WarlockFish/fedora-infra_ansible
1
0
Fork 0
mirror of https://pagure.io/fedora-infra/ansible.git synced 2026-04-25 11:00:32 +08:00
Code Issues Packages Projects Releases Wiki Activity
43,735 Commits 9 Branches 0 Tags
7da433b68cb95e8c12a81ea4eee00523e8a87cd6
Commit Graph

43735 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Aurélien Bompard
7da433b68c fedocal: use a more recent version of Python for prod too 
Signed-off-by: Aurélien Bompard <aurelien@bompard.org>
 2025-07-07 10:31:03 +02:00
Kevin Fenzi
cb3a9a1bc7 proxies: add a tag for meetbot-raw config 
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
 2025-07-06 18:50:35 -07:00
Carl George
0f03b772e5 roles/bodhi2: Disable delta RPMs for EPEL 8 
We previously disabled delta RPMs for EPEL 9 and 10.  It looks like we
should disable this for EPEL 8 as well, because Pulp 3 (part of
Satellite 6.17) intentionally blocks mirror syncing of repos with delta
RPMs.

Related: https://pagure.io/fedora-infra/ansible/pull-request/2391
Related: https://community.theforeman.org/t/content-management-suse-drpm-files-not-syncing/25243
Resolves: https://pagure.io/releng/issue/12790
 2025-07-06 22:26:04 +00:00
Kevin Fenzi
ae2432c684 people: increase quota for dusty ( ticket 12614 ) 
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
 2025-07-06 15:23:23 -07:00
František Zatloukal
7d54a27e26 Blockerbugs: project_alert_users: kparal 2025-07-06 21:54:25 +02:00
Yaakov Selkowitz
f3fcf8fe01 releng: fix /mnt/koji on compose-eln 
This appears to be a thinko; stg compose machines use /mnt/fedora_koji_prod.

Signed-off-by: Yaakov Selkowitz <yselkowi@redhat.com>
 2025-07-06 13:21:28 -04:00
Kevin Fenzi
3a9405cd77 bvirthost: increase default number of processes before alerting 
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
 2025-07-06 09:37:40 -07:00
Kevin Fenzi
2de770d167 secondary01: switch to nfsv4 mounts 
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
 2025-07-06 09:32:34 -07:00
Kevin Fenzi
cb3d019169 compose-eln: also need the /pub mount here 
We need this mount to sync eln composes out.

Signed-off-by: Kevin Fenzi <kevin@scrye.com>
 2025-07-06 08:06:33 -07:00
Mikolaj Izdebski
3ce996aac2 staging-sync/koschei: Update rawhide from f41 to f43 2025-07-05 11:24:23 +02:00
Mikolaj Izdebski
9dc5c9a47a staging-sync/koji: Don't mix up mount and umount 2025-07-05 11:10:11 +02:00
Mikolaj Izdebski
1723796680 staging-sync/koji: Make sure mount point directories are created 2025-07-05 11:06:20 +02:00
Mikolaj Izdebski
ebcac361a5 staging-sync/koji: Try to remount with plain old shell 2025-07-05 11:03:19 +02:00
Mikolaj Izdebski
2a3d83de77 staging-sync/koji: Fix re-mounting existing filesystems 2025-07-05 11:01:10 +02:00
Mikolaj Izdebski
3ca066f146 staging-sync/koji: Update /mnt/koji cleanup code 2025-07-05 10:55:58 +02:00
Kevin Fenzi
a387990755 value: we deployed value01 in rdu3, not value02 
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
 2025-07-04 14:09:33 -07:00
Yaakov Selkowitz
787cab1edd releng: fix group_names conditional for pungi-utils 
Signed-off-by: Yaakov Selkowitz <yselkowi@redhat.com>
 2025-07-04 20:40:52 +00:00
Mikolaj Izdebski
d3cf5104ed db-koji01.stg: increase disk size to match prod 2025-07-04 21:12:32 +02:00
Kevin Fenzi
6e16f8eeb6 certgetter: there is only one certgetter in rdu3, not one in every datacenter. Just specify the correct one or proxy playbooks fail for all the proxies in other datacenters 
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
 2025-07-04 11:07:41 -07:00
Aurélien Bompard
e0e8bd9816 FMN: adjust the cache lifetime to the rebuild duration 
Signed-off-by: Aurélien Bompard <aurelien@bompard.org>
 2025-07-04 19:46:40 +02:00
Mikolaj Izdebski
8da3e395b2 postgresql: Install mbuffer package 
mbuffer is required on db-koji for staging-sync playbook, but
installing it on all DB servers doesn't hurt.
 2025-07-04 19:44:32 +02:00
Aurélien Bompard
8d86dbc5e2 FMN: use the fqdn for log01 in collectd 
Signed-off-by: Aurélien Bompard <aurelien@bompard.org>
 2025-07-04 19:36:02 +02:00
Mikolaj Izdebski
c39cb62915 staging-sync/koji: Bump DB sequence numbers 
Starting IDs for task/build etc. must be greater than those in prod.
Also clean up some bogus constructs.
Comment about koji regen-repo is no longer valid since Koji implements
dynamic on-demand repo regeneration.
 2025-07-04 19:30:57 +02:00
Mikolaj Izdebski
b4e77e3a71 koji: Drop legacy mash.conf for httpd 2025-07-04 19:04:25 +02:00
Aurélien Bompard
ea26e54ca0 Drop the mirror_pagure_ansible_13 rabbitmq queue 
Signed-off-by: Aurélien Bompard <aurelien@bompard.org>
 2025-07-04 18:40:04 +02:00
Mikolaj Izdebski
5a926c4076 Koschei: deploy backend 
This reverts commit b1c16b683f.
 2025-07-04 18:30:31 +02:00
Mikolaj Izdebski
1c82e912d1 koji: Install GSSAPI keytab also in staging 2025-07-04 18:24:59 +02:00
Mikolaj Izdebski
b1c16b683f Koschei: deploy frontend only, skip backend for now 2025-07-04 17:54:19 +02:00
Jiri Kyjovsky
b7acf000ae copr-backend: actually keep 10gb space for root 2025-07-04 17:42:44 +02:00
Kevin Fenzi
7b3d95ee86 provision: needs to work in stg too 
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
 2025-07-04 08:20:30 -07:00
Greg Sutcliffe
08d3f43ab1 Zabbix: disable templates for now, revisit after DC move completed 
Signed-off-by: Greg Sutcliffe <fedora@emeraldreverie.org>
 2025-07-04 16:16:34 +01:00
Jiri Kyjovsky
a4050477c2 copr-backend: fix typo for powerful builders to claim we have power10 2025-07-04 17:06:12 +02:00
Jiri Kyjovsky
f41f602366 copr-backend: setup swap for powerful builders in osuosl 2025-07-04 17:06:12 +02:00
Dusty Mabe
348a3d72d4 Revert "proxies: move coreos cincinnati to rdu3 openshift" 
Now that the datacenter move has happened this is no longer
needed and I think is actually causing the endpoints to be
unreachable for servers trying to update.

This reverts commit 713af9ea3a.
 2025-07-04 14:51:54 +00:00
Michal Konecny
17d33b34cd [haproxy] Let's use the correct cert for ipa 
HAProxy had incorrect IPA certificate for staging. I'm not sure how that
even worked, but the issue was revealed when the IAD2 machines were
removed from cluster.
 2025-07-04 16:09:24 +02:00
Michal Konecny
ea141b651a Add zabbix01.stg to staging group 2025-07-04 14:35:52 +02:00
Michal Konecny
be2e3a6d35 [nfs] Fix the nfs related errors 
* intr parameter is now deprecated and it's no longer effective in
newer kernels

* nfs-server needs to run as well on Fedora otherwise the mount request
is rejected
 2025-07-04 10:38:17 +02:00
Adam Williamson
3fc124f048 Fix rabbitmq 'pungi' user creation on prod 
I think @kevin inadvertently broke this in
f0663ae52f - he mixed up the names
of environments ('production' / 'staging') and Koji instances
('primary' / 'secondary'), just in this one case (it's correct
everywhere else). This is causing the pungi user not to be
created, and that means compose hosts can't publish messages.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
 2025-07-03 23:43:05 -07:00
Adam Williamson
da51de6283 proxies-websites: add more tags 
So I can fix the worker06 bug for the sites it's annoying me on,
at least...

Signed-off-by: Adam Williamson <awilliam@redhat.com>
 2025-07-03 23:09:56 -07:00
Kevin Fenzi
079bb7a6b5 releng-compose: rawhide cron is a template 
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
 2025-07-03 19:48:05 -07:00
James Antill
2a76cdf5a9 distgit: Deploy fedora-messaging conf and certs. 
Signed-off-by: James Antill <james@and.org>
 2025-07-03 17:05:48 -04:00
Kevin Fenzi
1b67cfcf3b releng-compose: filter some more users that should be local 
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
 2025-07-03 13:58:56 -07:00
Michal Konecny
bc8d9ab946 [koji_builders] Don't recreate /mnk/fedora_koji 
This folder is already created during execution of nfs/client role,
recreating it will just make empty folder.
 2025-07-03 22:54:55 +02:00
Michal Konecny
5703285385 Add staging builders to staging group 2025-07-03 20:13:46 +00:00
Adam Williamson
7df8f49847 openqa: assign a tap2 x86 worker for lab 
Whoops, somehow overlooked this or lost it in an edit...

Signed-off-by: Adam Williamson <awilliam@redhat.com>
 2025-07-03 13:09:38 -07:00
Kevin Fenzi
28da1d6cf7 bodhi-backend: switch mirror to check for sync to the generic, rdu3 ones instead of hard coding iad2 external 
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
 2025-07-03 13:08:11 -07:00
Michal Konecny
df86e98c4d Add missing staging buildvm host vars 2025-07-03 21:22:25 +02:00
Adam Williamson
64ffac4caf haproxy: only proxy rabbitmq on rdu3 proxies 
Prior to 38d138e this condition existed with 'iad2' instead of
'rdu3'. @abompard took it out entirely, but that was wrong, it
makes the external proxies include this block. We need to put the
condition back with the correct data center name.

Signed-off-by: Adam Williamson <awilliam@redhat.com>
 2025-07-03 12:05:00 -07:00
Adam Williamson
9da2cfb6f2 haproxy: IPA certs don't depend on data center 
The IPA cert doesn't change when we move datacenters, because we
just replicate across. So it shouldn't have the datacenter in the
name. This should fix haproxy deployment (it was broken because
we didn't have an 'rdu3' file).

Signed-off-by: Adam Williamson <awilliam@redhat.com>
 2025-07-03 11:55:59 -07:00
Kevin Fenzi
1f05949fbc add some tags to login-registry 
Signed-off-by: Kevin Fenzi <kevin@scrye.com>
 2025-07-03 11:50:58 -07:00