cancel
Showing results for 
Search instead for 
Did you mean: 

Intermittent errors using FTP-adapter

KennethEriksen
Participant
0 Kudos

Hi all!

In our PI 7.00 SP14 environment we have several FTP-adapters running, both sender and receiver. For most of the time they work fine, but now and then the following error message can be found in the communication channel monitoring:

"Error occurred while connecting to the FTP server "myserver:21": java.net.SocketException: Software caused connection abort: recv failed"

This causes an alert to be triggered, which fills up my alert inbox with irrelevant error messages.

I have tried to update/change the FTP-server software without luck, and the FTP-server logs say nothing. Other posts on this forum do not solve my problem.

How do I solve this?

Is it related to the Java installed on the PI-server?

How can I exclude such error messages from alerts (adapter engine alerts cannot be filtered by error message as far as I know)?

Many thanks for your help!

Br,

Kenneth

Accepted Solutions (1)

Accepted Solutions (1)

Former Member
0 Kudos

Hi,

This problem is either caused by incorrect firewall / packet filter settings or an incorrect configuration of the FTP server. Also make sure that you have correctly specified the host name / IP address and port of the FTP server.

Changing the connection type from 'active' to 'passive' (or vice versa) might additionally help to work around the incorrect firewall configuration

Regards,

Suryanarayana

KennethEriksen
Participant
0 Kudos

Thanks for your response.

Both PI and the FTP-server are located on the same LAN and are not separated by a firewall/packet filter or any other "obstacle" on the network. The hostname/IP-address and port numbers are also correct as most of the time the communication channels work just fine and messages flow through. But sometimes the polling to the FTP-server returns the error stated in my first post.

Any other suggestions?

Answers (1)

Answers (1)

KennethEriksen
Participant
0 Kudos

Closes the thread as we switched to using NFS instead, and the problem was solved.