Loop end node problems

I have encountered a problem with a number of loop end nodes recently (always a 'standard' loop end node, but as part of either a chunk loop or table row to variable loop) where the loop fails part way through execution with the following error in the console:

ERROR Loop End Execute failed: Error while writing to buffer, failed to write to file "knime_container_20111103_2870664986600309219.bin.gz": C:\DOCUME~1\S98C4~1.ROU\LOCALS~1\Temp\knime_container_20111103_17814611933609587\col_0\000\000\0.bin.gz (The system cannot find the path specified)

Browsing to the file path shows that the folder "knime_container_20111103_17814611933609587" appears not to have been written in the temp folder.  This doesnt seem to invariably fail (several previous folders from the same loop execution have been written), and re-executing the failed loop end node seems to allow continuation of the loop from where it left off, but it is obviously not an ideal situation!


As an addendum to the initial post above, although the loop appears to finish executing normally, much of the output table is lost!


Hi Steve,

Thanks for reporting this. It looks like a bug that we fixed yesterday (fix available in v2.5). Does the problem go away if you add this line to the knime.ini file (and restart):


Hope this helps,


Thanks.  I have this running at the moment.  So far, 1500 of 7200 loop iterations completed, which is further than ever before, so looking promising as a solution.  I will let you know if it fails later!


As an aside, it would be a great feature to be able to manually restart a loop which has failed somewhere in the intervening nodes on the next iteration!  I suggest this as I am using some 3rd party nodes which sometimes unpredictable are unable to handle the external data files and crash, meaning I have to remove the failed file and try starting again from the beginning.

I have exactly same problem with newest KNIME version. Can not finish LOOP anyhow. Is there any vure for this?