[systemd-devel] forever loop during garbage collection

[Lennart Poettering]

On Wed, 10.12.14 15:22, Umut Tezduyar Lindskog (umut at tezduyar.com) wrote:

Sorry for the late reply.

> On Mon, Dec 8, 2014 at 8:09 PM, Lennart Poettering
> <lennart at poettering.net> wrote:
> > On Sun, 30.11.14 14:38, Umut Tezduyar Lindskog (umut at tezduyar.com) wrote:
> >
> >> Hi,
> >>
> >> We are experiencing an unbreakable loop in manager_dispatch_gc_queue.
> >> Problem happens when systemd runs in sysV compatibility mode (Porky
> >> enables this).
> >>
> >> Seems like manager_dispatch_gc_queue's while loop gets stuck and seems
> >> like unit_gc_sweep cannot make a decision about the unit. As a result,
> >> it marks the unit with offset_unsure and adds the unit back to gc
> >> queue.
> >>
> >> If I am reading the code correctly recursive unit_gc_sweep will never
> >> be able to remove the unit from the gc queue if it is referenced by
> >> another unit and if another unit is referenced by the unit.
> >>
> >> A is referenced by B
> >> B is referenced by A
> >
> > So in this case first A will be processed by the GC sweep, it will
> > follow the link to B while setting the state to IN_PATH and invoke the
> > GC sweep on that. B will then be set to IN_PATH too. GC sweep now
> > follows its link back, and up at A again, but this time return quickly
> > because its state is set to IN_PATH. Due to this, it will then set B's
> > state to UNSURE, and return to A, which in effect will now be set to
> > UNSURE too. Now, we return into GC queue dispatch call, which will
> > notice that it is UNSURE and uprgade that to BAD, and kill it because
> > there's nothin in the unit's dependency network that is clearly a
> > GOOD, and hence should be removed.
> >
> > The essence of cycle breaking here is really in
> > manager_dispatch_gc_queue() which uprgades UNSURE to BAD in the end. I
> > am not seeing how this could end up in an endless loop hence.
> 
> I have debugged it more and as you have said there is no bug in code
> but it takes so long to go out of unit_gc_sweep I thought there is a
> forever loop.
> 
> Attached is my patch on 216 and
> https://drive.google.com/file/d/0B_uiALgWpGXtZ0VidURxSnVhcDA/view?usp=sharing
> is a part of the log after patch.
> 
> It has been 3 hours since I issued "systemctl isolate" and according
> to the logs I can see that garbage collection logic is making it's way
> back up. I guess it will eventually resolve itself but after so many
> hours.

Hmm, so, you mean the code works correctly but scales really badly?
How many units do you have?

> 
> (Search for "-                     -" and it is happening every 300.000 lines)
> 
> Problem seemed to be introduced on "95ed329" - Move handling of sysv
> initscripts to a generator.

Hmm, how precisely do the deps look like the generator creates for
you?

Any chance you can run
"/usr/lib/systemd/system-generators/systemd-sysv-generator /tmp/foo
/tmp/foo /tmp/foo", and check what deps it precisely generates in
/tmp/foo?

I have never seen that the GC scales this badly...

> This is totally due to how sysV generator is linking services but I
> think slowness on GC can happen on a complex system with many units
> linked with each other.
> 
> Thoughts?

I am puzzled, quite frankly...

Lennart

-- 
Lennart Poettering, Red Hat