These are chat archives for spring-cloud/spring-cloud

31st
May 2018
Spencer Gibb
@spencergibb
May 31 2018 00:12
listOfServers should have a host and port
pradeepsingh1234
@pradeepsingh1234
May 31 2018 00:36
Whenever I restart my eureka server apigateway start throwing internal 500 error.. I thought that ribbon maintains the server list forever if Eureka is down.
pradeepsingh1234
@pradeepsingh1234
May 31 2018 00:58
what is the correct behavior?
Joshua Street
@jjstreet
May 31 2018 01:09
@spencergibb ah
Adding the port number causes my Zuul routes to no longer function
but the feign client functions properly
pradeepsingh1234
@pradeepsingh1234
May 31 2018 01:11
@spencergibb Whenever I restart my eureka server apigateway start throwing internal 500 error.. I thought that ribbon maintains the server list forever if Eureka is down.
Joshua Street
@jjstreet
May 31 2018 01:24
zuul immediately sends a response back from my hystrix fallback configuration
very strange
im disabling my fall back to see whats goin on
Joshua Street
@jjstreet
May 31 2018 01:44
i think i've just confused myself further
i have a configuration:
someService:
  ribbon:
    listOfServers: https://someService.cloud.com
this service someService is set by a filter in a Zuul proxy server
for requests
with the above configuration, it works just fine
i also have to use the same service identifier to poll for some data with a feign client
problem is, my service called someService could be named anything
Joshua Street
@jjstreet
May 31 2018 01:49
so i have another section of config
myApp.polling:
  services: someService
readyli
@readyli
May 31 2018 01:49
Distributed tracking what do you use?
zipkin?
Joshua Street
@jjstreet
May 31 2018 01:50
@readyli who are you asking?
my polling feign client does not work
readyli
@readyli
May 31 2018 01:52
everyone
Joshua Street
@jjstreet
May 31 2018 01:52
unless my listOfServers entries include the port number as @spencergibb recommends
so i'm at a loss on this issue
@readyli zipkin is a library for distributed tracing yess
readyli
@readyli
May 31 2018 01:54
thanks,learn technology from everyone and learn English by the way
power-man
@power-man
May 31 2018 02:20
@readyli we are family
readyli
@readyli
May 31 2018 02:22
图片.png
@power-man
power-man
@power-man
May 31 2018 02:24
@readyli oh my god, you find me!
@readyli how could do that?
readyli
@readyli
May 31 2018 02:27
haha
power-man
@power-man
May 31 2018 02:30
@readyli I am learning springboot 2.0 now.But something wrong when I build my project.like this

APPLICATION FAILED TO START


Description:

The Bean Validation API is on the classpath but no implementation could be found

Action:

Add an implementation, such as Hibernate Validator, to the classpath

I shoud clean my repo that others told me.
readyli
@readyli
May 31 2018 02:34
send me of your project,i'll see
console details
Joshua Street
@jjstreet
May 31 2018 02:36
i think i gotta put an issue in on this list of server stuff
once i create an example project
i can't get it to work no matter what i do
kandyjam
@kandyjam
May 31 2018 05:21
Hi,anyone knows how to connect to config server in version Finchley.RC2 ?
cloud:
config:
uri: http://localhost:8088
spring:
cloud:
config:
uri: http://localhost:8088
this config not work
Frederick
@silentFred
May 31 2018 06:44
Hi all. Anyone have experience with zuul using eureka (running Camden.SR7). Timeouts for days and most ribbon and hysterix config is not picked up
Marcin Grzejszczak
@marcingrzejszczak
May 31 2018 07:13
@silentFred maybe try bumping to anything relatively new (Edgware.SR3) or at least Dalston to see if things got fixed.
kandyjam
@kandyjam
May 31 2018 07:15
I get it, I forgot to add spring-cloud-starter-config dependency.
CH4:D
@chad_d_stud_twitter
May 31 2018 07:31
is it a good idea to put eureka in a docker container?
we have an issue with an AWS dockerized eureka instance. if we kill services registered to it, eureka still thinks they’re up. they actually STAY up. it never expires.
worth noting that the same eureka instance as a jar works fine in a local machine
Frederick
@silentFred
May 31 2018 07:35
@marcingrzejszczak will do. Our entire stack runs on Camden.SR7. Any risk to run zuul on a later version than our other components? (eureka and services)
it's unrealistic for us to bump all our services (we have 50 +)
big system
also everything still running on boot 1.5.10.RELEASE
pradeepsingh1234
@pradeepsingh1234
May 31 2018 13:30
@spencergibb as per info on this link client(ribbon) retains the eureka serverlist indefinite..spring-cloud/spring-cloud-netflix#1095
but I don't see samebehavior
whenever I shutdown/restart my eureka server, I see my apigateway start throwing 500 internal serer error
Is there any proeprty which force the ribbon to refresh the serverList while eureka is down?
Another question i have is there anyway to tell the eureka to ignore registry with certain keywords, like : localhost
Tyler Van Gorder
@tkvangorder
May 31 2018 13:49
@chad_d_stud_twitter It sounds like you might be running into Eureka entering self-preservation mode. We turn self-preservation mode "off" in our local development environments because we are often restarting/bouncing services.
CH4:D
@chad_d_stud_twitter
May 31 2018 14:42
@tkvangorder actually, we just discovered this is a Eureka bug where it doesn’t deregister a service that was involved in an unclean shutdown
You can replicate this issue when you SIGKILL a running spring boot instance
What we’re trying to figure out now is how to shutdown the app BEFORE docker restarts.
Stopping the docker instance directly doesn’t give the app the chance to shutdown and deregister itself
Sushant
@bsushant-athena
May 31 2018 16:25
@marcosbarbero spring-cloud-config-server is similar to Netflix Archaius ?
Marcos Barbero
@marcosbarbero
May 31 2018 16:33
Hi @bsushant-athena, it's similar but it has some differences, for instance Archaius has its own set of configuration files and loading priorities and it differs from spring apps. You can read more about that here
Tyler Van Gorder
@tkvangorder
May 31 2018 17:52
@chad_d_stud_twitter That actually makes sense, if a service is not gracefully shutdown, I would expect the service to "go away" after it misses consecutive heart beats. How long is your heart beat interval? I am not sure of your exact use case, but your down stream clients (ribbon?) need to be resilient to servers sticking around in Eureka but no longer being alive.
CH4:D
@chad_d_stud_twitter
May 31 2018 21:16
Our heartbeat is rather short @tkvangorder. 10s. But again this is a eureka issue. It stops sending heart beats for dead instances but doesn't register it. There's a GitHub issue for that that was filed last Feb 2017 and has remained open since April if this year.