| 4 |
|
#--------------------------------------------------------------------------------------------------- |
| 5 |
|
H=`basename $0` |
| 6 |
|
SAMPLE_LIST=$1 |
| 7 |
< |
MIT_LOCATION="/pnfs/cmsaf.mit.edu/t2bat/cms/store/user/paus" |
| 8 |
< |
CERN_LOCATION="/castor/cern.ch/user/p/paus" |
| 7 |
> |
|
| 8 |
> |
# start next request |
| 9 |
> |
remainingMax=0 |
| 10 |
|
|
| 11 |
|
# health checks |
| 12 |
|
if [ ".$1" == "." ] |
| 23 |
|
echo "" |
| 24 |
|
|
| 25 |
|
# check that queues are really empty |
| 26 |
+ |
dataset="XXX" |
| 27 |
|
startTime=$(date +%s) |
| 28 |
|
nowTime=$(date +%s); duration=$(($nowTime - $startTime)) |
| 29 |
< |
jobs=`condor_q -global cmsprod | grep downloadFiles.sh | wc -l` |
| 30 |
< |
while [ "$jobs" != "0" ] |
| 29 |
> |
jobs=`condor_q -global $USER -format "%d " ClusterId -format "%s " Cmd -format "%s \n" Args | grep $dataset | grep downloadFiles.sh | wc -l` |
| 30 |
> |
while [ $jobs -gt $remainingMax ] |
| 31 |
|
do |
| 32 |
|
echo " waiting since $duration sec == condor queue has $jobs jobs left"; sleep 60; echo "" |
| 33 |
< |
jobs=`condor_q -global cmsprod | grep downloadFiles.sh | wc -l` |
| 33 |
> |
jobs=`condor_q -global $USER $format | grep $dataset | grep downloadFiles.sh | wc -l` |
| 34 |
> |
#jobs=`condor_q -global $USER | grep downloadFiles.sh | wc -l` |
| 35 |
|
nowTime=$(date +%s); duration=$(($nowTime - $startTime)) |
| 36 |
|
done |
| 37 |
< |
echo " Queues are empty ($jobs) Let's get started." |
| 37 |
> |
echo " Queues are close to empty ($jobs) -- Let's get started." |
| 38 |
|
echo "" |
| 39 |
|
|
| 40 |
|
# all right, ready to download |
| 42 |
|
i=1 |
| 43 |
|
while [ $i -le $nSamples ] |
| 44 |
|
do |
| 45 |
+ |
startTime=$(date +%s) |
| 46 |
|
|
| 47 |
|
line=`head -$i $SAMPLE_LIST | tail -1` |
| 48 |
|
if [ "`echo $line | grep ^#`" != "" ] |
| 52 |
|
fi |
| 53 |
|
|
| 54 |
|
# define some quantities |
| 55 |
+ |
baseDir=`echo $line | tr -s ' ' | cut -d ' ' -f 1` |
| 56 |
|
book=`echo $line | tr -s ' ' | cut -d ' ' -f 2` |
| 57 |
|
version=`basename $book` |
| 58 |
|
dataset=`echo $line | tr -s ' ' | cut -d ' ' -f 3` |
| 59 |
+ |
targetDir=`echo $line | tr -s ' ' | cut -d ' ' -f 4` |
| 60 |
|
|
| 61 |
< |
#echo " downloadSample.sh $line" |
| 62 |
< |
mkdir -p /mnt/hadoop/cmsprod/$book/$dataset |
| 61 |
> |
|
| 62 |
> |
# stagein the sample if it is at CERN |
| 63 |
> |
if [ "`echo $baseDir | grep /castor/cern.ch`" != "" ] |
| 64 |
> |
then |
| 65 |
> |
echo " ssh paus@lxplus.cern.ch ./stageSample.py --dataDir=$baseDir/$book/$dataset" |
| 66 |
> |
ssh paus@lxplus.cern.ch ./stageSample.py --dataDir=$baseDir/$book/$dataset |
| 67 |
> |
fi |
| 68 |
> |
|
| 69 |
> |
# download the sample |
| 70 |
> |
echo " downloadSample.sh $line" |
| 71 |
|
downloadSample.sh $line |
| 72 |
|
|
| 73 |
+ |
# go into waiting loop |
| 74 |
|
nowTime=$(date +%s); duration=$(($nowTime - $startTime)) |
| 75 |
< |
jobs=`condor_q -global cmsprod | grep downloadFiles.sh | wc -l` |
| 76 |
< |
while [ "$jobs" != "0" ] |
| 75 |
> |
jobs=`condor_q -global $USER -format "%d " ClusterId -format "%s " Cmd -format "%s \n" Args | grep $dataset | grep downloadFiles.sh | wc -l` |
| 76 |
> |
while [ $jobs -gt $remainingMax ] |
| 77 |
|
do |
| 78 |
|
echo " waiting since $duration sec == condor queue has $jobs jobs left"; sleep 60; echo "" |
| 79 |
< |
jobs=`condor_q -global cmsprod | grep downloadFiles.sh | wc -l` |
| 79 |
> |
jobs=`condor_q -global $USER -format "%d " ClusterId -format "%s " Cmd -format "%s \n" Args | grep $dataset | grep downloadFiles.sh | wc -l` |
| 80 |
|
nowTime=$(date +%s); duration=$(($nowTime - $startTime)) |
| 81 |
|
done |
| 82 |
|
echo " Queues are empty ($jobs) --> cleaning up and making catalogs." |
| 83 |
|
echo "" |
| 84 |
|
|
| 85 |
|
# remove zero length files |
| 86 |
< |
echo "removeZeroLengthFiles.sh /mnt/hadoop/cmsprod/$book/$dataset" |
| 87 |
< |
removeZeroLengthFiles.sh /mnt/hadoop/cmsprod/$book/$dataset |
| 86 |
> |
echo "removeZeroLengthFiles.sh $targetDir/$book/$dataset" |
| 87 |
> |
removeZeroLengthFiles.sh $targetDir/$book/$dataset |
| 88 |
|
|
| 89 |
|
# finally make the corresponding catalog |
| 90 |
< |
#echo "catalog.sh -ceg $version $dataset --retry /mnt/hadoop/cmsprod" |
| 91 |
< |
catalog.sh -ceg $version $dataset --retry /mnt/hadoop/cmsprod |
| 90 |
> |
if [ ${#version} == 3 ] |
| 91 |
> |
then |
| 92 |
> |
echo "catalog.sh -ceg $version $dataset --retry $targetDir" |
| 93 |
> |
catalog.sh -ceg $version $dataset --retry $targetDir |
| 94 |
> |
else |
| 95 |
> |
echo "This is not a normal dataset, make a simple catalog." |
| 96 |
> |
echo "" |
| 97 |
> |
list $targetDir/$book/$dataset > /tmp/list.$$ |
| 98 |
> |
mkdir -p ~/catalog/t2mit/private/$book/$dataset |
| 99 |
> |
rm -f ~/catalog/t2mit/private/$book/$dataset/Files |
| 100 |
> |
touch ~/catalog/t2mit/private/$book/$dataset/Files |
| 101 |
> |
while read line |
| 102 |
> |
do |
| 103 |
> |
size=`echo $line | cut -d ' ' -f1` |
| 104 |
> |
file=`echo $line | cut -d ' ' -f2` |
| 105 |
> |
root -l -b -q $MIT_PROD_DIR/root/runSimpleFileCataloger.C\(\"$targetDir/$book/$dataset\",\"$file\"\) \ |
| 106 |
> |
2> /dev/null | grep CATALOG >> ~/catalog/t2mit/private/$book/$dataset/Files |
| 107 |
> |
done < /tmp/list.$$ |
| 108 |
> |
rm -f /tmp/list.$$ |
| 109 |
> |
fi |
| 110 |
|
|
| 111 |
|
i=$(( $i+1 )) |
| 112 |
|
|
