I've added additional syncing for handling the futures and I think that should fix the problem for now.
Carsten 2014-01-23 Carsten Ziegeler <cziege...@apache.org> > I'm not sure, if the current behaviour is optimal if the same health > checks are started in parallel - I have the feeling that some syncing is > missing. I'll try to have a look at this before the release > > Carsten > > > 2014/1/22 Carsten Ziegeler <cziege...@apache.org> > >> I've just committed a potential fix for SLING-3321 (Jira is currently >> down so I'll update the issue later). >> The result is now only put into the cache, if it's a real result - a >> timeout result is never cached. >> As soon as the future now finishes, the result is put into the cache and >> the future is removed from the list of futures >> >> Carsten >> >> >> 2014/1/22 Carsten Ziegeler <cziege...@apache.org> >> >>> Thanks, ok I'll start the vote either on Friday or next Monday. >>> >>> Carsten >>> >>> >>> 2014/1/22 Bertrand Delacretaz <bdelacre...@apache.org> >>> >>>> Hi, >>>> >>>> On Wed, Jan 22, 2014 at 2:18 AM, Carsten Ziegeler <cziege...@apache.org> >>>> wrote: >>>> > I'm planning to cut a health check core release, do we have any open >>>> issues? >>>> >>>> There's SLING-3321, I'm not happy with the current executor behavior >>>> w.r.t slow health checks and caching. >>>> >>>> We can release without fixing that right now IMO - even if we need to >>>> change the executor API to add an options parameters to the executor >>>> to better handle such cases, that would be a new method, so backwards >>>> compatible. >>>> >>>> -Bertrand >>>> >>> >>> >>> >>> -- >>> Carsten Ziegeler >>> cziege...@apache.org >>> >> >> >> >> -- >> Carsten Ziegeler >> cziege...@apache.org >> > > > > -- > Carsten Ziegeler > cziege...@apache.org > -- Carsten Ziegeler cziege...@apache.org
