#10072 Builds fail on buildvm-s390x-20.s390.fedoraproject.org: Couldn't resolve host name for http://kojipkgs01.fedoraproject.org/repos/f35-build/3809562/s390x/repodata/repomd.xml
Closed: Fixed 2 years ago by kevin. Opened 2 years ago by ppisar.

Koji host buildvm-s390x-20.s390.fedoraproject.org fails to download repository metadata https://koji.fedoraproject.org/koji/taskinfo?taskID=71433947:

DEBUG util.py:542:  Executing command: ['/usr/bin/dnf', '--installroot', '/var/lib/mock/f35-build-28285034-3809562/root/', '--setopt=install_weak_deps=0', '--disableplugin=local', '--disableplugin=spacewalk', '--disableplugin=local', '--disableplugin=spacewalk', 'groupinstall', 'build', '--setopt=tsflags=nocontexts'] with env {'TERM': 'vt100', 'SHELL': '/bin/bash', 'HOME': '/var/lib/mock/f35-build-28285034-3809562/root/installation-homedir', 'HOSTNAME': 'mock', 'PATH': '/usr/bin:/bin:/usr/sbin:/sbin', 'PROMPT_COMMAND': 'printf "\\033]0;<mock-chroot>\\007"', 'PS1': '<mock-chroot> \\s-\\v\\$ ', 'LANG': 'C.UTF-8', 'LC_MESSAGES': 'C.UTF-8', 'LD_PRELOAD': '/var/tmp/tmp.mock.g5atjbjb/$LIB/nosync.so'} and shell False
DEBUG util.py:444:  Unable to detect release version (use '--releasever' to specify release version)
DEBUG util.py:444:  No matches found for the following disable plugin patterns: local, spacewalk
DEBUG util.py:444:  Errors during downloading metadata for repository 'build':
DEBUG util.py:444:    - Curl error (6): Couldn't resolve host name for http://kojipkgs01.fedoraproject.org/repos/f35-build/3809562/s390x/repodata/repomd.xml [Could not resolve host: kojipkgs01.fedoraproject.org]
DEBUG util.py:444:    - Curl error (6): Couldn't resolve host name for http://kojipkgs-cache01.s390.fedoraproject.org/repos/f35-build/3809562/s390x/repodata/repomd.xml [Could not resolve host: kojipkgs-cache01.s390.fedoraproject.org]
DEBUG util.py:444:    - Curl error (6): Couldn't resolve host name for http://kojipkgs02.fedoraproject.org/repos/f35-build/3809562/s390x/repodata/repomd.xml [Could not resolve host: kojipkgs02.fedoraproject.org]
DEBUG util.py:444:  Error: Failed to download metadata for repo 'build': Cannot download repomd.xml: Cannot download repodata/repomd.xml: All mirrors were tried
DEBUG util.py:598:  Child return code was: 1

I tried building the package three times with the same result. I guess something is wrong with a network the machine is attached to.


Metadata Update from @humaton:
- Issue tagged with: medium-gain, medium-trouble, ops

2 years ago

Problem is that /etc/resolv.conf is broken on host

[root@buildvm-s390x-20 ~][PROD]# cat /etc/resolv.conf 
# Generated by NetworkManager
search s390.fedoraproject.org
nameserver 10.3.160.33

The following hosts are also broken:

buildvm-s390x-17.s390.fedoraproject.org | CHANGED | rc=0 >>
search s390.fedoraproject.org
buildvm-s390x-16.s390.fedoraproject.org | CHANGED | rc=0 >>
search s390.fedoraproject.org
buildvm-s390x-20.s390.fedoraproject.org | CHANGED | rc=0 >>
search s390.fedoraproject.org
buildvm-s390x-19.s390.fedoraproject.org | CHANGED | rc=0 >>
search s390.fedoraproject.org

I am not sure but the rest of the s390 may also be 'broke' with resolv.conf if they need to talk to each other.

search iad2.fedoraproject.org vpn.fedoraproject.org fedoraproject.org

should be

search s390.fedoraproject.org  iad2.fedoraproject.org  fedoraproject.org

I ran

sudo -i ansible-playbook /srv/web/infra/ansible/playbooks/groups/buildvm.yml -l buildvm-s390x-17.s390.fedoraproject.org:buildvm-s390x-16.s390.fedoraproject.org:buildvm-
s390x-20.s390.fedoraproject.org:buildvm-s390x-19.s390.fedoraproject.org -t resolvconf

and it fixed the resolv.conf on those hosts. It looks like networkmanager is setting this file up for these and other boxes so the problem will probably reassert itself in the future. Proper fix should be:
1. Determine if a resolv.conf_s390 is needed and have the s390 hosts use that
2. Determine what is needed to make /etc/resolv.conf static and not changed with networkmanager.
3. Fix ansible playbooks as appropriate.

As such I am leaving this ticket open for someone to review/close/complete.

Metadata Update from @kevin:
- Issue priority set to: Waiting on Assignee (was: Needs Review)

2 years ago

So do you want network manager to just set the correct ip for DNS in the first place or just tell it to ignore any dhcp DNS requests? This should be something I can work on if it is still needed.

Sorry, I just went ahead and fixed it.

we don't need a seperate resolv.conf for them, we just want it to keep using the static one we set, so I made sure NM was leaving it alone on those hosts.

Metadata Update from @kevin:
- Issue close_status updated to: Fixed
- Issue status updated to: Closed (was: Open)

2 years ago

Login to comment on this ticket.

Metadata
Boards 1
ops Status: Done