Details
-
Bug
-
Status: Resolved
-
Major
-
Resolution: Feedback Received
-
1.12.1
-
None
-
None
-
None
-
RHEL 7.8
java version 1.8.0_281
-
Important
Description
Hi All,
Hope everything is fine. I’m writing you as we are experiencing a problem in Nifi 1.12.1 on a RHEL 7.8 with java version "1.8.0_281" .
It is a single node, secured, without any load balancer upstream. However, it does have a s2s (it process provenance from another machine), ListSFTP + FetchSFTP processors as well as the ExecuteStreamCommand processor.
After few days running NiFi suddenly shuts down reporting the following message:
INFO [main]org.apache.nifi.bootstrap.Command Apache NiFi is running at PID XXXX but is not responding to ping requests
Looking at the log file we are facing the following WARN:
WARN [NiFi Web Server-XX-acceptor-X@XXXX -ServerConnector@XXXX{SSL,[ssl, http/1.1]}
{hostname:9443}] o.eclipse.jetty.server.AbstractConnector java.io.IOException: Too many open files
We tried to fix the problem my means of the best practice configurations (https://nifi.apache.org/docs/nifi-docs/html/administration-guide.html#configuration-best-practices) but the problem persists.
After some research we found out that a possible cause might be the use of the processor FetchSFTP that we use in our flow. It creates always new files during SFTP connections (https://issues.apache.org/jira/browse/NIFI-7222). We would like to substitute the .nar file you attached to jira but, unfortunately, it is no more available.
Based on your experience, is it possible that the bug found in NiFi 11.1.4 and 1.12.0 could still be present in the 1.12.1?
Thank you in advance for your support and let me know if you need further details