Hey Vinod,
Mesos: 0.23.0-1.0.ubuntu1404
Marathon: 0.10.1-1.0.416.ubuntu1404
Chronos: 2.4.0-0.1.20150828104228.ubuntu1404
Single master, 16 slaves, each with 6cpu 6mem, many (perhaps not all) had
more resources than mentioned in the logs (mesos offers with 6m mem etc..)
or what was required for th
Can you provide more details about your setup and master logs when this
started to happen? Typically, when a framework declines an offer it sets a
timeout filter which gives mesos a chance to aggregate free resources on a
slave into one big offer.
On Tue, Sep 29, 2015 at 6:31 AM, Eren Güven wrote
Hello,
I had an incident where mesos kept offering marathon insufficient resources
which resulted in lots of forever WAITING tasks in marathon. Below are some
example logs I saw. It kept repeating the same insufficient offers even
though the cluster had a lot of free resources like ~80cpus and 45G
3 matches
Mail list logo