Error "create stemcell failed" while Installing Pivotal Cloud Foundry® on vSphere

Products

• VMware Tanzu Application Service for VMs

Issue/Introduction

Symptoms:

Install Ops Manager, import Pivotal Application Service (and any other service tiles), configure those tiles correctly and finally click 'Apply Change' to start install. Then get error "create stemcell failed" followed by error details information:

create stemcell failed: <error details and Ruby trace log>

 

Environment

Cause

The root causes depend on error details. Please refer <Resolution> section for detailed error and corresponding resolutions. 

 

Resolution

Case 1

Started deploy micro bosh
Started deploy micro bosh > Unpacking stemcell
tar: image: Wrote only 8704 of 10240 bytes
tar: stemcell.MF: Cannot write: No space left on device
tar: Exiting with failure status due to previous errors
create stemcell failed: 'tar -zxf /var/tempest/stemcells/bosh-stemcell-2865.1-vsphere-esxi-ubuntu-trusty-go_agent.tgz -C /tmp/d20150517-1133-zvv3he/d20150528-27173-1mm4u6t/sc-20150528-27173-1n2baom' failed with exit status=2 [tar: image: Wrote only 8704 of 10240 bytes
 tar: stemcell.MF: Cannot write: No space left on device
 tar: Exiting with failure status due to previous errors
]:

Cause and Resolution

Ops Manager failed to extract (Bosh Director)stemcell into /tmp folder due to lacking of empty disk space. You can manually clear /tmp folder or just reboot Ops Manager VM, it gets emptied automatically on boot.

Case 2

create stemcell failed: execution expired:
/home/tempest-web/tempest/web/vendor/bundle/ruby/2.1.0/gems/httpclient-2.4.0/lib/httpclient/session.rb:805:in `initialize'

Cause and resolution

Mostly this is caused by firewall rules that Ops Manager/Bosh Director failed to connect to vCenter and ESX hosts at port 443. Please assure that Ops Manager/Bosh Director can connect to vCenter and ESX hosts as follows:

1. Ops Manager needs access to vCenter and all ESX hosts in the cluster at port 443
2. The Bosh Director needs access to vCenter and all ESX hosts in the cluster at port 443

Case 3

create stemcell failed: getaddrinfo: Temporary failure in name resolution (https://xxx.yyy.zzz:443):

Cause and resolution

ESXi host xxx.yyy.zzz can't be name resolved successfully. Please assure DNS servers used by Ops Manager/Bosh Director can name resolve xxx.yyy.zzz.

Case 5

create stemcell failed: undefined method `split' for nil:NilClass:
/home/tempest-web/tempest/web/vendor/bundle/ruby/2.1.0/gems/bosh_vsphere_cpi-1.2690.4.0/lib/cloud/vsphere/resources/util.rb:15:in `average_csv'
/home/tempest-web/tempest/web/vendor/bundle/ruby/2.1.0/gems/bosh_vsphere_cpi-1.2690.4.0/lib/cloud/vsphere/resources/cluster.rb:220:in `block in fetch_cluster_utilization'
/home/tempest-web/tempest/web/vendor/bundle/ruby/2.1.0/gems/bosh_vsphere_cpi

Cause and resolution

Host in vSphere cluster might get disconnected during PCF install. This can be resolved by restarting the install.