download_reliably.sh

#!/bin/bash

reliably_download()
{
    ssh_user=$1
    remote_host=$2
    remote_file=$3
    local_file=$4
# default to 3 tries
    max_tries=${5-3}
    limit=$((max_tries+1))

    local_file_md5=''
    remote_file_md5=`ssh -n ${ssh_user}@${remote_host} "md5sum ${remote_file} | awk '{print \\\$1}'"`
    ntries=1
    while [ \( "${local_file_md5}" != "${remote_file_md5}" \) -a \( "${ntries}" \< "${limit}" \) ]
    do
        echo "downloading ${host}:${remote_file} - attempt ${ntries}"
        scp ${ssh_user}@${remote_host}:${remote_file} ${local_file} >/dev/null
        local_file_md5=`md5sum ${local_file} | awk '{print $1}'`
        ((ntries += 1))
    done
    if [ "${local_file_md5}" == "${remote_file_md5}" ]
    then
        return 0
    else
        return 1
    fi
}


echo BEFORE
ls -l /tmp/test.xml
echo

reliably_download "hadoop" "chris.in.escapemg.com" "/tmp/original/test.xml" "/tmp/test.xml" 5
echo

echo AFTER
ls -l /tmp/test.xml