Skip to main content

I'm working FME Server 2014. I'm having trouble using a job server submitter to output the details on the output from the called workspace on the server. The output from the called workspace does get generated on the server however it doesa not appear in a single folder. Three of the writers (DWG, XLSX, TEXT) in the called workspace have a dataset fanout assigned to them while a fourth (CSV) has a feature_fanout in place. This seems to be at the root of the problem but I cannot determine exactly what the problem is. I have read about the use of the FME_SERVER_DEST_DIR to modify the location of the dataset fanout however the CSV fanout persists in going to a folder under C:\\Windows\\Temp\\... on the server, while the dataset fanouts go to a uuid subfolder under FME_SHAREDRESOURCE_TEMP. In my calling workspace I set up an attribute to define the fanout folder location using the FME_SHAREDRESOURCE_TEMP folder and a UUID, and passed this to the job submitter parameter for the fanout path on the called workspace. I have tried many variations on this and I cannot get the output to go into a single folder on the server. The calling workspace indicates it completed successfully but the log shows the called workspace has failed.

Has anyone else had a similar problem and if so how did you work around it?

Thx!

Hi @wallace

While obviously we'd like to see you upgrade to FME Server 2016 I think this is possible with FME Server 2014. I got this to work in FME Server 2014 SP5 as follows.

Assuming that MYFOLDER is the single folder you want to write to within the Resources/Data folder of FME Server.

1. In workbench create a private parameter called:

FME_SHAREDRESOURCE_DATA and set the value to a local temp director

2. For your writers where you want to use dataset fanout like DWG set the Fanout Directory to:

$(FME_SHAREDRESOURCE_DATA)/MYFOLDER

3. For writers where you want to use feature type fanout such as CSV set up your feature type fanout as you normally would. But upstream use an AttributeCreator to create an attribute called Folder and set this to MYFOLDER. Then use Fanout Dataset on this writer as well and set the fanout attribute to Folder (so all of these files will go to the same folder and set the Fanout Directory to $(FME_SHAREDRESOURCE_DATA) which is the linked parameter you created in Step 1.

PS the workspace I used is here:

https://www.dropbox.com/s/tj47g6keb9qtvcg/mapinfo2acad.fmw?dl=0


Hi Again @wallace

I read your post a little more carefully and realized that you want the folder to have uuid. Since we need both a parameter with uuid and an attribute for fanout I think we need to use a python scripted parameter. I modified the workspace above to include a Private Scripted parameter called myuuid.

So now the fanout directory parameter for the writers where you wanted dataset fanout like DWG would then look like this:

$(FME_SHAREDRESOURCE_DATA)/$(myuuid)

For the writers that use feature type fanout use an AttributeCreator to add an attribute _myuuid and get the value from the scripted parameter myuuid. Then use this as the fanout attribute keeping the fanout directory as $(FME_SHAREDRESOURCE_DATA)

I've shared the modified workspace here:

https://www.dropbox.com/s/4x4c7i6o83as4tm/mapinfo2acadUUUID.fmw?dl=0

Hope this helps


Reply