Bug 1708680 - Restart doesn't work(can't get ip) for rhel76 which has cloud-init config
Summary: Restart doesn't work(can't get ip) for rhel76 which has cloud-init config
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Container Native Virtualization (CNV)
Classification: Red Hat
Component: Virtualization
Version: 1.4.0-2
Hardware: All
OS: Unspecified
medium
medium
Target Milestone: ---
: 2.1.0
Assignee: Vladik Romanovsky
QA Contact: Denys Shchedrivyi
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2019-05-10 14:46 UTC by zhe peng
Modified: 2019-10-31 14:09 UTC (History)
11 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2019-10-31 14:08:47 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)
vm definition file (1.17 KB, text/plain)
2019-05-10 14:57 UTC, zhe peng
no flags Details
vm yaml output (1.40 KB, text/plain)
2019-05-13 08:23 UTC, zhe peng
no flags Details
vmi yaml output (2.26 KB, text/plain)
2019-05-13 08:47 UTC, zhe peng
no flags Details


Links
System ID Private Priority Status Summary Last Updated
Red Hat Product Errata RHEA-2019:3282 0 None None None 2019-10-31 14:09:01 UTC

Description zhe peng 2019-05-10 14:46:16 UTC
Description of problem:


Version-Release number of selected component (if applicable):


How reproducible:


Steps to Reproduce:
1.
2.
3.

Actual results:


Expected results:


Additional info:

Comment 1 zhe peng 2019-05-10 14:57:36 UTC
Created attachment 1566707 [details]
vm definition file

Comment 2 zhe peng 2019-05-10 15:08:55 UTC
Sorry for save a blank description, post here:
Description of problem:
create a rhel vm with cloud-config, after restart, vm can't get IP address.

Version-Release number of selected component (if applicable):
CNV 1.4 async

How reproducible:
always

Steps to Reproduce:
1.create a vm
2.start vm
#virtctl start vm
#oc get vmi
test-vm    2m        Running   10.129.0.21   cnv-executor-zpeng-node1.example.com

vm can get ip, then restart vm
#virtctl restart test-vm
after a while, check vm
#oc get vmi
test-vm    8m        Running                         cnv-executor-zpeng-node2.example.com

login to the vm, the interface didn't start

Actual results:
vm can't get ip

Expected results:
vm have ip address

Additional info:
Talk with Marcin, he suggest to file one bug and discuss if this is a real bug

Comment 3 Fabian Deutsch 2019-05-13 07:53:55 UTC
Zhe, please provide `oc get -o yaml $VMNAME` (replace $VMNAME) for the VM as well.

So far my assumption is that the VM uses a bridge/masquerade binding with the pod network, and after start, the VM is getting a new mac address from the network, which will break networking isde the VM, as inside the VM the ifcfg* scripts were tied to the MAC used for the first boot.

The resolution to this problem is:
1. for pod network have a stable address (requires https://jira.coreos.com/browse/CNV-1803)
2. For any additional network we depend on what the provide can provide

We need to wait for oc get poutput to udnerstand what the situation with this vm is (1 or 2)

Comment 4 zhe peng 2019-05-13 08:23:15 UTC
Created attachment 1567816 [details]
vm yaml output

Comment 5 Fabian Deutsch 2019-05-13 08:33:31 UTC
Thanks, I missed to say: oc get vmi $VMNAME - please this as well

Comment 6 zhe peng 2019-05-13 08:47:34 UTC
Created attachment 1567823 [details]
vmi yaml output

Comment 7 Fabian Deutsch 2019-05-13 09:08:21 UTC
Zhe, does it also happen if you create (and start and stop and start) a VM from the UI?

Comment 8 zhe peng 2019-05-13 11:01:11 UTC
Hi Fabian,
yes, I create a VM from UI wizard and still have this issue, I link vm yaml later.

Comment 9 zhe peng 2019-05-13 11:03:39 UTC
ui vm yaml file link: 
http://pastebin.test.redhat.com/762358

Comment 10 Fabian Deutsch 2019-05-13 12:14:05 UTC
Thanks.

Can you please also try this flow with CNV 2.0?

IIRC the default bdining mode was changed to masquerade in CNV 2.0 UI flows, which might avoid this issue.

Comment 11 Fabian Deutsch 2019-05-13 13:14:42 UTC
Zhe, the other question: As a workaround can you try to remove the HWADDR line from the ifcfg* file after the first boot?
After this, on later boots, the VM should still get an IP.

Comment 14 Fabian Deutsch 2019-05-14 12:18:27 UTC
dev and qe agreed to move this to 2.1, as it only impacts pod network interfaces (nto additional networks).

A known issue bug #1709850 was filed to mention it in the 2.0 docs.

Comment 17 Fabian Deutsch 2019-06-20 12:06:53 UTC
This will be fixed by https://jira.coreos.com/browse/CNV-1802

Comment 19 Denys Shchedrivyi 2019-09-23 21:19:54 UTC
Verified on hco-bundle-registry:v2.1.0-62

Comment 24 errata-xmlrpc 2019-10-31 14:08:47 UTC
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/RHEA-2019:3282


Note You need to log in before you can comment on or make changes to this bug.