2019-10-29 04:00:00 UTC - Abbie Dowd: It working now! :nerd_face: tada : Dominic Kim https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572321600140800?thread_ts=1572232613.126400&cid=C3TPCAQG1 ---- 2019-10-29 04:00:22 UTC - Abbie Dowd: It was the IP address. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572321622141100?thread_ts=1572232613.126400&cid=C3TPCAQG1 ---- 2019-10-29 04:00:26 UTC - Abbie Dowd: Thanks!! +1 : Dominic Kim https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572321626141300?thread_ts=1572232613.126400&cid=C3TPCAQG1 ---- 2019-10-29 13:10:04 UTC - Pepi Paraskevoulakou: hello guys i made a sequence of functions, the last of them using from docker hub the docker image openwhisk/python3aiactions, but when i call it from rest api i got this error. Anyone help please!! https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572354604142900 ---- 2019-10-29 13:32:38 UTC - Pepi Paraskevoulakou: actually the problem is that i made a zip file with a keras model (size 6,5 MB) and a docker conf including machine learning libs (keras, tensorflow) but the issue is that when i cal the rest api it stucks all the virtual machine, any experience? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572355958144900?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 13:37:01 UTC - Pepi Paraskevoulakou: it stucks like this https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572356221145100 ---- 2019-10-29 14:05:48 UTC - Pepi Paraskevoulakou: maybe openwhisk doesnt recognise .h5 files? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572357948145700 ---- 2019-10-29 14:07:52 UTC - Ali Tariq: how did you create the action? also, did you test your action using wsk-cli? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572358072145800?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 14:08:24 UTC - Pepi Paraskevoulakou: no, only from rest api https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572358104146000?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 14:08:35 UTC - Pepi Paraskevoulakou: do i need to do that first? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572358115146200?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 14:09:01 UTC - Pepi Paraskevoulakou: with .h5 it stucks all the virtual machine https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572358141146400?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 14:10:11 UTC - Ali Tariq: i would suggest testing it using wsk-cli, also please share how did you create the action - if i recall correctly, tensorflow is not present by default, and given its size, you are required to make a docker-hub image and link it to your function at creation time https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572358211146600?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 14:11:57 UTC - Ali Tariq: also, it cannot remain stuck for a long time - (there is max running time limit of 60 sec by default, after which your function will be killed) - so if it is indeed killed, we should be able to see the "limit exceeded" (or similar) - notification in the activations https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572358317146800?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 14:13:45 UTC - Pepi Paraskevoulakou: i created 3 functions, alla of them with a conf of a docker (the 2 first with a docker including ml libraries pandas,numpy), the third function is a zip with a pretrained model (model.h5) with configuration a docker which includes tensorflow, keras,pandas,numpy libs. When i create the sequence of these 3 functions and create a rest api to pass new features to my pretrained model (which is in third function) the vm stucks forever https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572358425147000?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 14:14:36 UTC - Pepi Paraskevoulakou: first_Function.py along with docker conf https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572358476147200?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 14:15:45 UTC - Pepi Paraskevoulakou: Second function as a zip with a .sav file and a docker conf including also ml libraries https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572358545147600?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 14:17:20 UTC - Pepi Paraskevoulakou: third function as a zip file along with the .h5 file and docker including ml libs (tensorflow,keras etc) https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572358640148000?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 14:18:09 UTC - Pepi Paraskevoulakou: this is all but i think the vm stucks when openwhisk tries to read .h5 file (6,5 MB) https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572358689148400?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 14:22:29 UTC - Pepi Paraskevoulakou: do i need h5py lib maybe? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572358949148600?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 14:38:52 UTC - Pepi Paraskevoulakou: I cannot believe that OpenWhisk can’t read h5 files https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572359932149400?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 14:44:55 UTC - Ali Tariq: its not about openwhisk, the question is, whether python runtime can read it (without any external library) - if it uses an external library, you might wanna include that (because openwhisk has limit packages by default) - https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572360295149600?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 14:47:06 UTC - Ali Tariq: but in any-case, it's unlikely to be stuck at any point - either its performing a time-heavy execution or waiting for something - and the best way to debug would be "1) run the program on local python (i assume thats the runtime you are using) runtime, - 2) then test it on wsk-cli " https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572360426149800?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 14:48:13 UTC - Ali Tariq: the reason i say this, if it's unable to read .h5 files, you will likely see an error or exception in the logs https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572360493150000?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 14:53:24 UTC - Pepi Paraskevoulakou: but it stucks the vm i can control it https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572360804150200?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 14:53:38 UTC - Pepi Paraskevoulakou: so i cannot read logs https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572360818150400?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 14:55:03 UTC - Pepi Paraskevoulakou: and i only need this to finish my thessis https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572360903150600?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 15:17:41 UTC - Pepi Paraskevoulakou: it cannot read h5 file it stucks all the vm i cannot do nothing https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572362261150800?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 15:18:30 UTC - Pepi Paraskevoulakou: and wsk cli stucks too https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572362310151000?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 15:18:59 UTC - Pepi Paraskevoulakou: https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572362339151200?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 15:22:58 UTC - Ali Tariq: what about, when you run the code on your local machine (no openwhisk) - how long does it take? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572362578151600?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 15:25:05 UTC - Pepi Paraskevoulakou: 4 seconds https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572362705151800?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 15:25:42 UTC - Pepi Paraskevoulakou: this is the function in my local machine https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572362742152000?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 15:30:13 UTC - Ali Tariq: can you try creating a new action with nothing but a logging statement in the main thread (comment out everything else in your main_loop) - keep the imports as it is and try running - look if you are able to see you logs ... then uncomment instruction by instruction to see where the fault originates. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572363013152400?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 15:32:41 UTC - Pepi Paraskevoulakou: maybe the problem is thatthe docker i use has tensorflow version 1.11 and my local machine has 2.0.0 so i stored my h5 model with new version of tensorflow which will cause problems in the action afterwards https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572363161152600?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 15:35:10 UTC - Ali Tariq: could be ... i have had many compatibility issues working on tensorflow versions, just update you docker image with the latest tf version https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572363310152800?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 15:35:41 UTC - Pepi Paraskevoulakou: you mean to create a new docker with tf new version right? +1 : Ali Tariq https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572363341153000?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 15:40:52 UTC - Ali Tariq: you can just rebuild the image - by modifying the Dockfile u used (just change the tensorflow installation to latest) https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572363652153200?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 15:43:46 UTC - Pepi Paraskevoulakou: i use this one <https://hub.docker.com/r/openwhisk/python3aiaction> how can i do this? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572363826153500?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 16:00:50 UTC - Pepi Paraskevoulakou: please.. https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572364850153700?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 16:03:23 UTC - Ali Tariq: oh .. so you didnt build the docker image (you have been using so far) yourself? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572365003153900?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 16:03:55 UTC - Ali Tariq: <http://jamesthom.as/blog/2017/08/04/large-applications-on-openwhisk/> https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572365035154100?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 16:04:01 UTC - Pepi Paraskevoulakou: i use for the third function the docker from the url i send to you https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572365041154400?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 16:04:13 UTC - Pepi Paraskevoulakou: yes i saw this article https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572365053154600?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 16:05:04 UTC - Ali Tariq: use the artical to built your own docker image - (you would need to add `pip install tensorflow` in the Dockerfile as well) https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572365104154800?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 16:06:30 UTC - Ali Tariq: push it to your docker hub - and you can use it ... the issue with tensorflow, it keeps changing rapidly ... its very likely the api you are using - might not be present in older version https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572365190155000?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 16:08:23 UTC - Pepi Paraskevoulakou: do you think that it will play with this upgrade? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572365303155400?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 16:08:52 UTC - Ali Tariq: before you go through that trouble - you can check if that is indeed the issue - you can downgrade your local tensorflow (remove tf from your local machine, and install the version you have on your docker image) https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572365332155600?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 16:09:17 UTC - Ali Tariq: and re-run your code https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572365357155800?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 16:09:32 UTC - Pepi Paraskevoulakou: yes you are right https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572365372156000?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 19:45:59 UTC - Pepi Paraskevoulakou: its a big trouble, tensorflow 1.11 doesnt support python 3.7 so i have to install older python version, omg its a true headache https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572378359156200?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 20:38:00 UTC - Ali Tariq: it would be alot easier to make that docker-image with latest tf, because, the api you are using in your code - might not even be supported in such an older version of tensorflow https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572381480156400?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 21:54:34 UTC - Klaus Pedersen: Hi Adrian - I have the same problem. I’m using swift also in a docker container, but getting “The action did not initialize or run as expected. Log data might be missing.” Did you found the problem? https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572386074156600?thread_ts=1571750437.034300&cid=C3TPCAQG1 ---- 2019-10-29 22:00:16 UTC - Pepi Paraskevoulakou: Yes but I don’t know what to write in dockerfile because I think that it needs special handling for OpenWhisk https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572386416158100?thread_ts=1572355958.144900&cid=C3TPCAQG1 ---- 2019-10-29 22:00:52 UTC - Pepi Paraskevoulakou: I am from the morning till now midnight trying to debugging this thing :sweat::sweat::sweat: https://openwhisk-team.slack.com/archives/C3TPCAQG1/p1572386452159300?thread_ts=1572355958.144900&cid=C3TPCAQG1 ----
