Large file transfer intermittently fails with XCOMU0505E

Products

• XCOM Data Transport
• XCOM Data Transport - Linux PC
• XCOM Data Transport - Windows

Issue/Introduction

There is an intermittent problem when attempting a large transfer of a file over 100 GB in size using XCOM for AIX 11.6 SP01 to XCOM for Windows 11.6 SP03.

=== 
2023/12/02 15:14:07 TID=193100 PRG=xcomtcp PID=41550114 IP=remote_host PORT=8045
    XCOMU0811I Starting Secure TCP/IP Connection.
 
2023/12/02 15:14:07 TID=193100 PRG=xcomtcp PID=41550114 IP=remote_host PORT=8045
    XCOMU0813I Secure TCP/IP Connection Requested.
 
2023/12/02 15:14:07 TID=193100 PRG=xcomtcp PID=41550114 IP=remote_host PORT=8045
    XCOMU0814I Secure TCP/IP Handshake Complete.
 
2023/12/02 15:28:28 TID=193100
    #XCOMU0505E Received a signal from TCP/IP. 
 
2023/12/02 15:28:32 TID=193100 [/tmp/file1 --> C:\temp\file1 at remote_host]
    XCOMU0047I Local transfer restarted .
 
2023/12/02 15:28:32 TID=193100 PRG=xcomtp PID=8454716 IP=remote_host PORT=8045
    XCOMU0811I Starting Secure TCP/IP Connection.
 
2023/12/02 15:28:32 TID=193100 PRG=xcomtp PID=8454716 IP=remote_host PORT=8045
    XCOMU0813I Secure TCP/IP Connection Requested.
 
2023/12/02 15:28:32 TID=193100 PRG=xcomtp PID=8454716 IP=remote_host PORT=8045
    XCOMU0814I Secure TCP/IP Handshake Complete.
 
2023/12/02 15:28:32 TID=193100 PRG=xcomtp PID=8454716 IP=remote_host PORT=8045
    XCOMU0818I Secure TCP/IP Connection Ended.
 
2023/12/02 15:28:32 TID=193100
    XCOMN0441I Remote restart request does not match any queue entry.
===

Cause

Error "XCOMU0505E Received a signal from TCP/IP" was reported on the AIX side of the transfer which is from the AIX TCP/IP layer indicating the connection has been closed for some reason. 

XCOM Engineering advised the root cause is likely due to some network glitch. The longer a transfer lasts the more likely it is a network problem will interrupt it. XCOM is quite capable of dealing with this situation, as long as its configuration is appropriate. In this example, a restart was attempted 4 seconds after the initial failure, but failed with the XCOM Windows error "XCOMN0441I Remote restart request does not match any queue entry".
The reason that XCOM Windows rejects the restart is not so much that the transfer doesn’t exist on the Windows side of the transfer, but that it is still in an active state. For restarts to be successful, the transfer has to be in a RESTARTABLE state on both sides of the transfer.

Resolution

Delay the restart time on AIX by setting the RETRY_TIME parameter to a larger value. In this case the RETRY_TIME was 0.
Edit the XCOM for AIX transfer configuration file:
Change: RETRY_TIME=0 to RETRY_TIME=60
Change: NUMBER_OF_RETRIES=0 to NUMBER_OF_RETRIES=10

The XCOM for AIX server will wait until RETRY_TIME has passed before trying the restart.
The restart will appear in the AIX xcom.log after the RETRY_TIME has passed e.g.:

===
2024/03/13 08:07:59 TID=021036
#XCOMU0505E Received a signal from TCP/IP.

2024/03/13 08:09:01 TID=021036 [/tmp/file1 --> C:\temp\file1 at remote_host]
XCOMU0047I Local transfer restarted .

2024/03/13 08:09:01 TID=021036 PRG=xcomtp PID=16580642 IP=remote_host PORT=8045
XCOMU0811I Starting Secure TCP/IP Connection.

2024/03/13 08:09:01 TID=021036 PRG=xcomtp PID=16580642 IP=remote_host PORT=8045
XCOMU0813I Secure TCP/IP Connection Requested.

2024/03/13 08:09:01 TID=021036 PRG=xcomtp PID=16580642 IP=remote_host PORT=8045
XCOMU0814I Secure TCP/IP Handshake Complete.

2024/03/13 08:10:13 TID=021036 PRG=xcomtp PID=16580642 IP=remote_host PORT=8045
XCOMU0818I Secure TCP/IP Connection Ended.

2024/03/13 08:10:13 TID=021036
XCOMU0011I Transfer ended; 7258 blocks (217720040 bytes) transmitted in 12 seconds (18143336 bytes/second)
===

If the restart is still not successful then the RETRY_TIME may need to be increased further.

Additional Information

XCOM Data Transport for UNIX/Linux 11.6.1 parameter pages:
RETRY_TIME
NUMBER_OF_RETRIES