2019-10-23 00:27:53 UTC - Rodric Rabbah: you can check the jenkins environment https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571790473037300 ---- 2019-10-23 00:27:59 UTC - Rodric Rabbah: basically you have to edit the hosts file https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571790479037500 ---- 2019-10-23 00:28:11 UTC - Rodric Rabbah: then you deploy the top level playbook +1 : zyqCSL https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571790491037800 ---- 2019-10-23 06:38:57 UTC - Adrian Schuepbach: Any idea what might cause this one-line error message?
I assume I have no error in my Python code, since it was working before. But assuming there would be one, like a syntax error or anything else, would I still see the "raised Exception" message in the OpenWhisk logs as in older OpenWhisk versions? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571812737038300?thread_ts=1571812737.038300&cid=C3TPCAQG1 ---- 2019-10-23 06:45:59 UTC - chetanm: Can you share the error message? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571813159038600 ---- 2019-10-23 06:48:37 UTC - Bill Zong: Not met before. The developer error would not print out why: ``` protected def logStream(transid: TransactionId, container: Container, logLimit: LogLimit, sentinelledLogs: Boolean, isDeveloperError: Boolean): Source[ByteString, Any] = { if (isDeveloperError) { logs.concat( Source.single( ByteString(LogLine(Instant.now.toString, "stderr", Messages.logWarningDeveloperError).toJson.compactPrint))) } } ``` Sometimes it’s an exception not caught. Add more logs to this action, or debug it locally. By the way, it would be better opening a thread for only one question. XD https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571813317038700?thread_ts=1571812737.038300&cid=C3TPCAQG1 ---- 2019-10-23 07:40:22 UTC - Adrian Schuepbach: sure, this is the only error message I get (see also yesterday's description above): `stderr: The action did not initialize or run as expected. Log data might be missing.` https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571816422039600 ---- 2019-10-23 07:44:43 UTC - Adrian Schuepbach: Thanks for the answer. Something must have changed when I upgraded OpenWhisk. The action code was working with an older OpenWhisk installation. Running the action code in a python2action container, which I start manually, works fine. This is why I wonder what is happening when I run it using `wsk -i action invoke`. Since it runs fine manually, but not iwith `wsk` and there is no additional output in the logs, it is really difficult to debug it. Adding more logs does not help, even the `print()` just below `def main(args)` doesn't show up in the logs. Even not, if I explicitely flush the stdout buffer. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571816683039700?thread_ts=1571812737.038300&cid=C3TPCAQG1 ---- 2019-10-23 07:51:02 UTC - Bill Zong: Currently, python3action is the default python runtime type. Would you plz show out the `wsk action get` info about the failed action? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571817062039900?thread_ts=1571812737.038300&cid=C3TPCAQG1 ---- 2019-10-23 08:07:28 UTC - Adrian Schuepbach: sure, this is the output: ``` $ wsk -i action get functionname ok: got action functionname { "namespace": "namespace", "name": "functionname", "version": "0.0.8", "exec": { "kind": "python:2", "binary": true }, "annotations": [ { "key": "provide-api-key", "value": false }, { "key": "exec", "value": "python:2" } ], "limits": { "timeout": 300000, "memory": 512, "logs": 10, "concurrency": 1 }, "publish": false } ``` https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571818048040100?thread_ts=1571812737.038300&cid=C3TPCAQG1 ---- 2019-10-23 08:08:48 UTC - Adrian Schuepbach: and this is its activation: ``` ok: got activation e7d1047e986a450091047e986a150040 { "namespace": "namespace2", "name": "functionname", "version": "0.0.8", "subject": "asq", "activationId": "e7d1047e986a450091047e986a150040", "start": 1571817902123, "end": 1571817925910, "duration": 23787, "response": { "status": "action developer error", "statusCode": 0, "success": false, "result": { "error": "The action did not return a dictionary." } }, "logs": [ "2019-10-23T08:05:25.912Z stderr: The action did not initialize or run as expected. Log data might be missing." ], "annotations": [ { "key": "path", "value": "namespace2/functionname" }, { "key": "waitTime", "value": 44860 }, { "key": "kind", "value": "python:2" }, { "key": "timeout", "value": false }, { "key": "limits", "value": { "concurrency": 1, "logs": 10, "memory": 512, "timeout": 300000 } }, { "key": "initTime", "value": 830 } ], "publish": false } ``` https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571818128040300?thread_ts=1571812737.038300&cid=C3TPCAQG1 ---- 2019-10-23 08:23:52 UTC - Bill Zong: ``` "result": { "error": "The action did not return a dictionary." } ``` Any idea about your action response? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571819032040900?thread_ts=1571812737.038300&cid=C3TPCAQG1 ---- 2019-10-23 08:28:59 UTC - Adrian Schuepbach: I thought about it, but I believe that since the action did not even initialize, this is just a side error. Even a toy example does not realiably work, like this example from the OpenWhisk website: ``` def main(args): name = args.get("name", "stranger") greeting = "Hello " + name + "!" print(greeting) return {"greeting": greeting} ``` It may work or may show the one-line message `stderr: The action did not initialize or run as expected. Log data might be missing.` and in this case also that the action did not return a dictionary (which is true, when it didn't initialize). https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571819339041100?thread_ts=1571812737.038300&cid=C3TPCAQG1 ---- 2019-10-23 08:50:07 UTC - Adrian Schuepbach: I just tried with Python 3, same behaviour https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571820607041400?thread_ts=1571812737.038300&cid=C3TPCAQG1 ---- 2019-10-23 08:53:05 UTC - Bill Zong: You said the action is OK in previous version, right? May be you should try the tagged runtime version instead of `nightly` build one? That will figure out whether it’s a executor problem. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571820785041700?thread_ts=1571812737.038300&cid=C3TPCAQG1 ---- 2019-10-23 08:56:48 UTC - Adrian Schuepbach: so I used the `incubator-openwhisk-deploy-kube` repository with helm to deply OpenWhisk on Kubernetes. It seems that while deploying, some scripts just clone the OpenWhisk repo (looks like they clone master). Also it seems that they pull images from docker hub (latest). How can I deploy a specific version of OpenWhisk on Kubernetes? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571821008041900?thread_ts=1571812737.038300&cid=C3TPCAQG1 ---- 2019-10-23 09:02:02 UTC - Bill Zong: I mean just replace the runtime config in helm value file ``` whisk: runtimes: "runtimes-fdn-debug.json" ``` Like this, using the fixed tagged python runtimes, then redeploy the openwhisk. To find out whether it’s a runtime problem. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571821322042100?thread_ts=1571812737.038300&cid=C3TPCAQG1 ---- 2019-10-23 09:02:57 UTC - Adrian Schuepbach: ok I will try it, thanks https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571821377042300?thread_ts=1571812737.038300&cid=C3TPCAQG1 ---- 2019-10-23 09:03:16 UTC - Bill Zong: By the way, the git clone repo could be set to specific commit by setting those: ``` whisk: versions: openwhisk: gitTag: ``` https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571821396042500?thread_ts=1571812737.038300&cid=C3TPCAQG1 ---- 2019-10-23 09:04:06 UTC - Adrian Schuepbach: aaah perfect, thanks https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571821446042700?thread_ts=1571812737.038300&cid=C3TPCAQG1 ---- 2019-10-23 13:00:03 UTC - Pepi Paraskevoulakou: hello again i am trying to create a sequence of functions but raises this error “error: Unable to create action ‘functions’: Sequence component does not exist. (code HTA27TtVgVLCe03NlyDZDdkOSV2M0DYu)” what i have to do? Run ‘wsk --help’ for usage. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571835603043000?thread_ts=1562247212.205400&cid=C3TPCAQG1 ---- 2019-10-23 13:13:29 UTC - Pepi Paraskevoulakou: need to install wskdeploy? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571836409043200?thread_ts=1562247212.205400&cid=C3TPCAQG1 ---- 2019-10-23 14:17:50 UTC - Adrian Schuepbach: I installed the older version I used before, but now I see in the activation log that it failed to run the container with a specific image: ``` $ wsk -i activation get b575cf16bd444e81b5cf16bd446e8147 ok: got activation b575cf16bd444e81b5cf16bd446e8147 { "namespace": "zrlio", "name": "asq_lambda_ow", "version": "0.0.1", "subject": "<mailto:[email protected]|[email protected]>", "activationId": "b575cf16bd444e81b5cf16bd446e8147", "start": 1571840117541, "end": 1571840117541, "duration": 0, "response": { "status": "whisk internal error", "statusCode": 0, "success": false, "result": { "error": "Failed to run container with image 'openwhisk/python2action:d6a8920'." } }, "logs": [], "annotations": [ { "key": "limits", "value": { "concurrency": 1, "logs": 10, "memory": 512, "timeout": 60000 } }, { "key": "path", "value": "zrlio/asq_lambda_ow" }, { "key": "kind", "value": "python:2" }, { "key": "waitTime", "value": 139 } ], "publish": false } ``` I tried I pull this specific version from docker hub manually on every node, but it did not help. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571840270043400?thread_ts=1571812737.038300&cid=C3TPCAQG1 ---- 2019-10-23 14:19:41 UTC - Adrian Schuepbach: and now OpenWhisk starts more and moe of these containers: `wskowdev-invoker-11-10-whisksystem-invokerhealthtestaction0` even though I only invoked my failing action once https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571840381043600?thread_ts=1571812737.038300&cid=C3TPCAQG1 ---- 2019-10-23 21:04:05 UTC - Adrian Schuepbach: ok, I did not touch the system for 7 hours and now I have 810 `wskowdev-invoker-11-95-whisksystem-invokerhealthtestaction0` instances https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571864645044300?thread_ts=1571812737.038300&cid=C3TPCAQG1 ---- 2019-10-23 21:28:51 UTC - mmarth: @Pepi Paraskevoulakou I dont know and I am afraid that posting on this old Slack thread will not give you any other eyeballs. Maybe start a new thread? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571866131044500?thread_ts=1562247212.205400&cid=C3TPCAQG1 ---- 2019-10-23 22:00:39 UTC - Adrian Schuepbach: mh, this is exactly the same behaviour that suddenly started to happen by itself last Wednesday on the old installation, which I had running for 225 days without interruptions and without updating anything. I wonder why after 225 days it sunndenly started to create more and more `wskowdev-invoker-11-95-whisksystem-invokerhealthtestaction0` instances. Anyways, it seems that the old version does not work anymore. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571868039044700?thread_ts=1571812737.038300&cid=C3TPCAQG1 ---- 2019-10-23 22:25:47 UTC - Adrian Schuepbach: now I only went slightely back to a version from September 05th. The error message is now `"error": "Failed to run container with image 'openwhisk/python2action:1.13.0-incubating'."` So it seems that when going back to older versions, there is a specific version coded to be used. After the first failed invocation, I pulled the image manually, since I was not sure if OpenWhisk's pull mechanism would pull these versions or "just" "latest", but it did not help, still the same error message. The good thing, no `wskowdev-invoker-11-95-whisksystem-invokerhealthtestaction0` instances so far https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1571869547044900?thread_ts=1571812737.038300&cid=C3TPCAQG1 ----
