When we run multiple KNIME batch executions in parallel, at least one of the executions fail. At the same time at least one succeeds for sure. The error message the failed execution throws is shown below.
- workspace is different for all these executions
- user is same for all
WARN main Node Node can’t be executed - Node “SDF Reader” not available from extension “KNIME Base Chemistry Types & Nodes” (provided by “KNIME AG, Zurich, Switzerland”; plugin “org.knime.chem.base” is installed)
What else could be shared by these parallel batch executions that I should take care of?
Any help would be really appreciated as this is blocking one of our main feature.
Please find attached a simple workflow (just with a SDF Reader and SDF Extractor nodes) and see below code for runner script that runs this workflow in batch 5 times simultaneously.
from subprocess import Popen
BATCH_COMMAND = ["/software/lib/Linux-x86_64/knime_4.1.3/knime", "-application",
if __name__ == '__main__':
for i in range(5):
cmd = 
jobname = random.randint(1, 999999)
proc = Popen(cmd, shell=False,stdin=None,stdout=None,stderr=None,close_fds=True)
In one of my runs, out of 5, 2 succeeded and 3 failed, in another run 3 succeeded and 2 failed.
The error is:
2020-09-24 16:30:16,842 : WARN : main : : Node : MISSING SDF Reader : 0:6 : Node can’t be executed - Node “SDF Reader” not available from extension “KNIME Base Chemistry Types & Nodes” (provided by “KNIME AG, Zurich, Switzerland”; plugin “org.knime.chem.base” is installed)
Please let me know if you want us to try something else?
Simple_KNIME_Wkf_for_batch.knwf (17.4 KB)
Ok. In an offline conversation with KNIME team, the issue turned out to be the conflicting configuration directory. For all these parallel batch executions, since I did not specify any configuration directory, it used the same default one under home directory.
Specifying a unique configuration directory via -configuration option fixed the issue.
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.