On Nov 2, 2013, at 8:36 AM, Sven Van Caekenberghe <sven@stfx.eu> wrote:
On 02 Nov 2013, at 12:23, Johan Brichau <johan@inceptive.be> wrote:
On 02 Nov 2013, at 05:51, Marcus Denker <marcus.denker@inria.fr> wrote:
Something is fundamentally wrong here.
Yes. And it is hurting the trust people have in smalltalkhub. This is not the way to promote the community either.
I hope there are at least monitoring and auto-restart services running!
Johan
I suspect that the problem is that a database (Mongo DB) is involved so that before each (potentially automatic) restart a human has to make sure that the database is not corrupt. If the application/webapp and the DB are really very well separated then that should be less of a problem.
the problem is external semaphores saturation and errors while cleaning them. Now, the real problem is that shit always happens when there is no one available to provide a real fix, so for now we are just restarting the service each time it crashes, and we will (bah, Nico will) provide a real solution after holidays :) Esteban
That being said, the situation should definitively be improved ;-)