Traefik (1.6.5 and 1.7.1) stops watching file after it detects Marathon is down #3974
Labels
area/provider/file
area/provider/marathon
kind/bug/confirmed
a confirmed bug (reproducible).
priority/P2
need to be fixed in the future
status/5-frozen-due-to-age
Milestone
Comments
jbdoumenjou
added
kind/bug/possible
|
Hi @daltonmatos, |
|
Sure @jbdoumenjou, I will re-run the tests and post it here. Thanks. |
|
Here is the raw log, just copied from the terminal. All I changed was to redact the domain name. Every |
|
@daltonmatos could you provide the full content (redacted) of your |
|
Hello @ldez, Here is the config: MaxIdleConnsPerHost = 10000
logLevel = "debug"
[lifeCycle]
graceTimeOut = 10
defaultEntryPoints = ["http"]
[entryPoints]
[entryPoints.http]
address = ":80"
[entryPoints.https]
address = ":443"
[entryPoints.https.tls]
[entryPoints.traefik]
address = ":8082"
[file]
watch = true
fileName = "/etc/ssl.toml"
[api]
dashboard = true
[metrics]
[metrics.prometheus]
entrypoint = "traefik"
buckets = [0.1,0.3,1.2,5.0]
[marathon]
endpoint = "http://172.18.0.31:8080"
watch = true
domain = "marathon"
groupsAsSubDomains = true
|
|
It's a problem in the logic, I debuged and reproduce the issue. If the first configuration loaded in [1] is The fix might be to add a loop trying to reload the first configuration from time to time if the initial was null and the watch is on. I'll try make a pr for this change. |
geraldcroes
added
kind/bug/confirmed
diegooliveira
pushed a commit
to diegooliveira/traefik
that referenced
this issue
Nov 1, 2018
If Traefik can't load the configuration when starting it will give up doing so until a new Marathon event gets fired. If no new event is fired then no configuration will ever be loaded, no matter if the watch is true or not. fix traefik#3974
diegooliveira
pushed a commit
to diegooliveira/traefik
that referenced
this issue
Nov 1, 2018
If Traefik can't load the configuration when starting it will give up doing so until a new Marathon event gets fired. If no new event is fired then no configuration will ever be loaded, no matter if the watch is true or not. fix traefik#3974
diegooliveira
pushed a commit
to diegooliveira/traefik
that referenced
this issue
Nov 1, 2018
If Traefik can't load the configuration when starting it will give up doing so until a new Marathon event gets fired. If no new event is fired then no configuration will ever be loaded, no matter if the watch is true or not. fix traefik#3974
This was referenced Nov 1, 2018
|
Closed by #4230. |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Do you want to request a feature or report a bug?
Bug
What did you do?
I'm trying to use the
file.watchfeature. The idea is to be able to reload SSL certificates without restaring traefik.I have the TLS config on a separated file, and here is my main config file (the part that matters):
I have two copies of "/etc/ssl.toml", each one with a different certificate, like this:
ssl1.toml:and
ssl2.toml:And then what I do is just a simple
cp, like this:And check the certificates were reloaded with a simple curl to localhost.
What did you expect to see?
I expected to see traefik change the certificates every time I did the
cp.What did you see instead?
Traefik changes the certificates correctly, until this message appears in the debug log:
As soon as this message appears, traefik does not detect the file change anymore and then the certificates stop being reloaded.
Output of
traefik version: (What version of Traefik are you using?)The version here is
1.6.5but this same behavior happens with1.7.1.What is your environment & configuration (arguments, toml, provider, platform, ...)?
{ "LifeCycle": { "RequestAcceptGraceTimeout": 0, "GraceTimeOut": 0 }, "GraceTimeOut": 0, "Debug": false, "CheckNewVersion": true, "SendAnonymousUsage": false, "AccessLogsFile": "", "AccessLog": null, "TraefikLogsFile": "", "TraefikLog": null, "Tracing": null, "LogLevel": "", "EntryPoints": {}, "Cluster": null, "Constraints": [], "ACME": null, "DefaultEntryPoints": [ "http" ], "ProvidersThrottleDuration": 2000000000, "MaxIdleConnsPerHost": 200, "IdleTimeout": 0, "InsecureSkipVerify": false, "RootCAs": null, "Retry": null, "HealthCheck": { "Interval": 30000000000 }, "RespondingTimeouts": null, "ForwardingTimeouts": null, "AllowMinWeightZero": false, "Web": null, "Docker": null, "File": null, "Marathon": null, "Consul": null, "ConsulCatalog": null, "Etcd": null, "Zookeeper": null, "Boltdb": null, "Kubernetes": null, "Mesos": null, "Eureka": null, "ECS": null, "Rancher": null, "DynamoDB": null, "ServiceFabric": null, "Rest": null, "API": null, "Metrics": null, "Ping": null, "ConfigFile": "" }If applicable, please paste the log output at DEBUG level (
--logLevel=DEBUGswitch)For now on, even changing the contents of
/etc/ssl.toml, traefik does not detect this change and does not reload the certificates.Important Note: I was only able to reproduce this when I started Traefik with Marathon Turned off. When I started traefik with Matathon already started, all works. Even when I remove every app from Marathon and then stop it. but if I later start marathon, Traefik does not start to detect the file changes, although it receives new updates from Marathon:
Let me see if you need any new information and I will provide it.
The text was updated successfully, but these errors were encountered: