Compare commits

...

12 Commits

Author SHA1 Message Date
3wc
7835b585fd chore: publish 3.1.1+v2.11.25 release 2025-06-11 18:23:53 +01:00
830559895e chore: publish 3.1.0+v2.11.24 release 2025-04-22 15:28:32 +02:00
ac53e9debe
chore: publish 3.0.0+v2.11.22 with release note 2025-04-22 09:27:08 +02:00
acb4c6960a
feat: Switch to endpoint-mode dnsrr instead of vip
The default docker swarm endpoint mode (vip) introduces unnecessary
indirection in the communication between services, namely the
docker-proxy and a dynamic haproxy endpoint container. This commit
switches the socket-proxy service to endpoint_mode: dnsrr by default and
the traefik service when using host-mode port publishing.

I would strongly recommend considering switching to host-mode port
publishing by default, especially as most coop-cloud deployments are
single-server.

See: toolshed/organising#648

Thanks to @mirsal. Rebased and merged from the following commit.
abbb3255f8
2025-04-22 09:04:43 +02:00
22578d1e8e chore: publish 2.10.0+v2.11.22 release 2025-04-01 16:00:35 +02:00
55ad530fb7 chore: publish 2.9.1+v2.11.14 release 2025-02-21 18:42:22 +01:00
54fe45da2f Revert max log max log retention 2025-02-21 18:40:16 +01:00
e21dbc655a fix default values and breaking configuration for LOG_MAX_AGE change 2025-02-20 14:42:13 -05:00
b9d825b5c5 publish new version 2025-02-19 17:21:22 -05:00
74b3ee6716 chore: publish 3.1.0+v2.11.14 release 2025-02-19 17:20:04 -05:00
14d5d79520 Merge pull request 'Expose max log retention in traefik.yml' (#51) from sixsmith/traefik:master into master
Reviewed-on: coop-cloud/traefik#51
Reviewed-by: marlon <marlon@riseup.net>
2025-02-19 22:15:04 +00:00
abbb3255f8
Switch to endpoint-mode dnsrr instead of vip
The default docker swarm endpoint mode (vip) introduces unnecessary
indirection in the communication between services, namely the
docker-proxy and a dynamic haproxy endpoint container. This commit
switches the socket-proxy service to endpoint_mode: dnsrr by default and
the traefik service when using host-mode port publishing.

I would strongly recommend considering switching to host-mode port
publishing by default, especially as most coop-cloud deployments are
single-server.

See: toolshed/organising#648
2024-11-27 16:42:35 +00:00
5 changed files with 7 additions and 6 deletions

View File

@ -10,7 +10,6 @@ LETS_ENCRYPT_EMAIL=certs@example.com
# DASHBOARD_ENABLED=true
# WARN, INFO etc.
LOG_LEVEL=WARN
LOG_MAX_AGE=1
# This is here so later lines can extend it; you likely don't wanna edit
COMPOSE_FILE="compose.yml"

View File

@ -3,7 +3,7 @@ version: "3.8"
services:
app:
image: "traefik:v2.11.14"
image: "traefik:v2.11.25"
# Note(decentral1se): *please do not* add any additional ports here.
# Doing so could break new installs with port conflicts. Please use
# the usual `compose.$app.yml` approach for any additional ports
@ -27,7 +27,6 @@ services:
environment:
- DASHBOARD_ENABLED
- LOG_LEVEL
- LOG_MAX_AGE
healthcheck:
test: ["CMD", "traefik", "healthcheck"]
interval: 30s
@ -48,12 +47,14 @@ services:
- "traefik.http.routers.${STACK_NAME}.tls.certresolver=${LETS_ENCRYPT_ENV}"
- "traefik.http.routers.${STACK_NAME}.service=api@internal"
- "traefik.http.routers.${STACK_NAME}.middlewares=security@file"
- "coop-cloud.${STACK_NAME}.version=2.9.0+v2.11.14"
- "coop-cloud.${STACK_NAME}.version=3.1.1+v2.11.25"
- "coop-cloud.${STACK_NAME}.timeout=${TIMEOUT:-120}"
- "backupbot.backup=${ENABLE_BACKUPS:-true}"
socket-proxy:
image: lscr.io/linuxserver/socket-proxy:1.26.2-r0-ls30
deploy:
endpoint_mode: dnsrr
environment:
- ALLOW_START=0
- ALLOW_STOP=0

1
release/2.9.1+v2.11.14 Normal file
View File

@ -0,0 +1 @@
Reverts max log retention

2
release/3.0.0+v2.11.22 Normal file
View File

@ -0,0 +1,2 @@
socket-proxy: switch to endpoint-mode dnsrr instead of vip
See https://git.coopcloud.tech/coop-cloud/traefik/pulls/50.

View File

@ -1,8 +1,6 @@
---
log:
level: {{ env "LOG_LEVEL" }}
maxAge: {{ env "LOG_MAX_AGE" }}
providers:
docker: