FBR: Make rabbitmq.fp.o go to proxy101/110 internally
by Patrick Uiterwijk
Can I get +1s for the patch to the "dns" repo underneath?
This should make "rabbitmq.fp.o" resolve to proxy101/proxy110 internally.
Patrick
diff --git a/fedoraproject.org.template b/fedoraproject.org.template
index b0e659c..eac02a0 100644
--- a/fedoraproject.org.template
+++ b/fedoraproject.org.template
@@ -507,9 +507,15 @@ teams IN TXT "v=spf1 a ip4:35.172.234.62 ?all"
teams-api IN A 35.172.234.62
teams-media-protected IN A 35.172.234.62
teams-static IN A 35.172.234.62
+{% if region == "PHX2" %}
+; Point to proxy101/proxy110
+rabbitmq IN A 10.5.126.8
+rabbitmq IN A 10.5.126.9
+{% else %}
; for now just point rabbitmq at proxy01/10 until/unless rabbitmq
hosts get vpn.
rabbitmq IN A 209.132.181.15
rabbitmq IN A 209.132.181.16
+{% endif %}
rabbitmq.stg IN CNAME admin.stg.fedoraproject.org.
rdu IN A 66.187.233.101
redhat.download IN A 209.132.181.23
5 years, 1 month
FBR: To add coreos-continuous permission to koji hub policy
by Mohan Boddu
Core OS folks wanted to have a separate koji tag to build and tag their
Core OS continuous builds. This has been discussed in the releng ticket -
https://pagure.io/releng/issue/8165.
Also, we tested it staging koji and has been working fine. Here's the
staging koji ansible commit:
https://infrastructure.fedoraproject.org/cgit/ansible.git/commit/?id=c07ff32
Once, this patch is deployed, I will make the necessary koji tags and
permissions as we did in staging koji (which are listed in the releng issue
linked above).
diff --git a/roles/koji_hub/templates/hub.conf.j2
b/roles/koji_hub/templates/hub.conf.j2
index 71792c7..5142491 100644
--- a/roles/koji_hub/templates/hub.conf.j2
+++ b/roles/koji_hub/templates/hub.conf.j2
@@ -96,6 +96,8 @@ tag =
# These two rules makes sure people can't build srpms in infra tags and tag
them into distribution tags
tag *infra* && fromtag *infra* && has_perm infra :: allow
fromtag *infra* :: deny
+ # CoreOS continuous builds, https://pagure.io/releng/issue/8165
+ tag f{{FedoraRawhideNumber}}-coreos-continuous
f{{FedoraBranchedNumber}}-coreos-continuous f{{Fedora
CycleNumber}}-coreos-continuous
f{{FedoraPreviousCycleNumber}}-coreos-continuous && has_perm coreos-conti
nuous && match action add unblock block :: allow
all :: allow
channel =
If you have any questions, please let me know.
Thanks.
5 years, 1 month
FBR: Enable central http logging for internal proxies
by Patrick Uiterwijk
Hi,
Can I get +1s for applying the underneath patch and running the
logserver.yml playbook?
This will make sure we also capture the proxy101/proxy110 http logs on log01.
Patrick
diff --git a/roles/base/templates/syncHttpLogs.sh.j2
b/roles/base/templates/syncHttpLogs.sh.j2
index d83f63d65..8cca68d5d 100644
--- a/roles/base/templates/syncHttpLogs.sh.j2
+++ b/roles/base/templates/syncHttpLogs.sh.j2
@@ -33,6 +33,11 @@ function syncHttpLogs {
syncHttpLogs {{host}}
{% endfor %}
+## Sync up all internal proxies
+{% for host in groups['proxies-internal'] %}
+syncHttpLogs {{host}}
+{% endfor %}
+
## Sync up all staging proxies
{% for host in groups['proxies-stg'] %}
syncHttpLogs {{host}}
5 years, 1 month
Need someone to run Thursday meeting
by Stephen John Smoogen
I will be gone Wed->Sun this week and will need someone to run the
Thursday Infrastructure meeting. The basics of the meeting are in the
gobby-0.5 on infinote.fedoraproject.org
On Wednesday, call for items to be added to the document on IRC in
#fedora-admin, #fedora-apps, #fedora-noc, #fedora-releng . Send out an
email around 20:00 UTC.
On Thursday, go to IRC an hour before meeting and remind people on
#fedora-admin, #fedora-apps, #fedora-noc, #fedora-releng . Start the
meeting and follow the form. After each section either mark sub-items
to be used next meeting or remove it.
Thanks
--
Stephen J Smoogen.
5 years, 1 month
FBR clean up repospanner on batcave01/pagure01
by Kevin Fenzi
Greetings.
Back a long time ago we setup a repospanner cluster for our ansible
repo. We used repospanner01, batcave01 and pagure01 in this cluster.
However, there were some things that needed fixing up before we could
use it, so we never used it.
Fast forward to todaay and we have setup a new cluster for src, but will
also just reuse it for ansible repo as well. That cluster has
repospanner01 (reinstalled), repospanner-osuosl01, and
repospanner-cc-rdu01 in it.
That is all great, however, we still have repospanner installed and
trying to cluster on batcave01 and pagure01. it's sending many messages
per second about it's missing repospanner01 cluster member being gone.
So, I would like to apply the following to ansible, stop and remove the
repospanner package from batcave01/pagure01. It shouldn't cause any
issues I wouldn't think.
+1s?
kevin
--
> diff --git a/playbooks/groups/batcave.yml b/playbooks/groups/batcave.yml
> index e0f7213..fe0f450 100644
> --- a/playbooks/groups/batcave.yml
> +++ b/playbooks/groups/batcave.yml
> @@ -31,7 +31,6 @@
> SSLCertificateChainFile: "{{wildcard_int_file}}"
> - openvpn/client
> - batcave
> - - { role: repospanner/server, when: inventory_hostname.startswith('batcave01'), node: batcave01, region: ansible, spawn_repospanner_node: false, join_repospanner_node: repospanner01.ansible.fedoraproject.org }
> - { role: nfs/client, when: inventory_hostname.startswith('batcave'), mnt_dir: '/srv/web/pub', nfs_src_dir: 'fedora_ftp/fedora.redhat.com/pub' }
> - { role: nfs/client, when: inventory_hostname.startswith('batcave01'), mnt_dir: '/mnt/fedora/app', nfs_src_dir: 'fedora_app/app' }
>
> diff --git a/playbooks/groups/pagure.yml b/playbooks/groups/pagure.yml
> index a6dd601..d9f8c29 100644
> --- a/playbooks/groups/pagure.yml
> +++ b/playbooks/groups/pagure.yml
> @@ -56,7 +56,6 @@
> roles:
> - pagure/frontend
> - pagure/fedmsg
> - - { role: repospanner/server, when: inventory_hostname.startswith('pagure01'), node: pagure01, region: ansible, spawn_repospanner_node: false, join_repospanner_node: repospanner01.ansible.fedoraproject.org }
>
> handlers:
> - import_tasks: "{{ handlers_path }}/restart_services.yml"
5 years, 1 month
FBR: Push a fix to clean_amis.py in fedimg box
by Sayan Chowdhury
I need to push a fix to run the clean_amis script. The diff is as follows:
diff --git a/roles/fedimg/files/clean-amis.py b/roles/fedimg/files/clean-amis.py
index d0a491824..8e8cfff91 100644
--- a/roles/fedimg/files/clean-amis.py
+++ b/roles/fedimg/files/clean-amis.py
@@ -116,7 +116,7 @@ def get_page(page, delta, topic, start=None, end=None):
return resp.json()
-def _get_two_week_atomic_released_compose_id(delta, start=None, end=None):
+def _get_two_week_released_atomic_compose_id(delta, start=None, end=None):
""" Returns the release compose ids for last n days """
topic = "org.fedoraproject.prod.releng.atomic.twoweek.complete"
@@ -136,7 +136,7 @@ def
_get_two_week_atomic_released_compose_id(delta, start=None, end=None):
for msg in messages:
# This is to support the older-format fedmsg messages
if "atomic_raw" in msg:
- released_atomic_compose_ids.append(["atomic_raw"]["compose_id"])
+ released_atomic_compose_ids.append(msg["atomic_raw"]["compose_id"])
# We are just trying here multiple archs to get the compose id
elif "aarch64" in msg:
released_atomic_compose_ids.append(
@@ -186,7 +186,7 @@ def _get_nightly_amis_nd(delta, start=None, end=None):
# Sometimes the compose label is None
# and they can be blindly put in for deletion
- if compose_label is None:
+ if not compose_label:
amis.append((compose_id, ami_id, region))
if compose_id in released_atomic_compose_ids:
@@ -209,7 +209,7 @@ def _get_nightly_amis_nd(delta, start=None, end=None):
def delete_amis_nd(deletetimestamp, dry_run=False):
""" Delete the give list of nightly AMIs
- :args deletetimestamp: the timestamp for the delete
+ :args deletetimestamp: the timestamp for the delete
:args dry_run: dry run the flow
"""
log.info("Deleting AMIs")
The PR to this diff is merged into releng repo[1].
+1s
[1] https://pagure.io/releng/pull-request/8193#request_diff
--
Sayan Chowdhury <https://words.yudocaa.in/>
GPG Fingerprint : 0F16 E841 E517 225C 7D13 AB3C B023 9931 9CD0 5C8B
5 years, 1 month