clusters and pools not interchangeable in knife vsphere vm find

[match-garricklinn]

What are you trying to do and what's not working?

We don't currently have any resource pools in vcenter, we do have clusters, but knife vsphere vm find demands that a pool be specified and does not take a cluster in place of a pool.

Please paste the command and output (run with -VV). Don't forget to X out any passwords or sensitive information

TX1-GLINN03ML:ma_siteserver garrick.linn$ knife vsphere vm find --pool "Production/Production-DA" --match-name "DA3SHWL021" -VV
INFO: Using configuration from /Users/garrick.linn/Documents/chef/automate-match/.chef/knife.rb
DEBUG: value for config item vsphere_host: da0shas000
DEBUG: value for config item vsphere_path: /sdk
DEBUG: value for config item vsphere_port: 443
DEBUG: value for config item vsphere_nossl:
DEBUG: value for config item vsphere_user: xxxxxx
DEBUG: value for config item vsphere_pass: xxx
DEBUG: value for config item vsphere_pass: xxx
DEBUG: value for config item vsphere_pass: xxx
DEBUG: value for config item vsphere_insecure: true
DEBUG: value for config item proxy_host:
DEBUG: value for config item proxy_port:
DEBUG: value for config item vsphere_dc: Clearview
Pool Production/Production-DA not found
TX1-GLINN03ML:ma_siteserver garrick.linn$

Other details, such as gem version and if you're working with a Linux or Windows VM

Running OSX 10.12.1 on my laptop chef-provisioning-vsphere (0.10.0) fog-vsphere (1.7.0) knife-vsphere (1.2.24)

[swalberg]

Thanks for letting us know about this. I will have a look. The method that looks for the resource doesn't support the slashes. I'll need to do some debugging on my system.

[swalberg]

Hi sorry for the delay. I tested this against our two clusters on one vcenter, and a pool inside one of those clusters, and it worked:

knife vsphere vm find --pool xxx-Prod --match-name converter # cluster
knife vsphere vm find --pool xxx-Lower --match-name cheftest # cluster
knife vsphere vm find --pool infrastructure_testing --match-name firewall-centos-72-f7a7d77f # pool inside xxx-Lower

We talked a bit in slack... just wanted to confirm you did try --pool "Production-DA". Is there perhaps another object with that same name that we're getting confused on?

[match-garricklinn]

Hi Sean...apologies should be coming from my side, I got knocked out by the flu for a few days and then got caught up in other things.

Hmm...I did try a few variations on that, but no dice. (We do actually have a resource pool now for a redis cluster but that's the only one so far.) Interestingly, we DO have a cluster named Clearview-Build-Staging which isn't throwing the "not found" error when passed to --pool, although it doesn't currently have any VMs in it. Biggest difference I can see there is the lack of a slash in the cluster name, although that may be a red herring since (if I understand you correctly) you were able to specify a pool/cluster picked out from a cluster name containing a slash?

`

TX1-GLINN03ML:automate-match garrick.linn$ knife vsphere pool list Pool: Production/Production_Tier1-DA/Redis TX1-GLINN03ML:automate-match garrick.linn$ knife vsphere vm find --pool Production-DA --match-name "DA3PMWS101" Pool Production-DA not found TX1-GLINN03ML:automate-match garrick.linn$ knife vsphere vm find --pool Clearview-Build-Staging TX1-GLINN03ML:automate-match garrick.linn$`

[rhadoo]

Hi, Unfortunately i am running of a fork of an old version and i didn't get to push anything back to main since the initial commit a year ago. I think i have experienced the issue with the slashes and have that already fixed in our production, i will check that tomorrow and get back with a PR if so.

[swalberg]

Hey @garrickmatch no, I can't use a slash. The code takes the pool and recursively searches for a pool or cluster with that name. There are other commands inside the gem that will split the name up on slashes and follow the hierarchy -- the fully/qualified/path isn't actually sent to the API.

Can you attach a screenshot of your cluster either here or in DM in Slack?

[swalberg]

Thanks @rhadoo I'd love to see that!

[rhadoo]

Hi, Sorry for the late reply. It seems i have added a --pool-path flag, which makes --pool to be treated as path, going to find_pool() instead of traverse_folders_for_pool(). I don't think i have made any other change related to this. I will pull current and try to add these.

[swalberg]

That's a great way to maintain backwards compatibility. Thanks!