Bug 1813019
| Summary: | UPI installation failed when using static Network config + DNS PTR not set. | ||
|---|---|---|---|
| Product: | OpenShift Container Platform | Reporter: | Martin Ouimet <mouimet> |
| Component: | RHCOS | Assignee: | Ben Howard <behoward> |
| Status: | CLOSED DUPLICATE | QA Contact: | Dusty Mabe <dustymabe> |
| Severity: | medium | Docs Contact: | |
| Priority: | medium | ||
| Version: | 4.2.0 | CC: | bbreard, dustymabe, imcleod, jligon, miabbott, nstielau, smilner |
| Target Milestone: | --- | ||
| Target Release: | 4.5.0 | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | If docs needed, set a value | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2020-06-18 20:31:03 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
@behoward this smells similar to BZ#1804426 or perhaps BZ#1763700. Could you please triage at your convenience and provide guidance? This does look like https://bugzilla.redhat.com/show_bug.cgi?id=1804426 to me. Can you retry with 4.4? This bug should have been fixed. I'll attempt to verify this one. Ok I went through and executed some reproducers for this. I verified that I *could* reproduce the problem with 4.2 (43.81.202003191953.0) and that the problem is *fixed* in both 4.4 (44.81.202004260825-0) and 4.5 (45.81.202006180929-0). I did this by running an install with the following command line arguments: ``` console=ttyS0 coreos.inst.install_dev=sda coreos.inst.ignition_url=http://192.168.122.1:8000/config.ign coreos.inst.image_url=http://192.168.122.1:8000/path/to/raw.gz ip=192.168.122.111::192.168.122.1:255.255.255.0:testhost::none:192.168.122.1 ``` And verifying: - 4.3 - the machine came up with the right hostname on first boot, but didn't have the correct hostname set after subsequent boots - 4.4 - the machine came up with the right hostname on first boot and subsequent boots - 4.5 - the machine came up with the right hostname on first boot and subsequent boots The media I used: - 4.3 rhcos-4.3.8-x86_64-metal.x86_64.raw.gz, rhcos-4.3.8-x86_64-installer.x86_64.iso - 4.4 rhcos-4.4.3-x86_64-metal.x86_64.raw.gz, rhcos-4.4.3-x86_64-installer.x86_64.iso - 4.5 rhcos-45.81.202006180929-0-metal.x86_64.raw.gz, rhcos-45.81.202006180929-0-installer.x86_64.iso I'm going to officially call this a duplicate of 1804426 and close it out as such. *** This bug has been marked as a duplicate of bug 1804426 *** |
Description of problem: Version-Release number of selected component (if applicable): 4.2 and 4.3 How reproducible: UPI installation of Openshift 4.2+ PXE boot configuration with static IP and static hostname Example of PXE configuration DEFAULT default TIMEOUT 20 PROMPT 0 LABEL default KERNEL http://192.168.2.8/rhcos-4.3.0-x86_64-installer-kernel APPEND bootdev=ens4 ip=192.168.3.22::192.168.3.1:255.255.255.0:master2.acocp.rhcasalab.com:ens4:none nameserver=192.168.3.8 rd.neednet=1 initrd=coreos/rhcos-4.3.0-x86_64-installer-initramfs.img console=tty0 console=ttyS0 coreos.inst=yes coreos.inst.install_dev=vda coreos.inst.image_url=http://192.168.3.8/rhcos-4.3.0-x86_64-metal.raw.gz coreos.inst.ignition_url=http://192.168.3.8/master.ign Steps to Reproduce: 1. Do not set PTR record in DNS 2. Configure static network in PXE including Hostname 3. Launch bootstrap and masters. Actual results: First boot of CoreOS shows the hostname set correctly. After reboot of CoreOS, hostname is set to localhost. Installation failed because all nodes has hostname localhost. Expected results: Given the fact we set the hostname static through PXE, I would expect hostname to not be altered by DNS PTR check. Additional info: I know good practices are to always have DNS PTR set up, but I got a customer having some constraint in a closed environment were setting PTR is an issue. I would like to understand why the reverse DNS check overrides the hostname originally set up by PXE. Thanks,