Requested RM_Server repository does not exist?
hi,
I tried to run a process on server from RM Studio... I saved the process on server repo and run it, my browser opens and says that some file from retrieve operator cannot be found.. the thing is, the data it is on my local RM Studio repository , and I even copied them into data folder on my RM Server repository..but when I want to retrieve them, it says there is no repository ?
here is the log:
Aug 07, 2016 9:16:23 AM <unknown> <unknown>
INFO: Process context is: Macros: [_ra_user : admin, _ra_jobID : 8]; Input: []; Output: []
Aug 07, 2016 9:16:23 AM <unknown> <unknown>
INFO: No filename given for result file, using stdout for logging results!
Aug 07, 2016 9:16:23 AM <unknown> <unknown>
INFO: Process //_LOCAL/process/160805-k-NN_test_train_perf starts
Aug 07, 2016 9:16:23 AM <unknown> <unknown>
SEVERE: Process failed: com.rapidminer.operator.UserError: Cannot retrieve repository data from entry '//RM_Server/data/Master3_0.3TestData'. Reason: Requested repository RM_Server does not exist..
Aug 07, 2016 9:16:23 AM <unknown> <unknown>
SEVERE: Here: Process[1] (Process)
subprocess 'Main Process'
+- Retrieve[0] (Retrieve)
+- Split Data[0] (Split Data)
+- Write Excel (2)[0] (Write Excel)
+- Write Excel[0] (Write Excel)
==> +- Retrieve testData[1] (Retrieve)
+- Select Attributes (3)[0] (Select Attributes)
+- Sample (2)[0] (Sample (Bootstrapping))
+- Retrieve Master3Klassen_nominal[0] (Retrieve)
+- Select Attributes[0] (Select Attributes)
+- Sample (Bootstrapping)[0] (Sample (Bootstrapping))
+- Multiply Trainings Data[0] (Multiply)
+- Optimize Parameters (Grid)[0] (Optimize Parameters (Grid))
subprocess 'Optimization Process'
| +- Validation[0] (X-Validation)
subprocess 'Training'
| | | +- Bagging[0] (Bagging)
subprocess 'Learning Process'
| | | +- MetaCost (2)[0] (MetaCost)
subprocess 'Learning Process'
| | | +- k-NN[0] (k-NN)
subprocess 'Testing'
| | +- Apply Model[0] (Apply Model)
| | +- Performance[0] (Performance (Classification))
| +- Log[0] (Log)
+- Set Parameters[0] (Set Parameters)
+- k-NN2[0] (k-NN)
+- Multiply Model[0] (Multiply)
+- Write Model[0] (Write Model)
+- Apply Model (2)[0] (Apply Model)
+- Performance (2)[0] (Performance (Classification))
+- Log Train Perfromance[0] (Log)
+- Apply Model (3)[0] (Apply Model)
+- Performance (3)[0] (Performance (Classification))
furthermore, I don't see any repository folder on my Rapidminer Server installation folder, so where are they stored? maybe I am mixing up something with local REpository folders and RM Server repo folders... however if I do browse from RM_Serve repository, a web window opens showing me the Serve folders from my Server repository:
com.rapidminer.operator.UserError: Cannot store data in repository at entry '//RM_Server/models/knnmodel'. Reason: Requested repository RM_Server does not exist..
where are those stored and what do I need to configure to get access to my data and processes?
Best Answers
-
Hi,
the most important difference in terms of repositories between Studio and Server is: Studio can have as many as you want, Server only has a single one, and it's hidden from your eyes. Reason is that you can name your repositories in Studio as you like - so multiple people will have different names for the same RM Server repository.
Usually you should resolve all locations relative to your process (after you have saved your process on server). That way, you can for example simply do a Store operator with "myData" as location and it will be saved right next to your process. Or "../myData" to store one directory above your process. Or "../../myData" to store two directories above your process and so on.
If for some reason you want to explicitly mention the hidden RM Server repository, you can do so. Use "//_LOCAL/myData" to store "myData" into the root folder of the Server repository. "//_LOCAL/" is the internal reference to the hidden RM Server repository.
Hope this helps make things a bit clearer!
Regards,
Marco
3 -
Hi,
without the leading /. So "results/myDataResults" would be correct
Regards,
Marco
0
Answers
-
Hi,
It looks like there are several misunderstandings about how to use the Server repositories. I thought that it might be easier to just post one very helpful video instead of trying to figure out all the details with dozens of messages going back and forth :smileywink:
So please watch this video here:
https://www.youtube.com/watch?v=I23co-N5Cz4
And then try again to work with your server. I think this should clarify most of your questions but please come back in case it still does not work for you after watching the video.
Cheers,
Ingo
0 -
ok thanks, but there is still not much explanation how the repositories work..
and btw, RapidAnalytics == Rapidminer Server (in later editions?)
retrieving Data from RM Server repository now works fine, but I have an issue when I want to use the store operator.. to save my results ...
If I set some store repository, it stll gives me an error:
com.rapidminer.operator.UserError: Cannot store data in repository at entry '//RM_Server/models/knnmodel'. Reason: Requested repository RM_Server does not exist..
I'm not getting behind how this works sorry...
edit: I just edited the repository from RM_Server to my local repository (didn't work), then I switched back to store results to my RM_Server repository folder, and this time, the warning triangle disappeared, whereas with the same configuration (on the first try), the triangle persisted (altough it was the same folder). That's weird, maybe it is a bug but this time it ran successfully and the results were also stored.
0 -
Yes, "RapidAnalytics" was the old name for "RapidMiner Server".
Could it be that you saved the process in the meantime on the Server and the path got resolved relatively to the process location?
Here is a tip when you work with Server or in general with multiple repositories: It is in general a good practice to first store any process (even if it is empty at this time) at the desired location before you start building the process. This way it is ensured that all locations are resolved in a relative fashion instead of using absolute paths. This way the process will still run if you move the process into a different repository (e.g. the Server).
Anyway, happy mining,
Ingo
0 -
Hi,
the most important difference in terms of repositories between Studio and Server is: Studio can have as many as you want, Server only has a single one, and it's hidden from your eyes. Reason is that you can name your repositories in Studio as you like - so multiple people will have different names for the same RM Server repository.
Usually you should resolve all locations relative to your process (after you have saved your process on server). That way, you can for example simply do a Store operator with "myData" as location and it will be saved right next to your process. Or "../myData" to store one directory above your process. Or "../../myData" to store two directories above your process and so on.
If for some reason you want to explicitly mention the hidden RM Server repository, you can do so. Use "//_LOCAL/myData" to store "myData" into the root folder of the Server repository. "//_LOCAL/" is the internal reference to the hidden RM Server repository.
Hope this helps make things a bit clearer!
Regards,
Marco
3 -
ok and if my process is in subfolder "process" and I want to store results in the results subfolder (on same hierarchy), so I have to save in "/results/myDataResults" ??
0 -
Hi,
without the leading /. So "results/myDataResults" would be correct
Regards,
Marco
0
