Red Hat Bugzilla – Bug 1194627
User shouldn't be allowed to create bond interface without specifying identifier when it is set to managed
Last modified: 2017-02-23 15:27:53 EST
Description of problem: when user add bond interface and without specifying mac, the new host form was submitted successfully. Ideally UI should auto pick the mac based on the 'attached_device'. or should throw validation error when user tries to create bond interface without specifying mac address. Version-Release number of selected component (if applicable): sat6.1 beta snap3 Satellite-6.1.0-RHEL-6-20150217.0 How reproducible: always Steps to Reproduce: 1. add an bond interface when creating new host and do not specify mac address 2. 3. Actual results: form was submitted successfully however provisioning template doesn't show bond interface configuration Expected results: As soon as user submits the from, UI should raise an error. as mac was not associated with bond interface OR UI should auto pick the mac based on attached_device selection. Additional info:
Created redmine issue http://projects.theforeman.org/issues/9495 from this bug
Mac is not mandatory even for managed bond. We don't set it anywhere, it's not part of bond config file. When bond is created on host, it get's MAC of one of bonded interface automatically. We don't even know which one it will be. We get the MAC reported back with first puppet facts upload. On the other hand, identifier is needed. Sachin, are you OK with changing the subject to add presence validation on identifier for managed bonds?
(In reply to Marek Hulan from comment #2) > Mac is not mandatory even for managed bond. We don't set it anywhere, it's > not part of bond config file. When bond is created on host, it get's MAC of > one of bonded interface automatically. We don't even know which one it will > be. We get the MAC reported back with first puppet facts upload. Ok, Thanks Marek for explaining this info. Its really helps and I strongly feel we should document this too. But would like to know why we need mac for Bond field when we know that its auto picked from one of attached_devices. So we should disable the mac field and filled it with mac of any of attached_devices while adding bond interface ? I still don't understand what would be the use case when user will fill mac while created bond interface. > On the > other hand, identifier is needed. Sachin, are you OK with changing the > subject to add presence validation on identifier for managed bonds? Yeah..thanks for pointing this. I'm fine with anyway. if you want I can open another bug for validating identifier presence.
> I still don't understand what would be the use case when user will fill mac while created bond interface. One use case is documentation purposes when user don't set bond to be managed. Second is using the bond information in puppet module (not provisioning template) since the all bond attributes are part of ENC. Users may have various reasons to specify the MAC in this case. > Yeah..thanks for pointing this. I'm fine with anyway. if you want I can open another bug for validating identifier presence. No, I'll just change the name so we keep the history of this issue on one place (plus it's already synced to redmine issue)
Moving to POST since upstream bug http://projects.theforeman.org/issues/9495 has been closed ------------- Marek Hulán Applied in changeset commit:e329155b185e3be5863bfd79582b61a6ecc9ae9a.
verified with: sat6.1 beta snap6 compose2 (Satellite-6.1.0-RHEL-6-20150311.1). Now if user creates managed bond interfaces without specifying identifier, UI throws error that identifier can't be blank. Thanks for fixing this. Please see the screenshot for verification.
Created attachment 1002670 [details] UI raises error when creating a bond interface without identifier
This bug is slated to be released with Satellite 6.1.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2015:1592