| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
We're waiting for the OpenStack VMs to be able to SSH into before
proceeding with the DNS playbook (because until 2.3 Ansible doesn't
handle that gracefully).
But if the private key for the SSH connection was only specified using
the `--private-key` argument to `ansible-playbook`, the test was
failing because it didn't take that possibility into the account.
Now it does.
Fixes #355
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
(#359)
* Move development back to dev branch
* Remove /etc/hosts autogenerate and dnsmaq resolution
* Fix loop on nodes
* Fix count
* Prepare for upstream commit
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
| |
* fix of readme format
* fix of readme format
|
| |
|
|
|
|
| |
Fix ssl-certificate role, so it doesn't fail when user provides custom
certificate, and add a check for already created ssl certificate in GCP, so the
following tasks could be skipped.
|
| |
|
|
|
|
|
|
|
|
|
| |
* testing using the shared validation plays
* Fix etcd validation in GCP
In GCP, etcd certificates are created with the short hostnames of the master
nodes, not with the fqdn.
* Set validate_etcd_short_hostname variable when calling etcd validation
|
| |
|
|
|
|
|
|
|
|
|
| |
* Enable epel
For the time of the transition, epel repo is required for the python-libcloud
package.
* Quote rh username and pool id
* Supply zone to the set-disk-auto-delete command
|
| | |
|
| |
|
|
|
|
|
|
|
|
| |
* enable io1
* fixing groups and using iops correctly
* linting
* resolving line spacing weirdness
|
| |
|
|
|
|
|
|
| |
* numeric variables must include quotes
* fix for inventory
* linting
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
* 3.5 release changes
* Update README.md
* Update README.md
* Update main.yaml
* Adding rhel_subscribe to all plays and changing variables in python wrapper
* change variable names in add-node.py
|
| |\
| |
| | |
IAM arn fix
|
| | |
| |
| | |
blankspace
|
| |/ |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* testing exports
* more facts to output
* support new gluster functionality
* gluster node plays
* accidently removed line
* should have been delete on terminate
* required console_port
* gluster prereqs
* rolling this back
* prereqs and firewalld fix
* linting
* cloudformation cleanup and variable fix
* adding intelligence to add node and linting
* missing :
* add node functionality without cloudformation facts
* rolling back unneeded items
* port fix and new instance group
* static inventory update to include storage
* handle if cloudformation facts and without cloudformation facts
* allow for multiple storage stacks
* 3.5 release changes
* name change
* starting crs coding
* crs addition
* openshift node to cluster ports
* linting
* naming failure
* need to quote pool name
* better infrastructure when rules
* fixes for skipping crs
* lb fix
* switching back to iptables
* heketi is not required to be on all servers
* gluster nodes need to be able to communicate
* missing if statement
* improper instance group configuration
* reversed variables
* linting
* rhsm timeout fix
* make crs rhsm serial
* prompt for name
* re-order and fix of when
* display node size
* prompt for rhsm values
* crs fixes
* better iptables rules
* force the use of iptables
* linting
* ensure firewalld disabled
* linting
* force an iptables reload
* better handling of iptables
* ugly rule placement but works
* more port work
* linting
* instance group clean up
* teardown variable add
* adding the ports correctly
* Readme update
* increase size of default instance and network manager
* Update README.md
|
| |
|
|
|
|
|
|
|
|
| |
* 3.5 release changes
* Update README.md
* Update README.md
* Update main.yaml
|
| |
|
|
|
|
| |
* block ansible 2.3 for the current time
* Update main.yaml
|
| |
|
|
|
|
|
|
| |
* New ssl-certificate role
* Migrate core infrastructure to ansible
* Fixed pylint errors
|
| |
|
|
|
|
|
|
| |
* timeout increase for subscription-manager
* Update main.yaml
Needed to place this higher in the file
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Remove RHN and dependencies from Heat templates
The templates will just deploy the OpenStack resources from now on. Any
configuration or dependency installation will be handled by Ansible.
* Add outputs to the Heat templates
These outputs will be consumed by Ansible so we'll no longer need the
Python scripts and templates to generate the inventory.
* Add `become: true` to the bind-server playbook
The playbook requires root privileges and this was achieved in
bin/full-service.sh by passing `--become` to `ansible-playbook`.
Since we'll be invoking the playbook directly, we just ask for sudo
where necessary instead.
* Add the "all in one" ansible playbook
* Update the readme
* Remove the old files
The move to a full ansible installation means most of the files under
`bin` and the heat environment files are no longer needed.
* Fix yamllint errors
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Deprecate Iscsi and add registry storage account
* Cleanup of additional storage references
* Enable the launch of the development instance
* setup scripts to support forking and upstream via user
* Fix registry storage account name
* Fix development branch to be master on the upstream
* Fix infranode default vmsize
* Fix vmsize
* Pass the registry storage key to bastion
* Move to reference for registrystorage
* Fix Reference
* Fixup registry key parameter
* Remove output section
* Fix parameters on bastion
* Fix depends storage name
* Move to array reference for keys
* Fix storage api version for keys for registry
* Change to samcogan style reference
* Fix paren
* Add registryConnect to parameters
* Move the grabbing the keys to bastion level
* Fix case
* Fix branch
* Fix up listkeys reference
* Fix storage name
* Fix Mispelling
* Switch from variable to parameter
* Fix case issue and simpilfy
* Fix reference
* Fix missing parameter
* Found bug - ListKeys not supported in variable - moved to parameter
* Add in accountname, and restructure yum installs plus add in nodejs for azure cli and scripting
* Fixup order and remove force
* Do direct references
* Remote quotes and verify values recieved for registry
* Add additional debug
* replace with bash for setup script
* Add array processing to work arround bash bug
* Add all the variables needed for pv/pvc creation using azure
* Build registry and ocp azure configs
* Run the setupregistry playbook
* Remove misc reference
* Fix install script parens
* Fix epel for sendmail
* Fix quota to work on all nodes
* Fix quota
* Fix typo on openshift.local.volumes
* Change order on quota setup
* Fix sendmail
* Move to 7.3
* Work around resource disk move about bug with ansible error recovery
* Start move of quota to node based accounts
* Cleanup zerovg script and make sure all ocp have it
* remove docker-storage-setup command, but update file for ansible scripts use
* Make sure updateok is last
* Roll back to 7.2 because of rhn issues
* Fix to handle multiple open disk in node setup scripts
* Get rid of setupregistry yml and do direct
* Initial setup for azure pvc
* Add autorun of azure setup
* Change to short host names to be compatible with azure pvc
* Fix module path to include openshift ansible library
* Fix azure-pvc-setup output path
* Fix hanging issue on azure pvc setup
* Update default sizes to Standard_DS4_v2
* Add storage accounts for azure pvc
* Add top level availability sets
* Fix bastion size to be in variable and smaller
* Fix newStorageAccountVolume1 that was missing
* Fix api version
* Fix redundant host groups
* disable running the azure pvc script for testing
* Reduce to 1 node and 1 master
* Fixup to do all masters and nodes for azure
* Revert to rhel 7.3
* Change the azure pvc setup to 1, and initial implementation of master lb, and master av set
* Fix openshiftMasterHostname name issue
* Fix openshiftMasterPublicIpDnsLabel
* Fix paren issue
* Fix openshiftMasterPublicIpDnsLabel issue
* Fix openshiftInfraHostname
* Cleanup InfraLbPublicIpDnsLabel
* Fix nested parameter
* Add api versions into variables
* Fix openshiftMasterPublicIpDnsLabel
* Remove duplicate wildcardzone entry
* Add missing parameters to master2 and 3
* Fix parameter
* Add infranode load balancer and avset
* Add nod avset and cleanup order of variables and resources
* Fix domain issue on api
* Fix fulldomain
* Limit to one master update at a time
* Timed setup of nodes
* Add great time between restarts
* Fix escapes for loop in pvc deploy
* Fix node counter
* Fix escapes
* Revert "Fix escapes"
This reverts commit 72987c30438196a8b939d7ea4ca35f96f5b74035.
* Change to short hostnames using generated /etc/hosts
* Test using dnsmasq configured by ocp
* Fixup to allow dnsmasq to do forward and reverse lookup to enable azure storage to function properly
* Fix to make sure bastion is included
* Break the bastion updates and the rest of nodes
* Get rid of redundant lines and use better host selection for /etc/host and dnsmasq updates
* Fix nodenames on master and node
* Get rid of timing and pacing for azure pvc setup
* Add direct node name statements in /etc/ansible/hosts
* Switch to OCP 3.5
* Try 3.5 without storage setup
* Add back etcd for 3.5
* Cleanup node name references
* Add workaround to delete nodes so they can recreate
* Fixup so defers_to will work
* Add nip.io for wildcard domain
* Cleanup output section
* Fix NIP
* Fix extra paren on resource name
* Cleanup storage names
* Fix IP Address for router
* Add in script to create vhd container
* Fixup vhds for azure pv
* fix sending log via email
* Move creating storage profile earlier
* Remove pips from masters
* Fix up master.json to remove pip reference
* Cleanup Log when Mailed
* Cleanup up name used to send mail
* Remove all reference to pip
* Remove pip from infranode
* Remove pip references from parameters
* Cleanup on depends to make deploy more consistent
* Add some sleep to make sure network stable
* Fix storage class substitution
* Fixup to make sure masters are unschedulable
* Remove extranous line
* Make sure the set schedulable is last
* Fix switch on adm
* Prep for upstream commit
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* rename add-node.json
* added config file editting for wrapper script
* modifying some files that shouldnt have changes
* mend
* adding add-node.py
* Adding new files for add node
* Adding infra scale up
* linting
|
| |\
| |
| | |
specify origin version
|
| |/ |
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Set image family for clean image
* Don't use PREFIX where it's not needed
Some resources can be shared between multiple OCP clusters.
* Migrate network creation
* Rename local playbook to prereq
* Renamed network deployment
* Migrate creation of gold image to Ansible
* Don't mark checking task as changed in rhsm-subscription role
|
| |\
| |
| | |
Increate timeout. Resovles #301
|
| | | |
|
| | | |
|
| | | |
|
| |/ |
|
| |\
| |
| | |
make the installer setup the registry
|
| | |
| |
| | |
Missed /registry to /exports
|
| | | |
|
| | | |
|
| | | |
|
| | | |
|
| | | |
|
| | | |
|
| | | |
|
| | | |
|
| | | |
|
