Thank you very much,
This is exactly the answer I was searching for!
- Thodoris
> On 22 May 2018, at 21:10, Joseph Wu wrote:
>
> 1) DRF is based on the _current_ allocation of resources (from the master's
> perspective) rather than a historical allocation of resources.
1) DRF is based on the _current_ allocation of resources (from the master's
perspective) rather than a historical allocation of resources.
2) So when a new cluster is started, all frameworks will have a current
allocation of 0. And assuming all else (like quotas, roles, and weights)
are
Hi Hendrik,
> Is it normal that the reserved resources are only available a bit after
> the task ended?
Yes, that's normal since we don't block the forwarding of the terminal
status update behind the allocation of the freed resources. Since the
latter can take some time, we opt to forward the
Your tip helped :-) I'm seeing lines like this in the log:
mesos-master[3781]: I0926 23:36:11.292654 3805 hierarchical.cpp:927]
Framework 12da1693-ec1b-4eb6-bec9-6e6ef1b9fe11-0007 filtered slave
12da1693-ec1b-4eb6-bec9-6e6ef1b9fe11-S0 for 5secs
This is caused by Mesos doing a 5sec filter if
Thanks for the tip. No, I did not set any filters.
On 26.09.2016 09:53, Guangya Liu wrote:
This depends, it may be caused by that one of the offer is declined
and then got a filter, this will cause the filtered offer will be
offered after the expire time. Did you set filters when decline
This depends, it may be caused by that one of the offer is declined and
then got a filter, this will cause the filtered offer will be offered after
the expire time. Did you set filters when decline offer? If you are
enabling GLOG_v=2 for mesos master, you will get some log as "Framework xxx
6 matches
Mail list logo