The httpd logs for IPA are large and are not compressed with default
logrotate config. This will add a logrotate config that compress them.
Signed-off-by: Michal Konecny <mkonecny@redhat.com>
These will be replacing the rhel9 versions next week on the outage on
thursday. The basic process will be to sync data from the old ones
then enable checksums with pg_checksum, then upgrade to postgresql18.
Finally the old ones will be taken down and the new ones will take over
the old name/ip (to avoid firewall complications).
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
This commit puts a proxy redirect into place, which will redirect all
traffic to docs.fp.o/ai/ to docs.fp.o/ai-ml/. This matches the name of
the SIG and helps us with the change in the Antora component name, which
was discussed in today's AI/ML SIG meeting. The redirect helps to make
sure we don't lose any web traffic for folks who have used the old link
since the site was launched in March 2026.
ref: ai-ml/docs#4
Signed-off-by: Justin Wheeler <jwheel@redhat.com>
/etc/copr/copr.conf:206: SyntaxWarning: "\." is an invalid escape sequence. Such sequences will not work in the future. Did you mean "\\."? A raw string is also an option.
"pattern": "^@python/python3\.\d+/.*/root$",
The default logging is INFO, and it's very noisy.
While we aren't actively debugging anything, lets set it back to WARN.
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
This instance will eventually replace db-datanommer02
(next thursday hopefully). In the mean time we can start syncing the
data to it so we can update timescaledb on 02 and still have a way to
back out in case of doom.
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
Revert commit 6ef9d68439 which removed resultsdb references without
first deleting the OpenShift resources, leaving them orphaned.
Add openshift/object-delete roles gated with [never, delete-resultsdb]
tags to properly clean up the resultsdb resources from OpenShift.
To delete the resources, run:
sudo rbac-playbook -l os_control_stg[0] -t delete-resultsdb openshift-apps/testdays.yml
sudo rbac-playbook -l os_control[0] -t delete-resultsdb openshift-apps/testdays.yml
Assisted-by: Claude Opus 4.6 (via OpenCode)
Right now the default number of open files is 1024 and running a bunch
of playbooks hits that limit. So, just increase it to 64k to avoid that.
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
db-datanommer01.stg is now rhel10 and postgresql18 and timescaledb
2.30.0.
We dont need the other two transitory ones left around now.
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
Enable SSH key synchronization from Fedora Accounts OIDC for staging
environment to test with multiple keys.
Changes:
- Add attribute-ssh-public-key: ssh_key for staging
- Add fas-attributes scope to OPENID_CONNECT_SCOPES for staging
- Forgejo code already handles both string (1 key) and array (2+ keys)
Testing: User with 2 SSH keys in FAS to verify array handling.
This one is getting scm-commits archives page by page (now on page
176504). Thats causing a lot of db query overhead.
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
This is an aws ip that is scraping bodhi with things like:
"GET /updates/?type=security&rows_per_page=200&page=51 HTTP/1.1" 200 2176010 "-" "python-requests/2.34.2"
Where the page is up to thousands after a while.
These are very large and hit the db server very hard when it has to
search for really really old security updates.
Signed-off-by: Kevin Fenzi <kevin@scrye.com>