I would like to have RStudio running on my local machine (os x) and the R executable on a remote computer.
I'm aware that I could run RStudio Server on the remote machine and connect to it using a web interface, however I abhor using web interfaces for things like this due to the delays and limited ability to move windows and use shortcuts.
Since RStudio isn't stand alone but points to an R executable elsewhere on the local machine (in a location that can be varied), it seems like in theory this pointer could be to a remote location. (Is there some reason this is incorrect?)
How might I accomplish this?
We're writing software for an embedded platform for which we also need automated on-device testing.
When a push/tagging is made to the repository a GitLab pipeline is started. The first step is to build the software.
The next step is to execute a custom written Python script (doing the testing) on a machine running locally in our office. This machine is physically connected to the embedded device via USB, hence the testing-part of the CI pipeline needs to be executes on this machine exclusively. The Python script needs to be executes multiple times with different parameters, but may (and should) run concurrently if possible.
Is this setup possible with GitLab runners, and if so, how do I configure it to only run this exact part of the pipeline locally?
Thanks in advance
I have set up a two node NFS cluster for running Open MPI applications.
One of them has Intel processors (the master) while the other one has AMD processors. Both are quadcore.
I have installed Open MPI on the same location on both systems (in /use/local)
I am trying to run a simple Hello World example with 4 processes(two on each).
I have tried and tested running these applications locally on both nodes. They run perfectly well.
However, if I pass on a hostfile consisting of these nodes while running (I use mpiexec) it throws errors.
If I compile on Intel, the job does not run on AMD node.
If I compile on AMD, the job does not run on Intel node.
Where could be the problem?
I wanted to launch an instance with high availability with out having risk factor i.e, an instance will be launched in multiple regions(zones) that to sync the state like database(master-slave). When some applications got installed, same should reflect in another region/zone also(mostly image format). Can we do that?.
I have checked some links based on this. I got a confusion after reading all the docs.
Host-aggregate/Cell in openstack
Nova evacuate command
Buildbot tool
Exactly what is the difference among. VM replication & syncing is possiblein Openstack?
To the best of my knowledge, Open Stack does not support VM replication for now.
There is a component called Remus under the Xen project, which could potentially used by manual configuration as Open Stack supports Xen (https://www.xenproject.org/directory/directory/projects/70-remus.html). But it seems to be slow and unstable.
The newest approach is called reversed virtual machine replication (http://dl.acm.org/citation.cfm?id=2996894&CFID=918229768&CFTOKEN=85577813), this one seems to be very interesting and some critical problems in VM replication is well defined and elegantly solved. However, I did not find the open source project for it.
I have some questions:
Is it possible to install openstack on a Notebook with a 4GB DD3 Ram? Because the website says it needs atleast 8GB of RAM.
They say it requirs a double-QuadCore , I assue that means Octacore. Can we install that on a Quadcore?
They say that there is no possibility to install it on a NAS . Did you find any where if there is a possibility to do?. I dint find any even after asking our friend(google).
All in all, is it at-all possible to install on it a notebook/Desktop?
That advice is for production environments,
so 1)If you just want to play around your notebook will do fine. I had a succesful test-run on a 1.2 Ghz 1GB Netbook. It became incredibly slow when it launched it's first instance...
With a Double Quadcore they actually mean two seperate Quad-cores, as in two quad-core xeon processors on a single motherboard
So 2) yes you can install it on a quad-core.
3) a NAS device running openstack an openstack storage service seems to be unlikely indeed. You will most likely need more computing power.However If your NAS supports NFS or SSH or sth you can probably mount this drive and use it for storage.
4) You can perfectly build a all-in-one openstack test setup on your notebook. Performance will be low, but acceptable for testing.
It depends on what you mean by "install OpenStack". OpenStack itself is an extremely modular framework consisting on many services (Compute, Networking, Image service, Block Storage, Object Storage, Orchestration, Telemetry, ...). On top of that, a typical production deployment of OpenStack also requires several components, like load balancers, caching systems, firewalls, web servers and others. It is definitely possible to install a minimal openstack system, even on an average laptop.
The simplest way to run OpenStack on a laptop/desktop is to use Devstack, a shell script that installs all services from source and run them (by default) on a single machine. It is customizable enough to provide very good testing ground; it's used by OpenStack developers as well as the OpenStack QA team to test latest developments against "real" systems.
To avoid messing up your system, it's generally recommended to install OpenStack in a VM. From devstack doc:
DevStack should run in any virtual machine running a supported Linux release. It will perform best with 2Gb or more of RAM.
As of the time of this writing (Jan 2015), supported distros are:
Ubuntu (latest LTS)
Fedora
CentOS
Regarding NAS: you can of course use it, but "outside" Openstack apis, by providing mount points to your vms. It's even mandatory if you want to support live migration.