[ClusterLabs] Informing RAs about recovery: failed resource recovery, or any start-stop cycle?

[Andrew Beekhof]

On Wed, Jun 8, 2016 at 6:23 PM, Adam Spiers <aspiers at suse.com> wrote:
> Andrew Beekhof <abeekhof at redhat.com> wrote:
>> On Wed, Jun 8, 2016 at 12:11 AM, Adam Spiers <aspiers at suse.com> wrote:
>> > Ken Gaillot <kgaillot at redhat.com> wrote:
>> >> On 06/06/2016 05:45 PM, Adam Spiers wrote:
>> >> > Adam Spiers <aspiers at suse.com> wrote:
>> >> >> Andrew Beekhof <abeekhof at redhat.com> wrote:
>> >> >>> On Tue, Jun 7, 2016 at 8:29 AM, Adam Spiers <aspiers at suse.com> wrote:
>> >> >>>> Ken Gaillot <kgaillot at redhat.com> wrote:
>> >> >>>>> My main question is how useful would it actually be in the proposed use
>> >> >>>>> cases. Considering the possibility that the expected start might never
>> >> >>>>> happen (or fail), can an RA really do anything different if
>> >> >>>>> start_expected=true?
>> >> >>>>
>> >> >>>> That's the wrong question :-)
>> >> >>>>
>> >> >>>>> If the use case is there, I have no problem with
>> >> >>>>> adding it, but I want to make sure it's worthwhile.
>> >> >>>>
>> >> >>>> The use case which started this whole thread is for
>> >> >>>> start_expected=false, not start_expected=true.
>> >> >>>
>> >> >>> Isn't this just two sides of the same coin?
>> >> >>> If you're not doing the same thing for both cases, then you're just
>> >> >>> reversing the order of the clauses.
>> >> >>
>> >> >> No, because the stated concern about unreliable expectations
>> >> >> ("Considering the possibility that the expected start might never
>> >> >> happen (or fail)") was regarding start_expected=true, and that's the
>> >> >> side of the coin we don't care about, so it doesn't matter if it's
>> >> >> unreliable.
>> >> >
>> >> > BTW, if the expected start happens but fails, then Pacemaker will just
>> >> > keep repeating until migration-threshold is hit, at which point it
>> >> > will call the RA 'stop' action finally with start_expected=false.
>> >> > So that's of no concern.
>> >>
>> >> To clarify, that's configurable, via start-failure-is-fatal and on-fail
>> >
>> > Sure.
>> >
>> >> > Maybe your point was that if the expected start never happens (so
>> >> > never even gets a chance to fail), we still want to do a nova
>> >> > service-disable?
>> >>
>> >> That is a good question, which might mean it should be done on every
>> >> stop -- or could that cause problems (besides delays)?
>> >
>> > No, the whole point of adding this feature is to avoid a
>> > service-disable on every stop, and instead only do it on the final
>> > stop.  If there are corner cases where we never reach the final stop,
>> > that's not a disaster because nova will eventually figure it out and
>> > do the right thing when the server-agent connection times out.
>> >
>> >> Another aspect of this is that the proposed feature could only look at a
>> >> single transition. What if stop is called with start_expected=false, but
>> >> then Pacemaker is able to start the service on the same node in the next
>> >> transition immediately afterward? Would having called service-disable
>> >> cause problems for that start?
>> >
>> > We would also need to ensure that service-enable is called on start
>> > when necessary.  Perhaps we could track the enable/disable state in a
>> > local temporary file, and if the file indicates that we've previously
>> > done service-disable, we know to run service-enable on start.  This
>> > would avoid calling service-enable on every single start.
>>
>> feels like an over-optimization
>> in fact, the whole thing feels like that if i'm honest.
>
> Huh ... You didn't seem to think that when we discussed automating
> service-disable at length in Austin.

I didn't feel the need to push back because RH uses the systemd agent
instead so you're only hanging yourself, but more importantly because
the proposed implementation to facilitate it wasn't leading RA writers
down a hazardous path :-)

>  What changed?  Can you suggest a
> better approach?

Either always or never disable the service would be my advice.
"Always" specifically getting my vote.

>
>> why are we trying to optimise the projected performance impact
>
> It's not really "projected"; we know exactly what the impact is.  And
> it's not really a performance impact either.  If nova-compute (or a
> dependency) is malfunctioning on a compute node, there will be a
> window (bounded by nova.conf's rpc_response_timeout value, IIUC) in
> which nova-scheduler could still schedule VMs onto that compute node,
> and then of course they'll fail to boot.

Right, but that window exists regardless of whether the node is or is
not ever coming back.
And as we already discussed, the proposed feature still leaves you
open to this window because we can't know if the expected restart will
ever happen.

In this context, trying to avoid the disable call under certain
circumstances, to avoid repeated and frequent flip-flopping of the
state, seems ill-advised.  At the point nova compute is bouncing up
and down like that, you have a more fundamental issue somewhere in
your stack and this is only one (and IMHO minor) symptom of it.

>
> The masakari folks have a lot of operational experience in this space,
> and they found that this was enough of a problem to justify calling
> nova service-disable whenever the failure is detected.

If you really want it whenever the failure is detected, call it from
the monitor operation that finds it broken.
I'm arguing that trying to do it only failure is an over optimization
and probably a mistake.

> So this
> feature has been in masakari for a long time already.  We are simply
> aiming for feature parity in a converged upstream implementation.
>
>> when the system is in terrible shape already?
>
> I'm not sure what you mean by that, but it's impossible to do anything
> about it unless you at least give some details, if not propose a
> solution ;-)  But openstack-dev@ would be a better list for that.
>
>> >> > Yes that would be nice, but this proposal was never intended to
>> >> > address that.  I guess we'd need an entirely different mechanism in
>> >> > Pacemaker for that.  But let's not allow perfection to become the
>> >> > enemy of the good ;-)
>> >>
>> >> The ultimate concern is that this will encourage people to write RAs
>> >> that leave services in a dangerous state after stop is called.
>> >
>> > I don't see why it would.
>>
>> Previous experience suggests it definitely will.
>>
>> People will do exactly what you're thinking but with something important.
>> They'll see it behaves as they expect in best-case testing and never
>> think about the corner cases.
>> Then they'll start thinking about optimising their start operations,
>> write some "optimistic" state recording code and break those too.
>>
>> Imagine a bug in your state recording code (maybe you forget to handle
>> a missing state file after reboot) that means the 'enable' does't get
>> run.  The service is up, but nova will never use it.
>>
>> > The new feature will be obscure enough that
>> > noone would be able to use it without reading the corresponding
>> > documentation first anyway.
>>
>> I like your optimism.

As a general rule, people do not read documentation.
They see an option, decide what it does based on the name and move on
if some limited testing appears to confirm their theory.

>>
>> >> I think with naming and documenting it properly, I'm fine to provide the
>> >> option, but I'm on the fence. Beekhof needs a little more convincing :-)
>> >
>> > Can you provide an example of a potential real-world situation where
>> > an RA author would end up accidentally abusing the feature?
>>
>> You want a real-world example of how someone could accidentally
>> mis-using a feature that doesn't exist yet?
>>
>> Um... if we knew all the weird and wonderful ways people break our
>> code we'd be able to build a better mouse trap.
>
> So what are you suggesting?  That we should deliberately avoid making
> any progress, based on nebulous fear of other people making stupid
> mistakes in ways that we can't even think of?

Its not nebulous, we've seen for many years now how these things get
used and how hard it is to walk back design errors.  You don't need to
maintain it or deal with the people 5 or 10 years from now that are
still getting themselves into trouble as a result, so of course you
set the bar lower.

Heck, I've seen how long it took people to stop abusing "target_rc" in
their monitor functions (hint, it wasn't voluntary -
https://github.com/beekhof/pacemaker/commit/46a65a8 ).

>  I'm totally open to
> other ideas, but I'm not hearing any yet.

I'm saying that this is not progress.
That even the usecase the feature was designed for shouldn't use it.
That the service should be unconditionally disabled on stop and
enabled on start.
That the desire for conditional enabling/disabling is rooted in the
unnecessary recovery optimization of an unrecoverable system.

Clear enough?