Github user chanlee514 commented on the issue:
https://github.com/apache/incubator-predictionio/pull/328
Right now, the commands throw java.io.FileNotFoundException regarding
engine.json. May be better if we throw an error message saying the commands
should be executed inside engine d
Github user pferrel commented on the issue:
https://github.com/apache/incubator-predictionio/pull/328
Ideally `pio build` and `pio train` should run anywhere by supplying any
params needed.
@chanlee514 metadata still is needed in a service like Elasticsearch so
every place th
Github user chanlee514 commented on the issue:
https://github.com/apache/incubator-predictionio/pull/328
@pferrel: I agree, and that would be possible with a general metadata
registry server which would be the next step. The reason I started by removing
manifest.json is because the pr
Github user dszeto commented on the issue:
https://github.com/apache/incubator-predictionio/pull/328
@pferrel For `pio build` we probably can enhance it such that it can take a
working directory for compilation to run. If we look at some other build tools,
they also rely on having cer
Github user pferrel commented on the issue:
https://github.com/apache/incubator-predictionio/pull/328
ok, this looks like a good step
@dszeto we would not do `pio build` we would do `sbt build` in the
directory that has the template and this is fine IMO. The confusion users ha
Github user pferrel commented on the issue:
https://github.com/apache/incubator-predictionio/pull/328
One more question. After push how do we run the PredictionServer or train
on multiple machines? In the past this required copying the manifest.json and
making sure binaries are in the
Github user pferrel commented on the issue:
https://github.com/apache/incubator-predictionio/pull/328
@dszeto asked: "Do you currently have any production deployments that rely
heavily on engine IDs and versions? That would be a bigger immediate concern
regarding this change."
Github user dszeto commented on the issue:
https://github.com/apache/incubator-predictionio/pull/328
@chanlee514 @pferrel I have moved the discussion to dev@.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your projec
Github user pferrel commented on the issue:
https://github.com/apache/incubator-predictionio/pull/328
too much noise in @dev, I have a simple question.
If I train on one machine, how do I deploy on another. Currently I copy the
manifest--well actually the entire directory--fr
Github user pferrel commented on the issue:
https://github.com/apache/incubator-predictionio/pull/328
@chanlee514 said:
If the absolute filepath is the same in the machines (since SHA-1 hash of
filepath is used as engineVersion), there would be no changes in the above
Github user pferrel commented on the issue:
https://github.com/apache/incubator-predictionio/pull/328
The file path is required to be the same on any 2 machines and is
explicitly in the metadata. If you have the deploy machine dir in a different
path than the train machine dir (actual
Github user jimlyndon commented on the issue:
https://github.com/apache/incubator-predictionio/pull/328
+1
Actually ran into this issue a couple of months ago. Luckily myself and our
dev ops guy immediately realized the issue (abs paths required) and fixed our
process: We bu
Github user pferrel commented on the issue:
https://github.com/apache/incubator-predictionio/pull/328
Actually the manifest may be gone but the requirement of the same absolute
path is maintained, I believe. A hash of the path to the template code is used
to create the id and is part
Github user chanlee514 commented on the issue:
https://github.com/apache/incubator-predictionio/pull/328
@pferrel: Yes the hash of engine directory path is used as engineVersion
(which kind of acts as the id).
However, I'm working on making updates to this PR so that users ca
Github user pferrel commented on the issue:
https://github.com/apache/incubator-predictionio/pull/328
Cool, that's awesome. This news means the scope has changed, which is great
but brings up the visibility of the id.
Could I suggest that the user-visible engine-instance-id b
If the absolute filepath is the same in the machines (since SHA-1 hash of
filepath is used as engineVersion), there would be no changes in the above
workflow. You can just copy the engine directory and deploy. If not, you
can also provide engine-id, engine-version as CLI arguments.
There's some ad
16 matches
Mail list logo