When automating file transfers via ncftp in bash scripts, network instability or temporary server issues can cause failures. Unlike human operators who retry manually, scripts need explicit error handling.

The $? variable stores the exit code of the last executed command:

ncftpput -DD -z -u user -p password server /remote /local/file
echo $? # Prints 0 for success, non-zero for failure

The original approach has syntax issues. Here's a corrected version:

until [ $? -eq 0 ]; do
    ncftpput -DD -z -u user -p password server /remote /local/file
done

Basic loops risk infinite retries. Add these improvements:

max_retries=5
retry_count=0

until [ $? -eq 0 ] || [ $retry_count -ge $max_retries ]; do
    ((retry_count++))
    ncftpput -DD -z -u user -p password server /remote /local/file
    
    if [ $? -ne 0 ]; then
        echo "Attempt $retry_count failed, retrying in 5 seconds..."
        sleep 5
    fi
done

[ $retry_count -ge $max_retries ] && echo "Max retries reached" >&2

For production systems, implement progressive delays between retries:

max_retries=8
base_delay=2
max_delay=300

for ((i=1; i<=max_retries; i++)); do
    ncftpput -DD -z -u user -p password server /remote /local/file
    
    if [ $? -eq 0 ]; then
        echo "Transfer succeeded"
        break
    fi
    
    delay=$((base_delay ** i))
    [ $delay -gt $max_delay ] && delay=$max_delay
    
    echo "Attempt $i failed, retrying in $delay seconds..."
    sleep $delay
done

For reusable code across scripts:

retry_command() {
    local cmd="$@"
    local max_attempts=3
    local attempt=1
    
    until $cmd || [ $attempt -ge $max_attempts ]; do
        echo "Command failed, retrying..."
        ((attempt++))
        sleep 1
    done
    
    return $?
}

retry_command ncftpput -DD -z -u user -p password server /remote /local/file

Differentiate between temporary and permanent failures:

ncftpput -DD -z -u user -p password server /remote /local/file
exit_code=$?

case $exit_code in
    0) echo "Success" ;;
    1) echo "Usage error - not retryable" >&2 ;;
    2) echo "Connection failed - will retry" ;;
    *) echo "Unknown error $exit_code" >&2 ;;
esac

When automating file transfers with ncftp, network hiccups or temporary server issues can cause failures. Instead of manually restarting failed transfers, we can create a robust bash script that automatically retries until success.

In bash, every command returns an exit status code. A value of 0 indicates success, while non-zero means failure. We can leverage this in our retry logic.

Here's a simple implementation that keeps retrying until the transfer succeeds:

#!/bin/bash

until ncftpput -DD -z -u user -p password remoteserver /remote/dir /local/file
do
    echo "Transfer failed, retrying in 5 seconds..."
    sleep 5
done

For production use, we should add several improvements:

#!/bin/bash

MAX_RETRIES=10
RETRY_DELAY=30
RETRY_COUNT=0

until [ $RETRY_COUNT -eq $MAX_RETRIES ] || ncftpput -DD -z -u user -p password remoteserver /remote/dir /local/file
do
    ((RETRY_COUNT++))
    echo "Attempt $RETRY_COUNT failed, waiting $RETRY_DELAY seconds..."
    sleep $RETRY_DELAY
done

if [ $RETRY_COUNT -eq $MAX_RETRIES ]
then
    echo "Transfer failed after $MAX_RETRIES attempts."
    exit 1
else
    echo "Transfer completed successfully."
    exit 0
fi

We can make the script more sophisticated by checking specific error codes:

#!/bin/bash

MAX_RETRIES=5
RETRY_DELAY=10

for ((i=1; i<=MAX_RETRIES; i++))
do
    ncftpput -DD -z -u user -p password remoteserver /remote/dir /local/file
    EXIT_CODE=$?
    
    case $EXIT_CODE in
        0)  echo "Transfer successful"
            exit 0
            ;;
        1)  echo "Local file error"
            exit 1
            ;;
        2)  echo "Connection error (retryable)"
            sleep $RETRY_DELAY
            continue
            ;;
        3)  echo "Authentication error"
            exit 1
            ;;
        *)  echo "Unknown error (retryable)"
            sleep $RETRY_DELAY
            continue
            ;;
    esac
done

echo "Maximum retries reached"
exit 1

For better handling of persistent issues, we can implement exponential backoff:

#!/bin/bash

MAX_RETRIES=8
INITIAL_DELAY=5
BACKOFF_FACTOR=2

DELAY=$INITIAL_DELAY

for ((i=1; i<=MAX_RETRIES; i++))
do
    echo "Attempt $i of $MAX_RETRIES..."
    ncftpput -DD -z -u user -p password remoteserver /remote/dir /local/file
    
    if [ $? -eq 0 ]
    then
        echo "Transfer successful"
        exit 0
    fi
    
    echo "Waiting $DELAY seconds before retry..."
    sleep $DELAY
    DELAY=$((DELAY * BACKOFF_FACTOR))
done

echo "Transfer failed after $MAX_RETRIES attempts"
exit 1