: Scheduled restart job, restart counter
> > > is at 2.
> > > Mar 13 21:44:04 minardi systemd[1]: Stopped CFEngine Execution Scheduler.
> > > Mar 13 21:44:04 minardi systemd[1]: cf-execd.service: Consumed 4.924s CPU time.
> > > Mar 13 21:44:04 minardi systemd[1]: cf-execd.service: Found left-over process 60518 (cf-execd)
> > > in control group while starting unit. Ignoring.
> > > Mar 13 21:44:04 minardi systemd[1]: This usually indicates unclean termination of a previous
> > > run, or service implementation deficiencies.
> > > Mar 13 21:44:04 minardi systemd[1]: Started CFEngine Execution Scheduler.
> > Ciao!
> >  -- bronto
> > 
> -- 
> You received this message because you are subscribed to the Google Groups "help-cfengine" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to
> [email protected].
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/help-cfengine/fd7a9d2e-dca5-4f48-b5c8-17d070e3dec1n%40googlegroups.com
> .

-- 
You received this message because you are subscribed to the Google Groups "help-cfengine" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
To view this discussion on the web visit https://groups.google.com/d/msgid/help-cfengine/303ee5c164550184fbb9b5b3d2981094f4ee6d0c.camel%40northern.tech.