ciscosystems / cisco-openstack-docs Goto Github PK

Says it all.

Container list doesn't show objects and space consumption

When on boarding, need to allow selection/definition of role

role:
Admin
SwiftOperator
Member

project:
{new project}

Change out the Ubuntu skin for a Cisco skin when deploying using the Ubuntu packaged Horizon.

Need to validate the currently available HA models, and verify one against our current OpenStack systems models.

I just did a cursory review of hte onboarding scripts. I see a number of things we need to address before we can release them for general consumption:

There's no shebang line. They'll only work if people are using bash or another shell that happens to use the same array syntax as bash.

Everything is hardcoded. Even the admin_token, endpoints and the name of the csv files. This is not acceptable. These things need to be passed in as command line arguments or environment variables.

The tenant description should be either configurable or at the very least be based on the user's name in the CSV.

The temp_passwd shouldn't be the same for all users and should be randomised,

Using telnet to send e-mail is a very bad idea. Use the mail(1) command line utility.

There are a lot of seemingly superfluous parentheses and square brackets whose purpose I don't quite understand.

They lack any kind of help functionality. If you run them, they're immediately going to make a lot of changes that the users might not be on board with (pun intended).

Glance currently uses the controller for it's data store. This works for small numbers of images and snapshots, but will not scale. Since we set up Swift, we should be able to leverage that as our principal data store.

Failed tempest tests do not clean up the tenants and users created by it. The following changes are made to enable clean up of the tenants and users in case of failures.

CiscoSystems/tempest@c7a6c2d

https://github.com/CiscoSystems/tempest/tree/tempest-clean-up

need to validate and setup multi-local-disk model with Swift

We need to determine why promiscuous mode (or a patched KVM) is needed when deploying openstack

and create the puppet site example for this (and cobbler?)

Assigned edgar in place of Rohit getting into the system.

We need to deal with the default deployment disk space so that we don't run out of disk space and kill mysql because we uploaded an image file.

In order to better control changes into running systems
Leverage Soren's build system.

Failed tempest tests do not clean up the tenants and users created by it. The following changes are made to enable clean up of the tenants and users in case of failures.

CiscoSystems/tempest@c7a6c2d

https://github.com/CiscoSystems/tempest/tree/tempest-clean-up

Collectd is set up, need to automate the configuration of a Nagios instance on the control node (and look at HA issues for the controller for Nagios).

Swift is functioning and a subset of systems users can read/write to the system, but not all users.

need to add:

CiscoSystems/puppet-cobblerd
CiscoSystems/puppet-snmpd
saz/puppet-ssh

Have been blocked since yesterday. Initially it was a glance issue. Now its a public IP issue. Arvind said he can ping my VMs from the cloud controller which means its a public IP issue or a lab issue.