You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@openwhisk.apache.org by OpenWhisk Team Slack <ra...@apache.org> on 2019/02/26 07:17:11 UTC

[slack-digest] [2019-02-25] #general

2019-02-25 01:02:38 UTC - Mark Deuser: @Andrei Alecu - it's likely you are missing some parameters on the post request.  use the `-v` option with the cli to see the parameters the cli passes to the createApi endpoint
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551056558080200?thread_ts=1551048160.079900&cid=C3TPCAQG1
----
2019-02-25 07:16:04 UTC - Markus Thömmes: <https://gravitational.com/blog/serverless-on-kubernetes/> :+1:
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551078964080700
----
2019-02-25 10:26:22 UTC - Dominic Kim: If there are a huge number of invokers with less invoker memory, how can controllers schedule bigger memory actions?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551090382081800
----
2019-02-25 10:27:09 UTC - Dominic Kim: For example, if invoker machines have 10GB memory, and there are 100 controllers, each controller will have 128MB slot for each invoker.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551090429082700
----
2019-02-25 10:27:19 UTC - Dominic Kim: Then how can they schedule 512MB actions?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551090439083100?thread_ts=1551090439.083100&cid=C3TPCAQG1
----
2019-02-25 10:29:16 UTC - Andrei Alecu: @Mark Deuser this is with the serverless framework using `serverless deploy`. `wsk` itself does deploy things properly. Did I misunderstood your suggestion?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551090556083200?thread_ts=1551048160.079900&cid=C3TPCAQG1
----
2019-02-25 10:35:59 UTC - Andrei Alecu: I have a problem with openwhisk on kubernetes on my dev pc, I'm running on Docker Desktop via its built in kubernetes and installed via helm. Whenever I restart docker, the openwhisk pods don't come back up properly - some are stuck in `Init:0/1` status. The only way to get it back up is do run `helm delete` and then reinstall it. Any suggestions?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551090959085200
----
2019-02-25 10:36:52 UTC - Andrei Alecu: 
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551091012085300
----
2019-02-25 10:45:11 UTC - Markus Thömmes: They can't under the current sharding strategy. When that was first proposed Christian and I mentioned that we'll need to introduce vertical sharding at a certain scale of the controllers. That has not yet been implemented though (nor are there plans to do so that I'm aware of)
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551091511085600?thread_ts=1551090439.083100&cid=C3TPCAQG1
----
2019-02-25 10:57:44 UTC - Andrei Alecu: found the problem, my persistentvolumes are all on /tmp so the couchdb gets cleared on reboot - need to figure out a solution for this
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551092264086800
----
2019-02-25 10:57:45 UTC - Andrei Alecu: <https://github.com/apache/incubator-openwhisk-deploy-kube/issues/339>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551092265087000
----
2019-02-25 12:39:07 UTC - Erez Hadad: Hi folks. Regarding my "overhead" tool: <https://github.com/IBM/overhead>       - Based on the bi-weekly feedback, I was requested to change the name. I thought of "owperf". Makes sense?
+1 : James Thomas
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551098347088800
----
2019-02-25 12:55:05 UTC - Rodric Rabbah: @Jonathan Springer any errors or failures in the app logs to upload travis logs to box?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551099305089500
----
2019-02-25 12:55:11 UTC - Rodric Rabbah: ill clear out some room now
+1 : Sven Lange-Last
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551099311089700
----
2019-02-25 12:55:14 UTC - Rodric Rabbah: fyi @Sven Lange-Last
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551099314089900
----
2019-02-25 12:58:39 UTC - Bui Vo Quoc Bao: Hello, I tried cloudant event provider <https://github.com/apache/incubator-openwhisk-package-cloudant>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551099519090400
----
2019-02-25 12:59:43 UTC - Bui Vo Quoc Bao: Everything works fine. However, I cannot create a trigger
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551099583091300
----
2019-02-25 12:59:51 UTC - Bui Vo Quoc Bao: ```wsk trigger create myCloudantTrigger --feed /_/myCloudant/changes --param dbname testdb```
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551099591091500
----
2019-02-25 13:00:10 UTC - Bui Vo Quoc Bao: This command always returns an error
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551099610092000
----
2019-02-25 13:01:03 UTC - Mark Deuser: @Andrei Alecu - i may have misunderstood.  from the following responses i thought you were manually using curl to perform the equivalent cli `wsk api create` command
&gt; Manually running `wsk -i api create /hello GET /guest/myapi-staging-hello`
unfortunately, i'm not familiar with the serverless framework offering, but it seems the rest api is not being called with all of the necessary parameters
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551099663092200?thread_ts=1551048160.079900&cid=C3TPCAQG1
----
2019-02-25 13:01:04 UTC - Bui Vo Quoc Bao: “The requested resource does not exist.”
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551099664092400
----
2019-02-25 13:01:14 UTC - Bui Vo Quoc Bao: I can read/write json to testdb
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551099674092700
----
2019-02-25 13:01:37 UTC - Bui Vo Quoc Bao: I’m not sure what is wrong here
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551099697093000
----
2019-02-25 13:02:31 UTC - Rodric Rabbah: you’re using couchdb? or cloudant?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551099751093300
----
2019-02-25 13:02:43 UTC - Rodric Rabbah: did you check the logs from the the action or the provider?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551099763093800
----
2019-02-25 13:02:48 UTC - Bui Vo Quoc Bao: cloudant from IBM, my openwhisk is local
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551099768094100
----
2019-02-25 13:03:38 UTC - Bui Vo Quoc Bao: here is the log
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551099818094300
----
2019-02-25 13:03:40 UTC - Bui Vo Quoc Bao: <https://github.com/apache/incubator-openwhisk-package-cloudant/issues/195>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551099820094500
----
2019-02-25 13:06:08 UTC - Bui Vo Quoc Bao: I can write json to testdb: ```bin/wsk action invoke -i /_/myCloudant/write --blocking --result --param dbname testdb --param doc "{\"_id\":\"doc1\",\"text\":\"Hello, how are you today?\"}"```
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551099968095000
----
2019-02-25 13:06:36 UTC - Bui Vo Quoc Bao: and I can read it: ```bin/wsk action invoke -i /_/myCloudant/read --blocking --result --param dbname testdb --param id doc1```
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551099996095400
----
2019-02-25 13:18:05 UTC - Dominic Kim: Thanks
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551100685095600?thread_ts=1551090439.083100&cid=C3TPCAQG1
----
2019-02-25 13:21:12 UTC - Rodric Rabbah: you installed your own `myCloudant/changes` package or is this a binding of the whisk.system package?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551100872098000
----
2019-02-25 13:22:05 UTC - Rodric Rabbah: check the logs for the `changes` activation that may tell you something
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551100925099100?thread_ts=1551100925.099100&cid=C3TPCAQG1
----
2019-02-25 13:22:19 UTC - Rodric Rabbah: @Jason Peterson or someone from ibm might need to jump in
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551100939099500
----
2019-02-25 13:22:24 UTC - Andrei Alecu: Still having problems with openwhisk pods not surviving a docker restart, my PersistentVolumes are now persistent (non /tmp based) and kafka doesn't start due to some file corruption issues. It could be Docker Desktop for Windows related somehow. I've attached full logs at <https://github.com/apache/incubator-openwhisk-deploy-kube/issues/435>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551100944099600
----
2019-02-25 13:23:40 UTC - Bui Vo Quoc Bao: yes, with the following command: ```bin/wsk package bind --insecure /guest/cloudant myCloudant -p username xxx-bluemix -p password xxx -p host xxx-bluemix.cloudantnosqldb.appdomain.cloud```
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551101020100100
----
2019-02-25 13:24:35 UTC - Bui Vo Quoc Bao: as suggested in the documentation
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551101075100400
----
2019-02-25 13:28:17 UTC - Bui Vo Quoc Bao: Sorry, but how can I do that? A flag in bin/wsk command?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551101297100500?thread_ts=1551100925.099100&cid=C3TPCAQG1
----
2019-02-25 13:32:22 UTC - Rodric Rabbah: fwiw you should not use --insecure with ibm cloud functions
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551101542101200
----
2019-02-25 13:32:59 UTC - Rodric Rabbah: check the activation log for the changes action (`wsk activation list changes` i think)
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551101579101900
----
2019-02-25 13:33:43 UTC - Rodric Rabbah: @Mark Deuser does plain `wsk` cli work with bluemix? ^^
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551101623102300?thread_ts=1551101623.102300&cid=C3TPCAQG1
----
2019-02-25 13:37:16 UTC - Bui Vo Quoc Bao: The command ```bin/wsk activation list changes -i``` shows nothing :white_frowning_face:
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551101836103800
----
2019-02-25 13:38:46 UTC - Bui Vo Quoc Bao: I used the wsk command in verbose mode, this REST call failed
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551101926105400
----
2019-02-25 13:38:50 UTC - Bui Vo Quoc Bao: ```[POST]	<https://172.17.0.1/api/v1/namespaces/guest/actions/myCloudant/changes?blocking=true&amp;result=false>```
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551101930105600
----
2019-02-25 13:39:11 UTC - Mark Deuser: it can - working against cloud-foundry based namespaces..  api management will require the user uaa token so that would be a manual step to retrieve and configure that.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551101951105700?thread_ts=1551101623.102300&cid=C3TPCAQG1
----
2019-02-25 13:43:11 UTC - Bui Vo Quoc Bao: Req Body:
```{"authKey":"xxx","dbname":"testdb","lifecycleEvent":"CREATE","triggerName":"/_/myCloudantTrigger"}```
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551102191106200
----
2019-02-25 13:43:39 UTC - Bui Vo Quoc Bao: Response body received: ```{"activationId":"0c519debf8014f00919debf8013f00e2","annotations":[{"key":"path","value":"guest/cloudant/changes"},{"key":"waitTime","value":16},{"key":"kind","value":"nodejs:6"},{"key":"limits","value":{"concurrency":1,"logs":10,"memory":256,"timeout":90000}},{"key":"initTime","value":147}],"duration":190,"end":1551101413360,"logs":[],"name":"changes","namespace":"guest","publish":false,"response":{"result":{"error":{"code":"71723e8dd6ae83a16845a368e50b57b6","error":"The requested resource does not exist."}},"status":"application error","success":false},"start":1551101413170,"subject":"guest","version":"0.0.2"}```
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551102219106400
----
2019-02-25 13:53:33 UTC - Rodric Rabbah: @Bui Vo Quoc Bao I’m confused. Are you running your own openwhisk or using ibm cloud functions?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551102813107800
----
2019-02-25 13:54:31 UTC - Rodric Rabbah: the apihost above is local - I’m not sure I understand your setup 
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551102871109000
----
2019-02-25 13:55:41 UTC - Bui Vo Quoc Bao: I’m running my own openwhisk, but I use cloudant from IBM
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551102941109600
----
2019-02-25 13:56:37 UTC - Bui Vo Quoc Bao: So the setup is not correct?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551102997109900
----
2019-02-25 14:01:21 UTC - Rodric Rabbah: and does your openwhisk have an external ip?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551103281110800
----
2019-02-25 14:02:54 UTC - Bui Vo Quoc Bao: No.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551103374111000
----
2019-02-25 14:04:06 UTC - Bui Vo Quoc Bao: So Cloudant from IBM need to “see” my openwhisk, right?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551103446111600
----
2019-02-25 14:11:26 UTC - Rodric Rabbah: if you expect the cloudant feed to fire triggers against your openwhisk, then yes you need a public ip
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551103886112400
----
2019-02-25 14:12:26 UTC - Rodric Rabbah: i suspect now from your error log that you don’t have the cloudant feed running locally and that’s why the “resource does not exist”, but even if you point to ibm’s it wont work for your setup
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551103946113900
----
2019-02-25 14:13:33 UTC - Bui Vo Quoc Bao: Okay, the documentation doesn’t say anything about this and I thought my openwhisk does “polling” for any changes
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551104013114600
----
2019-02-25 14:14:06 UTC - Bui Vo Quoc Bao: I will find another way…
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551104046115000
----
2019-02-25 14:14:12 UTC - Bui Vo Quoc Bao: Thanks you
+1 : Rodric Rabbah
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551104052115300
----
2019-02-25 14:14:36 UTC - Rodric Rabbah: the polling happens in the cloudant feed provider - this is a separate microservice which listens to the database changes and fires openwhisk triggers.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551104076115900
----
2019-02-25 15:06:03 UTC - Andrei Alecu: Onwards to another issue :slightly_smiling_face:

Docker Desktop on Windows:

```sh
$ wsk -i action invoke hello --result
{
    "payload": "Hello, World!"
}

$ wsk -i api create /hello get hello --response-type json
ok: created API /hello GET for action /_/hello
<https://192.168.65.3:31001/api/23bc46b1-71f6-4ed5-8c54-816aa4f8c502/hello>

$ curl -k <https://localhost:31001/api/23bc46b1-71f6-4ed5-8c54-816aa4f8c502/hello>
{"code":502, "message":"Oops. Something went wrong. Check your URI and try again."}
```
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551107163118300?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 15:06:49 UTC - Andrei Alecu: I cannot seem to get the api gateway to execute actions
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551107209118800
----
2019-02-25 15:08:35 UTC - Andrei Alecu: the api gateway log seems to have this in it:

```
==&gt; /var/log/api-gateway/gateway_error.log &lt;==
2019/02/25 15:05:18 [error] 38#0: *28 connect() failed (111: Connection refused) while connecting to upstream, client: 10.1.0.96, server: , request: "GET /api/23bc46b1-71f6-4ed5-8c54-816aa4f8c502/hello HTTP/1.1", upstream: "<https://127.0.0.1:31001/api/v1/web/guest/default/hello.json>", host: "owdev-apigateway.openwhisk.svc.cluster.local:8080"
2019/02/25 15:05:18 [warn] 38#0: *28 upstream server temporarily disabled while connecting to upstream, client: 10.1.0.96, server: , request: "GET /api/23bc46b1-71f6-4ed5-8c54-816aa4f8c502/hello HTTP/1.1", upstream: "<https://127.0.0.1:31001/api/v1/web/guest/default/hello.json>", host: "owdev-apigateway.openwhisk.svc.cluster.local:8080"
2019/02/25 15:05:18 [crit] 38#0: *28 connect() to [::1]:31001 failed (99: Address not available) while connecting to upstream, client: 10.1.0.96, server: , request: "GET /api/23bc46b1-71f6-4ed5-8c54-816aa4f8c502/hello HTTP/1.1", upstream: "https://[::1]:31001/api/v1/web/guest/default/hello.json", host: "owdev-apigateway.openwhisk.svc.cluster.local:8080"
2019/02/25 15:05:18 [warn] 38#0: *28 upstream server temporarily disabled while connecting to upstream, client: 10.1.0.96, server: , request: "GET /api/23bc46b1-71f6-4ed5-8c54-816aa4f8c502/hello HTTP/1.1", upstream: "https://[::1]:31001/api/v1/web/guest/default/hello.json", host: "owdev-apigateway.openwhisk.svc.cluster.local:8080"
```
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551107315119200
----
2019-02-25 15:09:17 UTC - Andrei Alecu: running on kubernetes, this is from the api gateway pod log
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551107357119700
----
2019-02-25 15:10:04 UTC - Andrei Alecu: It's a fresh installation, and a simple hello world. I'm starting to think there are not a lot of people trying to use openwhisk on windows dev machines :slightly_smiling_face:
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551107404120500
----
2019-02-25 15:12:32 UTC - Andrei Alecu: I'm assuming this is because of the self signed certificate, because I can open <https://127.0.0.1:31001/api/v1/web/guest/default/hello.json> if I bypass the cert:

```
$ curl -k <https://127.0.0.1:31001/api/v1/web/guest/default/hello.json>
{
  "payload": "Hello, World!"
}
```

How do I get the api gateway to connect to <https://localhost> without caring about the certificate?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551107552121500
----
2019-02-25 15:13:55 UTC - Andrei Alecu: Or, the gateway should probably try to connect to <https://192.168.65.3:31001> as an upstream server, not localhost, from within the pod, as it shows in the log. What am I doing wrong?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551107635122400
----
2019-02-25 15:14:52 UTC - Andrei Alecu: I believe Docker Desktop for Windows should be exactly the same as Docker Desktop for Mac functionality wise
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551107692123100
----
2019-02-25 15:17:34 UTC - Rodric Rabbah: is localhost the same as 192.168.65.3?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551107854123200?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 15:22:03 UTC - Andrei Alecu: well, no. I cannot see 192.168.65.3 from my machine. It's a docker/kubernetes internal ip, on docker windows and mac everything is on localhost
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551108123123400?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 15:22:29 UTC - Andrei Alecu: I'm using this guide: <https://github.com/apache/incubator-openwhisk-deploy-kube#configuring-the-cli-for-kubernetes-on-docker-for-mac-and-windows>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551108149123600?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 15:35:09 UTC - Ali Tariq: hi, apologies for the late response - but i have not been able to "make quick-start" on the latest version. Its terminates with some error (pic attached). Since temporary errors keep occurring in the main branch, i had saved a stable state for openwhisk in the past, and used it every time something like this occurs. I ran openwhisk instance (docker-compose), but then i was able to create functions but any command with linking API to the function resulted in the error i talked about previously.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551108909123800?thread_ts=1550961506.072100&cid=C3TPCAQG1
----
2019-02-25 15:35:50 UTC - Ali Tariq: 
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551108950124000?thread_ts=1550961506.072100&cid=C3TPCAQG1
----
2019-02-25 15:39:41 UTC - Andrei Alecu: 192.168.65.3 comes from `kubectl describe nodes | grep InternalIP`
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551109181124400?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 15:47:14 UTC - Dave Grove: the apigateway is running inside the kube cluster, so it should work to connect to 192.168.65.3:31001.  You don’t want localhost (that will try to connect to the container that apigateway is running in, not to the windows host)
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551109634124600?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 16:14:43 UTC - Andrei Alecu: @Dave Grove yes I know, but the api gateway log seems to show a 127.0.0.1 upstream, so it's not connecting to the proper ip. Or at least that's what I take from it
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551111283124800?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 16:17:47 UTC - Andrei Alecu: my assumption could be wrong, but why would `curl -k <https://localhost:31001/api/23bc46b1-71f6-4ed5-8c54-816aa4f8c502/hello>` fail and `curl -k <https://localhost:31001/api/v1/web/guest/default/hello.json>` work?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551111467125000?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 16:22:23 UTC - Bui Vo Quoc Bao: Okay, I have tried github events. Does github need to see my openwhisk?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551111743125900
----
2019-02-25 16:23:13 UTC - Rodric Rabbah: any external service that’s going to fire a trigger against your openwhisk will need a public ip
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551111793126900
----
2019-02-25 16:23:23 UTC - Rodric Rabbah: you can tunnel to it from your router perhaps?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551111803127200
----
2019-02-25 16:27:06 UTC - Bui Vo Quoc Bao: Are there any event sources where my openwhisk does the polling for changes?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551112026127900
----
2019-02-25 16:29:21 UTC - Rodric Rabbah: what are you trying to accomplish?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551112161128400
----
2019-02-25 16:29:30 UTC - Rodric Rabbah: you can deploy the alarm service locally
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551112170128700
----
2019-02-25 16:30:51 UTC - Bui Vo Quoc Bao: I just want to try the execution path: external event-&gt;action
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551112251129300
----
2019-02-25 16:32:15 UTC - Shazron Abdullah: poll for changes: use alarm (see example <https://github.com/apache/incubator-openwhisk/blob/master/docs/packages.md> **Creating and using trigger feeds**)
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551112335129900
----
2019-02-25 16:32:59 UTC - Rodric Rabbah: You can fire a trigger manually 
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551112379130200
----
2019-02-25 16:33:36 UTC - Rodric Rabbah: But anything from the public internet can’t reach your openwhisk unless you make it visible to the internet 
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551112416131300
----
2019-02-25 16:41:28 UTC - Bui Vo Quoc Bao: got it. thanks again… :slightly_smiling_face:
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551112888131600
----
2019-02-25 18:39:25 UTC - Andrei Alecu: still stuck on this, any pointers would be really appreciated
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551119965132000?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 19:09:10 UTC - Dave Grove: remind me, what exactly is your mycluster.yaml
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551121750132200?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 19:16:10 UTC - Andrei Alecu: ```
whisk:
  ingress:
    type: NodePort
    apiHostName: 192.168.65.3
    apiHostPort: 31001

nginx:
  httpsNodePort: 31001
    
invoker:
  containerFactory:
    impl: "docker"
```
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551122170132400?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 19:17:01 UTC - Dave Grove: that looks quite reasonable
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551122221132600?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 19:17:24 UTC - Andrei Alecu: I attached some additional logs in the <#C3TPCAQG1|general> channel from the api gateway pod
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551122244132800?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 19:17:48 UTC - Andrei Alecu: from what I understand, it seems like nginx tries to forward to localhost from within that pod, which seems wrong
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551122268133000?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 19:23:50 UTC - Dave Grove: I can reproduce at least a similar problem on Docker for Mac.   The output from wget suggests a certificate problem
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551122630133200?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 19:23:53 UTC - Andrei Alecu: I just followed the basic guide. I even reset docker to factory settings and started fresh with everything so other than using Windows I'm not sure what else would affect this.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551122633133400?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 19:24:04 UTC - Dave Grove: daves-mbp:openwhisk dgrove$ wget --no-check-certificate <https://localhost:31001/api/23bc46b1-71f6-4ed5-8c54-816aa4f8c502/hello>
--2019-02-25 14:23:10--  <https://localhost:31001/api/23bc46b1-71f6-4ed5-8c54-816aa4f8c502/hello>
Resolving localhost (localhost)... ::1, 127.0.0.1
Connecting to localhost (localhost)|::1|:31001... connected.
WARNING: cannot verify localhost’s certificate, issued by ‘CN=*.localhost,O=OpenWhisk,L=Yorktown,ST=NY,C=US’:
  Self-signed certificate encountered.
    WARNING: certificate common name ‘*.localhost’ doesn’t match requested host name ‘localhost’.
HTTP request sent, awaiting response... 502 Bad Gateway
2019-02-25 14:23:10 ERROR 502: Bad Gateway.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551122644133600?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 19:31:06 UTC - Dave Grove: this has worked in the past.  not sure what has happened
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551123066133800?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 19:32:07 UTC - Andrei Alecu: it seems like a basic usage scenario, so it should affect everyone, seems weird that I couldn't find any reports unless it's a very recent problem
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551123127134000?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 19:46:31 UTC - Dave Grove: The regression tests invoke apigateway from within the cluster (`helm test owdev`).  Those still work for me.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551123991134200?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 19:47:18 UTC - Dave Grove: The problem seems specific to docker for mac/windows odd networking setup and invoking from external to the cluster.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551124038134400?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 19:47:45 UTC - Andrei Alecu: Tests do work for me as well on Windows:
```
$ helm test owdev --cleanup
RUNNING: owdev-tests-package-checker
PASSED: owdev-tests-package-checker
RUNNING: owdev-tests-smoketest
PASSED: owdev-tests-smoketest
```
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551124065134600?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 20:04:18 UTC - Sam Hjelmfelt: How often are the docker hub images updated? It would be great to get an image with commit #6e883f9
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551125058135800
----
2019-02-25 20:04:30 UTC - Sam Hjelmfelt: from Saturday
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551125070136000
----
2019-02-25 20:07:08 UTC - Andrei Alecu: <https://localhost:31001/api/v1/web/guest/default/hello.json> works in my case though, I got that url from the gateway logs
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551125228136100?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 20:08:16 UTC - Andrei Alecu: is that sort of endpoint documented anywhere? could it be used as an alternative?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551125296136300?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 20:09:31 UTC - Dave Grove: right,.  that;s invoke the action “directly”  nginx ==&gt; controller ==&gt; invoker.    The encoded URL goes nginx ==&gt; apigateway ==&gt; nginx ==&gt; controller and that seems to be unhappy about the self-signed  nginx certificagte
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551125371136500?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 20:09:52 UTC - Andrei Alecu: would it matter that the certificate is for *.localhost instead of just localhost?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551125392136700?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 20:10:40 UTC - Dave Grove: i suspect it might, but i don’t really understand all the convolutions of ssl
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551125440136900?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 20:12:25 UTC - Dave Grove: if the machinery is working, it should be there already.  supposed to auto-publish images tagged with githash multiple times a day via a Jenkins job
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551125545137800
----
2019-02-25 20:13:43 UTC - Rodric Rabbah: is the build on master passing?
yes it is
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551125623138200
----
2019-02-25 20:14:40 UTC - Rodric Rabbah: hmm
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551125680138600
----
2019-02-25 20:14:50 UTC - Rodric Rabbah: @Carlos Santana?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551125690138800
----
2019-02-25 20:15:07 UTC - Rodric Rabbah: just realized i dont actually know where to check jenkins
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551125707139200
----
2019-02-25 20:24:49 UTC - Andrei Alecu: is this a problem with kubernetes or something else?
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551126289139300?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 20:25:23 UTC - Andrei Alecu: I wasn't able to use openwhisk with the openwhisk-toolbox docker thing on Windows. Only kubernetes seemed like it worked.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551126323139500?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 20:26:04 UTC - Andrei Alecu: I mean openwhisk-devtools
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551126364139700?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 20:26:11 UTC - Andrei Alecu: via docker compose
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551126371139900?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 20:28:21 UTC - Dave Grove: scala compilation is failing
thankyou : Rodric Rabbah
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551126501140300
----
2019-02-25 20:28:23 UTC - Dave Grove: <https://builds.apache.org/view/O/view/OpenWhisk/job/OpenWhisk-DockerHub/>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551126503140500
----
2019-02-25 20:28:36 UTC - Dave Grove: <https://builds.apache.org/view/O/view/OpenWhisk/job/OpenWhisk-DockerHub/1761/console>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551126516140700
----
2019-02-25 20:28:55 UTC - Rodric Rabbah: i used docker compose occasionally on mac though - and tested this weekend most recently
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551126535141200?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 20:29:32 UTC - Dave Grove: perhaps the jekins slave machine got “upgraded” to pick up Java 11 or something horrible like that
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551126572141900
----
2019-02-25 20:30:13 UTC - Dave Grove: 
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551126613142100
----
2019-02-25 20:30:31 UTC - Rodric Rabbah: oy that - needs a gralew clean if the jenkins space isn’t wiped
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551126631142700
----
2019-02-25 20:32:51 UTC - Rodric Rabbah: from the error message, the upstream server is `<https://127.0.0.1:31001>`
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551126771143300?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 20:32:54 UTC - Rodric Rabbah: that looks wrong
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551126774143500?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 20:35:11 UTC - Rodric Rabbah: i would edit the nginx conf to force the ip address to be correct for the upstream server first make confirm things work
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551126911143800?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 20:35:22 UTC - Rodric Rabbah: then need to figure out why that value is set to localhost in the gateway
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551126922144000?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:04:15 UTC - Andrei Alecu: I found this: <https://github.com/apache/incubator-openwhisk-apigateway/blob/bed6e65d9100397abd70d43b01f0436ef8fa9e7d/conf.d/managed_endpoints.conf#L83>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551128655144200?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:04:25 UTC - Andrei Alecu: seems to be where the upstream is being set, but I'm not sure how localhost ends up there
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551128665144400?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:05:08 UTC - Andrei Alecu: unless the lua code overwrites that variable somehow, it seems like it should be a hardcoded ip
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551128708144600?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:06:08 UTC - Rodric Rabbah: set $upstream <https://172.17.0.1>;
        
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551128768144900?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:06:23 UTC - Rodric Rabbah: That’s what youre seeing I guess 
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551128783145400?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:06:28 UTC - Andrei Alecu: kubectl exec owdev-apigateway-567db97f47-bh9t4 -n openwhisk cat /etc/api-gateway/conf.d/managed_endpoints.conf
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551128788145600?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:06:57 UTC - Andrei Alecu: I can see that on my pod the file is identical to the one linked above, so it wasn't changed by some initialization script
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551128817145800?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:07:22 UTC - Andrei Alecu: I'm not sure what that 172.17.0.1 ip is and why it gets replaced by localhost
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551128842146000?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:08:45 UTC - Rodric Rabbah: 172.17 would work I think but it’s getting replaced  
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551128925147000?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:09:05 UTC - Rodric Rabbah: Actually I shouldn’t say that I’m not as familiar with kube
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551128945147600?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:28:01 UTC - Andrei Alecu: I think it's important that this works though: <https://localhost:31001/api/v1/web/guest/default/hello.json>

How/where is that routed to? It's served by the same api gateway
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551130081147800?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:28:35 UTC - Andrei Alecu: and if that's the case, then the redirect from the other url to this one should be internal to the apigateway nginx instance, so it should work with localhost
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551130115148000?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:28:47 UTC - Andrei Alecu: so it might still point to a certificate error
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551130127148200?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:31:01 UTC - Rodric Rabbah: No the webaction is served from a different nginx. They can be the same in principle and using docker-compose they are the same nginx. 
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551130261149400?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:34:31 UTC - Rodric Rabbah: The api gw is optional in my view. You can in fact get subdomain mappings with the webactions (like guest.locahost/...) so URLs are nicer and more secure. I haven’t tried this with the openwhisk kube deploy but know in general it works 
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551130471152000?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:45:35 UTC - Dave Grove: i hadn’t played with invoking api gw routes externally when deployed to kube for Mac.  There’s something not configured correctly on the internal routing with this networking setup.  It does work when the apigw route is invoked internal to the cluster (checked by the smoke-test job) and works externally on other targets (for example IKS).
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551131135152400?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:48:45 UTC - Andrei Alecu: @Rodric Rabbah pointed me to <https://github.com/apache/incubator-openwhisk/blob/master/docs/webactions.md>, so this will get me unstuck for now I think. Api Gateway would still be nice to work because I'd like to eventually use the serverless-openwhisk plugin for the serverless framework, which uses api gateway to set up routes. But at least I can play with openwhisk for now.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551131325152700?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:49:42 UTC - Andrei Alecu: @Dave Grove should I open an issue on github for this or is it not necessary? I see that the kube repository is your baby :slightly_smiling_face:
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551131382152900?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:50:51 UTC - Dave Grove: an issue isn’t a bad idea.  since you are unblocked I am not going to dit into it more today.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551131451153100?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:55:06 UTC - Andrei Alecu: thank you very much for your time, I'm happy at least it wasn't pebkac on my end :smile: opened <https://github.com/apache/incubator-openwhisk-deploy-kube/issues/436>
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551131706153300?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 21:55:28 UTC - Dave Grove: thanks
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551131728153500?thread_ts=1551107163.118300&cid=C3TPCAQG1
----
2019-02-25 22:15:11 UTC - Ali Tariq: Okay - i have been using automated scripts for setup every time on new linux VM - which didn't include npm (maybe its a rather recent edition) which was causing the runtime issue and new .wskprops file compared to previous /bin/wsk. Its, working now :smile: - Thanks for your reply, it was helpful.
https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1551132911153700?thread_ts=1550961506.072100&cid=C3TPCAQG1
----