I'm trying to use model trained offline onto the server. Thus I'm looking to move a wordlist and a model from a local repository to a server repository. However I get the following error: "Error on copying model Cause: cannot upload object: internal Server Error"
Any ideas? I have the Text Processing extention installed on the server already
naive bayes? That should not cause any problems. How are you storing the model? For debugging purposes, first try just right-clicking the model tab in the results pane and just manually save it on your server. That SHOULD work no problem. Like this:
The solution is most likely hidden in the log files of the server.
Just open the server.log directly after reproducing the error, take a look at the error message.
I would assume two possible reasons:
a) Installation of Text Processing Extension failed it can't work with the WordList
b) (more likely) you are using MySQL as database backend for RM Server and MySQLs max_allowed_packet size is smaller than the serialized IOObject. Then the writing to the database will fail and there is returned an internal error. You can increase this setting in your MySQL config and need to restart both MySQL and RM Server. Directly set it to 1G, which is the maximum value.
Professional consulting for your Data Science problems
I think this is the same issue as that one :
I have posted logs on that other thread.
PS : On my side max_allowed_packed_size was already at 1G.
I am refering to this issue and also to the other thread which I think is related (see the link below).
I can confirm that your workaround works, i.e. right-clicking the model tab and saving it in the remote repository this way.
This is a very complex issue but in a nutshell
All using 7.6.1 across the board whith fresh MySQL reinstall.
I can open a new post to clearly identify the context in which this happens.
I already copied logs on the other thread so let me know if this can be helpful.
hello @sebas -
Yes it seems that something rather irregular is going on. Could you please open a new thread in this forum so that the dev team can follow along? Any data sets and processes you could repost, in addition to log files and configurations, would be helpful so we can try to replicate exactly what's going on here. I monitor this new Bug/Issue forum pretty tightly (so people don't just post normal questions here) but this will go through.