Initial size of DEA / Warden

[tal....@gmail.com]

Very novice question.

Was wondering please if someone can answer or direct me to a documentation regarding the resources (memory/CPU/disk) required by a DEA and each Warden (running the VM / Linux OS). I am interested to understand DEA/warden resource overhead.

Thank you.

[James Bayer]

tal,

DEA's have:

- a ruby process for the DEA (will switch to golang with the diego initiative)

- a ruby process for warden (will switch to golang with the diego initiative)

- a golang process for the Directory Server

- a ruby process for the BOSH agent (about to switch to golang).

i believe we currently allocate approximately 1-2GB for CF process overhead and use 32GB of app memory in the hosted VM configurations.

pivotal's "Pivotal CF" has a docs page with some of the vSphere defaults here [1] that use smaller 16GB DEAs.

i hope this helps.

[1] http://docs.gopivotal.com/pivotalcf/customizing/requirements.html

To unsubscribe from this group and stop receiving emails from it, send an email to vcap-dev+u...@cloudfoundry.org.

--

Thank you,

James Bayer

[tal....@gmail.com]

James, this is very helpful - thank you!

If I may follow-up on this ... I wonder please how much CPU specifically the DEA as a whole (including JVM, control, warden, etc.) is eating up from the allocated resources (I see "2 CPU" on the chart)? ... in other words how much net CPU is left for the applications.

Thanks!

Tal

[James Bayer]

the system uses very little CPU for system processes. the system is configured to give preference to system processes rather than app container processes. app container processes are assigned a CPU share allocation proportional to the amount of memory they are configured for within max/min limits [1]. the amount of CPU required for application container processes to operate well depends completely on the application.

[1] https://github.com/cloudfoundry/dea_ng/blob/cc247d8cdbb6c94e2e877cfd09f6455c43cf0efa/spec/unit/starting/instance_spec.rb#L661

[tal....@gmail.com]

Thanks again.

I guess that as a novice the gotcha moment for me is the understanding that unlike IaaS where the application would probably be provisions with an image into its own VM running its own OS, here many application would run on the same DEA VM, and the virtual resources isolation would occur on the Ubuntu cgroups layer. Thus when considering horizontal scaling with this PaaS, the DEA, VM and OS overhead is negligible. Whereas another replica of the application with horizontal scaling on an IaaS would mean increasing resource waste and overhead.

If this is a correct analysis (and please correct me if I am wrong) its a huge issue as cost is a primary driver for cloud transformation (virtual resource waste / overhead = higher cloud costs). I would even recommend adding to the scaling page / information.

[James Bayer]

tal, yes, paas should have vastly improved density and cost savings. see james watters blog [1] on the topic.

[1] http://wattersjames.com/2013/05/09/economics-of-application-virtulization-on-aws/