I am learning the bosh and cloudfoundry , I am curiosity about the vm network setting in the vsphere.
How the bosh setup the network (ip, gateway) for the vm in the vsphere. i know the bosh will fire a bosh-agent/nats-agent in the target vm to change the value, but how it first connect to the agent ?
because i am not familiar with go, ruby , so cannot figure out the code .
https://bosh.io/docs/vsphere-cpi.html explains interactions with VSphere, and related bosh configuration.
How the bosh setup the network (ip, gateway) for the vm in the vsphere.
Are you looking to understand how to configure bosh, or how the CPI interacts with vsphere?
The docs above answer the first part. As for how bosh CPI operates under the cover is uses the vsphere/SDK APIS documented here, https://www.vmware.com/support/developer/vc-sdk/visdk41pubs/ApiReference/index-methods.html
Related
I recently created a redis instance in Google Cloud Platform Memorystore and was trying to connect to it using Cloud Shell.
telnet 10.0.0.3 6379
Trying 10.0.0.3...
telnet: Unable to connect to remote host: Connection timed out
But always I'm getting a connection timeout issue. As per the documentation it states the compute instances which are in the same project should be able to connect to the instance. Did anyone also encounter the same issue ? and are there any resolutions for testing it locally ?
Cloud Shell is very limited. The port that you are trying to connect to is not allowed for outgoing connections. The documentation you shared is referring to a GCE (Google Compute Engine) VM instance, yes Cloud Shell provides a temporary instance but it is very limited. At the moment you can only access it from Compute Engine VM instances, Kubernetes Engine clusters, and App Engine (flexible & standard) as well as Cloud-Functions, you will need to create a GCE VM and follow the documentation that you mentioned.
I have installed Cloudera Express on Google cloud platform vm instance using Cloudera Manager.All the web services are running but web UI links of all services (Hue, Hbase, Spark etc.) are not loading.I can't reach the login page also .It gives the error as given below:
This site can’t be reached
instance-1.c.cluster-183105.internal’s server DNS address could not be found.
DNS_PROBE_FINISHED_NXDOMAINpls find the screenshot of error here
You're configurations in Cloudera Manager are pointing to internal DNS records.
So, for example, Cloudera Manager does a health check against the web server, and says they are reachable because it is also on that internal network.
Your machine cannot resolve the internal DNS records unless you tunneled into that network.
The solution here is to
whitelist your network in GCP so that you can reach the cluster without exposing it to the open internet
You must go through each component and change the web interface address to be the public DNS
Or setup and use an SSH tunnel or VPN to any machine in the cluster
Note, Cloudera Director and Cloudera Atlus are two offerings for running and configuring CDH in the cloud
I am pretty new to cloudfoundry. I am still trying to understand how exactly it works.
Say if I have three VMs. VM 1 is running on server A.
VM 2 and 3 are running on server B.
If I wanted to use a single CloudFoundry Instance on those three, would it work?
And if not, how could I use Cloudfoundry on multiple servers or at least multiple VMs? I know I can use BOSH to set them up, but do I still have to manage each instance seperately?
Thank you:
Jannis
BOSH will deploy VMs for you, you typically don't deploy Cloud Foundry onto existing VMs. BOSH supports deploying to several infrastructures. The core supported infrastructures include AWS, vSphere, OpenStack, and vCloud Air/vCloud Director. There are also community-provided "Cloud Provider Interfaces" for IBM SoftLayer, Azure, Google Compute Engine, and more.
Cloud Foundry is meant to be run as a distributed service, i.e. on multiple VMs. Typically those VMs will be on multiple different hosts, hardware racks, servers, datacenters, what have you. And BOSH is designed to facilitate deploying and managing distributed services like Cloud Foundry. So no, you do not need to manage individual VMs separately.
You can read more about BOSH and Deploying Cloud Foundry.
I've been following the instructions for deploying cloud foundry on openstack and was having a problem with step that uploads the bosh stemcell:
$ bosh upload stemcell http://bosh-jenkins-artifacts.s3.amazonaws.com/bosh-stemcell/openstack/bosh-stemcell-latest-openstack-kvm-ubuntu.tgz
...
Error 100: Unable to connect to the OpenStack Compute API. Check task debug log for details.
...
E, [2013-09-21T09:02:11.359958 #2587] [task:1] ERROR -- : No route to host - connect(2) (Errno::EHOSTUNREACH) (Excon::Errors::SocketError)
I can ssh to the instance running Micro Bosh and confirm that it can ping the compute host, but it can't connect via tcp/http.
I've described the error in more detail here:
http://openstack.redhat.com/forum/discussion/625/ingress-issue-from-spawned-instance-to-compute-host#Item_1
It appears to basically be an openstack firewall/iptables configuration issue between spawned instance running Micro Bosh and controller/compute host running the compute API, which I can only temporarily fix via iptables. But I was surprised not to find any other cloud foundry related posts pointing to this issue, and was wondering if anyone has seen this issue and found a workaround?
I am evaluating cloudfoundry (private cloud option) with Openstack as an IaaS candidate.
I have following setup in mind, but looks like I am missing some connections -
I will have OpenStack installed
On one VM on openstack (ubuntu 10.4 image), I will install cloudfoundry cloud_controller
On other multiple VMs on openstack, I will install cloudfoundry DEAs
And this I understand as called a multiple hosts installation of cloud foundry
Now when I push an application to cloudfoundry using VMC (with 5 instances request), One of the Cloudfoundry DEA will spawn 5 VMs on itself and deploy/run the app on all 5 cloudfoundry VMs
That means I have 5 instances of my app running
I can access the app through a single URL and cloudnfoundry controller/router will route the request to one of the running instances of my app
Now for scaling the infrastructure, I can reconfigure my openstack instances and restart them with new configuration (i.e. more volume, more RAM etc)
And for scaling the application, I can simply add more instances to the cloudfoundry vmc push command
Sorry for the writeup but pls suggest if this is a valid understanding (also if you guys have better options - basically we are looking at a scalable application and infrastructure for developers)
Thanks Much,
Vcap OSS questions are best directed to the vcap dev site and I would suggest you start there.