Repository navigation
Error publishing large twbx files #754
Description
While publishing a large twbx file with latest 0.14 version of tableau-server-client using the below code,
wb_item = server.workbooks.publish(wb_item,target_file,'Overwrite')
I am getting the following error.
File "/usr/local/lib/python3.9/site-packages/tableauserverclient/server/endpoint/fileuploads_endpoint.py", line 45, in read_chunks
chunked_content = file.read(CHUNK_SIZE)
ValueError: read of closed file
The same code works for 0.13 and below versions. And also it works for smaller workbooks, even with extracts.
Activity
bcantoni commentedon Dec 7, 2020
@AnandShiva thanks for reporting this. Can you confirm what version of Tableau Server you are connecting to?
Also, do you have a sense of what size is causing you the problem? If nothing else, maybe you can give the size of the one which fails and the size of a smaller one which passes.
AnandShiva commentedon Dec 7, 2020
@bcantoni I tried with the file of size 800MB and it failed. When I tried with a workbook of 70 MB it passed.
I am facing this issue with the latest 0.14 version of the package. The tableau server is the 2020.03 version.
bcantoni commentedon Dec 7, 2020
Ok thanks, that's helpful.
This might be related to changes which came in to 0.14 via #704.
avishaymahluf commentedon Dec 8, 2020
Hi,
I'm hitting the same issue when trying to publish a large hyper file (1.2GB).
client version - 0.14, server version - 2020.3.1.
noobling commentedon Dec 8, 2020
Do we have a resolution for this? It renders the package pretty unusable...
bcantoni commentedon Dec 8, 2020
I can repro this as well. It seems like any file >64MB will trigger this, so this path likely wasn't tested as part of #704.
A possible workaround until this is fixed is to stick with v0.13 of the library.
AnandShiva commentedon Dec 8, 2020
One thing which I am doing to still keep using the 0.14 package and still not face the issue is, create a small size twbx extract workbook file (~70MB) pointing to a temp smaller data source, change the data source to actual data source programmatically via tableau document API, and publish that.
After publishing trigger workbook refresh via
server.workbooks.refresh(workbook)which will update the workbook to larger data source extract clearing older smaller data source.Or of course, you can just do
pip3 install tableauserverclient==0.13and continue what you are doing anyway.Fixes issue #754 by moving file read logic inside generator
Merge pull request #757 from tableau/fix_754
Merge pull request #757 from tableau/fix_754
shinchris commentedon Dec 10, 2020
Fix for this is released with v0.14.1.
Update site properties and functions (#777)
MANIFEST.in: Add docs and test data (#780)
Squashed commit of the following:
Fixes issue tableau#754 by moving file read logic inside generator
Merge pull request tableau#757 from tableau/fix_754
Update site properties and functions (tableau#777)
MANIFEST.in: Add docs and test data (tableau#780)
Merge pull request tableau#757 from tableau/fix_754