configs/google_com_proxy.v2.yaml is not working anymore

[winguse]

I find I cannot turn http traffic to upstream https server any more. By digging into details, I also find the example config is not working as well.

repo steps:

$ envoy -c google_com_proxy.v2.yaml

and check the result:

$ curl -i http://127.0.0.1:10000
HTTP/1.1 503 Service Unavailable
content-length: 57
content-type: text/plain
date: Wed, 31 Jan 2018 00:30:09 GMT
server: envoy

upstream connect error or disconnect/reset before headers

[alyssawilk]

Hm, it works locally for me with
git pull
bazel build source/exe/envoy
bazel-bin/source/exe/envoy-static -c configs/google_com_proxy.v2.yaml

Can you verify you're running against head, and post with Envoy's logs?
Just to sanity check, I assume a direct curl http://www.google.com works for you?

[winguse]

sorry for i did not provide significant information.

I am testing on a aws ec2 machine and a direct curl http://www.google.com works.

as i check the the cluster info, i find this:

$ curl 127.0.0.1:9901/clusters
version_info::static
service_google::default_priority::max_connections::1024
service_google::default_priority::max_pending_requests::1024
service_google::default_priority::max_requests::1024
service_google::default_priority::max_retries::3
service_google::high_priority::max_connections::1024
service_google::high_priority::max_pending_requests::1024
service_google::high_priority::max_requests::1024
service_google::high_priority::max_retries::3
service_google::added_via_api::false
service_google::[::]:0::cx_active::0
service_google::[::]:0::cx_connect_fail::1
service_google::[::]:0::cx_total::1
service_google::[::]:0::rq_active::0
service_google::[::]:0::rq_error::1
service_google::[::]:0::rq_success::0
service_google::[::]:0::rq_timeout::0
service_google::[::]:0::rq_total::0
service_google::[::]:0::health_flags::healthy
service_google::[::]:0::weight::1
service_google::[::]:0::region::
service_google::[::]:0::zone::
service_google::[::]:0::sub_zone::
service_google::[::]:0::canary::false
service_google::[::]:0::success_rate::-1

[2018-01-31T00:31:58.289Z] "POST / HTTP/1.1" 503 UF 0 57 0 - "-" "curl/7.47.0" "73a1de92-820e-4dcc-b634-37c9a9f5170a" "www.google.com" "[2607:f8b0:400a:809::200e]:443"

it looks envoy only looking into ipv6 address of google. but my instance do not have ipv6 network. and the document is tell us default is AUTO.

and the dns query result as below (looks ipv4 by default):

$ dig google.com

; <<>> DiG 9.10.3-P4-Ubuntu <<>> google.com
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 51420
;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 1

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
;; QUESTION SECTION:
;google.com.                    IN      A

;; ANSWER SECTION:
google.com.             46      IN      A       216.58.211.174

;; Query time: 0 msec
;; SERVER: 10.0.0.2#53(10.0.0.2)
;; WHEN: Wed Jan 31 00:33:24 UTC 2018
;; MSG SIZE  rcvd: 55

by setting dns_lookup_family: V4_ONLY it works now.

[winguse]

but why the cluster info is not the currect ips of google?

$ curl 127.0.0.1:9901/clusters
version_info::static
service_google::default_priority::max_connections::1024
service_google::default_priority::max_pending_requests::1024
service_google::default_priority::max_requests::1024
service_google::default_priority::max_retries::3
service_google::high_priority::max_connections::1024
service_google::high_priority::max_pending_requests::1024
service_google::high_priority::max_requests::1024
service_google::high_priority::max_retries::3
service_google::added_via_api::false
service_google::0.0.0.0:0::cx_active::0
service_google::0.0.0.0:0::cx_connect_fail::0
service_google::0.0.0.0:0::cx_total::0
service_google::0.0.0.0:0::rq_active::0
service_google::0.0.0.0:0::rq_error::0
service_google::0.0.0.0:0::rq_success::0
service_google::0.0.0.0:0::rq_timeout::0
service_google::0.0.0.0:0::rq_total::0
service_google::0.0.0.0:0::health_flags::healthy
service_google::0.0.0.0:0::weight::1
service_google::0.0.0.0:0::region::
service_google::0.0.0.0:0::zone::
service_google::0.0.0.0:0::sub_zone::
service_google::0.0.0.0:0::canary::false
service_google::0.0.0.0:0::success_rate::-1

[mattklein123]

@winguse 0.0.0.0 is due to type: LOGICAL_DNS in the config. (Please read about it in the docs).

For the v4 vs v6 thing, perhaps you could add a YAML comment to the config file that might help someone who encounters this in the future? Thank you for letting us know the problem.

[mattklein123]

Comment added to sample config.