Is there a size limit for GP services?

[JonMorris2]

I'm trying to upload a GP tool to our server and the staging was reported as a success, but the GP service does not have any tasks. From the REST services directory:

Test/lfpv1r (GPServer)

Service Description: LFP v1

Tasks:

None

Execution Type: esriExecutionTypeAsynchronous

Result Map Server: Test/lfpv1r

If you click on the map server link:

Error: Service Test/lfpv1r/MapServer not found

Code: 500

Is there a maximum size for the python script? I've managed to upload simple test tools ok, but the script I'm trying to upload is 6.5k lines long, with 90 input parameters (don't ask!). Even a cut-down version at 2.5k lines fails.

I can split the tool into stages and upload each one as a separate service, which is probably what I'll end up doing, but I just wondered if there was some kind of size limit that was preventing me from uploading the full tool.

[JonMorris2]

Cross post in Geoprocessing​

[VinceAngelo]

Ninety parameters seems like eighty too many.  I'd recommend constructing a properties file to manage the inputs, then passing in the path to the properties file as one parameter.

- V

[JonMorris2]

Yes, I'd agree with you there, but it's a complicated geological process we're trying to model. I'm not a geologist so I'd rather not mess with the script too much, but I would prefer to break it up into smaller stages, and that's probably what I'll end up doing. We'll be calling this GP service through the javascript api, so ultimately parameters will be passed in as one object.

I just wondered if there was any kind of fixed size limit under the hood or if anyone had had similar problems when trying to publish large scripts. Or if anyone was doing this kind of thing at all, because I've had very little feedback so far!

[ChrisSmith7]

I wonder... what would happen if you uploaded one of the simple, working tools with enough random comments to reach 6,500 lines? You could also add 90 dummy parameters... If the bloated test tool works locally, but experiences the same behavior when published to the server, then we may be able to rule out script-specific issues... this would lend credence to a script size issue. Maybe also try an unbloated test tool with 90 dummy params.