On Sun, Dec 04, 2011 at 04:51:57PM -0800, Tom wrote:
> >
> > I'm tempted to remove the check, I have the impression it was with
> > previous versions of redis-rb/redis.
>
> Hope you had a good weekend... I don't have any hard evidence, but the
> check was definitely valuable for us. I tried to add some logging to
> it - which failed - but it told me that redis was returning Strings at
> times, even in #keys_to_a

Hello Tom,

I had a good weekend, I hope you had a good one too.

I had a second look at #keys_to_a and noticed I added it recently. I will
keep this defensive code around.

> > Tell me if you need me to test with a redis on a distinct host and I'll
> > take the time to try (hopefully reproducing the issue).
>
> Not sure if that's going to help... and I'm still not any smarter on
> which system is the actual the problem (redis, redis-rb or ruote-
> redis)... We assume it's redis, but then again, that's just a guess.

I've pushed the "defensive mget":

  
https://github.com/jmettraux/ruote-redis/commit/0e01a2e585e554e734f35808e58b20a273248c15

I hope it helps.

For the initial issue you reported, if it reappears, I wonder how we could
harden ruote-redis to deal with that. The symptom is an empty participant
list, the redis "get" returned nil.

Maybe you could monkey-patch ruote-redis with something like to force a
second "get":

---8<---
class Ruote::Redis::Storage

  def get_configuration(key)

    2.times do
      if r = get('configurations', key)
        return r
      else
        r
      end
    end
  end
end
--->8---


Please keep me informed, cheers,

--
John Mettraux - http://lambda.io/processi

-- 
you received this message because you are subscribed to the "ruote users" group.
to post : send email to [email protected]
to unsubscribe : send email to [email protected]
more options : http://groups.google.com/group/openwferu-users?hl=en

Reply via email to