apache zookeeper - VM architecture for running HA Mesos Cluster -


i'm reading mesos architecture docs which, ironically, don't specify components supposed run on vms/physicals.

it looks like, run mesos in ha, need several categories of components:

  1. mesos masters
  2. zookeeper instances (quorum)
  3. hadoop clusters (job nodes? name nodes?)

but there's never mention of how many need of each type.

so ask: how many vms/physicals need run mesos ha, , components should deployed each?

did have @ ha docs? run mesos in ha, you'll need mesos masters , zookeeper. hadoop-related configurations out of scope mesos ha itself.

to have ha setup, you'll need uneven number of nodes masters , zookeeper (because of quorum mechanism). in our case, we're running 3 master , 3 zookeeper nodes on 3 machines (one master , 1 zookeeper instance per machine), , number of mesos slaves/agents on different machines.

theoretically, slaves/agents can run on same machines masters/zookeepers well. guess matter of preferences , availability of machines, , sla needs.

if want run large-scale production setup, make lot of sense separate master , zookeeper instances.

further references:


Comments

Popular posts from this blog

c - How to retrieve a variable from the Apache configuration inside the module? -

c# - Constructor arguments cannot be passed for interface mocks -

python - malformed header from script index.py Bad header -