This was just added to help debug some startup problems, so we can drop
it now. Also, it causes the playbook to be non idempotent (ie, this task
runs and changes everytime the playbook it run, causing a lot of noise
in the check/diff runs).
/run/zabbix should get created by systemd, so no need to muck with it.
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
We intend to use disk encryption and udev network interface
naming on all hosts with the re-deployment, so this attempts to
set up for that.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
An explicit provides for /usr/libexec/platform-python was added to
platform-python in RHEL 8.
https://access.redhat.com/errata/RHSA-2024:6975
This solves the original reason for needing the filelists option, so it
should be safe to remove. This will make bootstrap chroot creation
faster for all builds, not just EPEL 8.
https://pagure.io/releng/issue/12199
Signed-off-by: Carl George <carlwgeorge@gmail.com>
We've finally ironed out the issues in stage with this, so this is the
configuration to enable it in production.
This should be rolled out in conjunction with disabling the bash script
that currently handles image pushes.
Ansible `pause` module is only running once per playbook. This is a hacky way to
make it run for each host.
Signed-off-by: Michal Konecny <mkonecny@redhat.com>
When introducing the new validation mechanism I accidentally removed check for
ipa_initial from ipa-replica-install. This commit is adding it back.
Signed-off-by: Michal Konecny <mkonecny@redhat.com>
As found on in https://pagure.io/fedora-infrastructure/issue/12149 the current
method of checking if the ipa-replica-install have to be executed is not
reliable enough.
Let's instead check for log file that is created during the installation. This
will need manual intervention in case the ipa-replica-install needs to be run
again, but it will be much more reliable otherwise.
Signed-off-by: Michal Konecny <mkonecny@redhat.com>
The whoosh engine doesn't seem to be maintained anymore and when generating the
index it got stuck when the size was about 20 GB. Xapian on the other hand looks
more stable and much quicker when generating the index.
Unfortunately the xapian-haystack package is not available in Fedora/EPEL yet
and waiting for review to be finished.
Link to review bug: https://bugzilla.redhat.com/show_bug.cgi?id=2313507
Signed-off-by: Michal Konecny <mkonecny@redhat.com>
Skopeo fails to login with:
FATA[0000] authenticating creds for "registry.stg.fedoraproject.org": creating new docker client:
missing client certificate registry-fedoraproject.cert for key registry-fedoraproject.key
It seems it is rather picky about the suffix, even though the docs
mentioned both crt and cert being used.
Now that the fulltext search index is finally generated, we need to resize the
LVM again. That should be the last time.
Signed-off-by: Michal Konecny <mkonecny@redhat.com>
This reverts commit 15dbcbb7ac, which was
a revert of commit 5e0ad1134d (pr #2200).
Adam Williamson got rid of the need for buildah so hopefully we won't
run into lots of permission issues.
proxy01/10 were bumped to 16gb a while back, so make sure the ansible
host vars match that.
proxy101/110 (the internal ones used by hosts inside iad2) had only 8g
and were looking pretty loaded. So, bump them up to 16gb as well in
ansible. I already manually increased them to 16g, but this way if we
reinstall them they will be correct.
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
