Question

Job restarts itsef on FME Server


Badge +1

Hi,

While running job on FME Server I noticed it is getting restarted.

Workspace I'm trying to run has Smallworld reader and GML writer (+few attribute manipulation transformers). It starts at e.g. 10:00, and after a few time Smallworld console returns stream exhausted traceback at e.g. 10:42. At the same time my job (started at 10 o'clock) is now running and started at 10:42.

This repeats itself few time when it finally gives failure status with no description in log of that job.

If I run same workspace in Desktop it results in FME.exe crash.

Thanks,

Želimir


11 replies

Userlevel 1
Badge +18

Can you provide the workspace and logging. Can you test if it also crashes when you only read data with the smallworld reader? Are you running 64 or 32 bits?

Badge +1

Can you provide the workspace and logging. Can you test if it also crashes when you only read data with the smallworld reader? Are you running 64 or 32 bits?

I've also done test sollely with readers and it also crashes. We're running on 32 bit. Workspace and log is also attached.

 

export-sw-gml.fmw

 

fmejob-114822log.txt
Userlevel 1
Badge +18
I've also done test sollely with readers and it also crashes. We're running on 32 bit. Workspace and log is also attached.

 

export-sw-gml.fmw

 

fmejob-114822log.txt
Did you follow the instructions from this link? https://knowledge.safe.com/articles/107/fme-and-smallworld-configuration-options.html

 

 

Badge +1
yup, we've been doing it for years and many other workspaces. And this is first time this kind of error occured.
Badge +2

1. Have you tested the workspace using desktop on the same server/machine where the server is installed?

2. Is your desktop and server versions the same?

Badge +1

1. Have you tested the workspace using desktop on the same server/machine where the server is installed?

2. Is your desktop and server versions the same?

Yes and yes :)

 

 

We've detected that problem might be in one feature type (gis.building_location_regular). If I use this workspace reading everything but that feature type... All ok. If I run it only with building_location_regular... Also ok. If I run it together... crash

 

 

 

 

Badge +2

1. Have you tested the workspace using desktop on the same server/machine where the server is installed?

2. Is your desktop and server versions the same?

For the immediate why don't you create 2 workspaces and use a fmeServerjobsubmitter transformer in a 3rd worksoace to run them consecutively . not sure if you can do with this temporary solution

 

 

Badge +1
For the immediate why don't you create 2 workspaces and use a fmeServerjobsubmitter transformer in a 3rd worksoace to run them consecutively . not sure if you can do with this temporary solution

 

 

yes, thank you. we've been working on few workarounds also inside code and FME. I'd really like to find out why is this happening...

 

 

 

Userlevel 1
Badge +18
yup, we've been doing it for years and many other workspaces. And this is first time this kind of error occured.
Can you log any feature that's coming out of the reader before it crashes?

 

Use a GeometryValidator to check and fix.

 

 

Badge +11
Hi @zzupljanin - FYI.. FME Server will resubmit jobs that fail due to FME Engine crashes for a maximum of 3 times. This parameter is set by default and can be modified in a configuration file. Any changes will require a restart of FME Server.

 

 

<FMEServer_InstallDir>/Server/fmeServerConfig.txt

 

(~Line 422) MAX_FAILED_TRANSACTION_REQUEST_RETRIES=3

 

 

"Maximum number of times a failed transaction request will be automatically resubmitted for processing, after which the failed request is cancelled. A value of < 0 means that no maximum limit is imposed, and failed transaction requests will be automatically resubmitted indefinitely. A value of 0 means that automatic resubmission is disabled."

 

Badge +11
I've also done test sollely with readers and it also crashes. We're running on 32 bit. Workspace and log is also attached.

 

export-sw-gml.fmw

 

fmejob-114822log.txt
@zzupljanin - Re: "[...] and after a few time Smallworld console returns stream exhausted traceback[...]"

 

 

Can you include a snippet of the traceback from the Smallworld Console?

 

Reply