| 1 |
#!/bin/bash
|
| 2 |
#---------------------------------------------------------------------------------------------------
|
| 3 |
# Go through a number of download requests
|
| 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"
|
| 9 |
|
| 10 |
# health checks
|
| 11 |
if [ ".$1" == "." ]
|
| 12 |
then
|
| 13 |
echo "";
|
| 14 |
echo "usage: $H <sampleList>";
|
| 15 |
echo ""
|
| 16 |
exit 1
|
| 17 |
fi
|
| 18 |
|
| 19 |
# say what we do
|
| 20 |
echo ""
|
| 21 |
echo " ==== Download queue from $SAMPLE_LIST ===="
|
| 22 |
echo ""
|
| 23 |
|
| 24 |
# check that queues are really empty
|
| 25 |
startTime=$(date +%s)
|
| 26 |
nowTime=$(date +%s); duration=$(($nowTime - $startTime))
|
| 27 |
jobs=`condor_q -global $USER | grep downloadFiles.sh | wc -l`
|
| 28 |
while [ $jobs -gt 15 ]
|
| 29 |
do
|
| 30 |
echo " waiting since $duration sec == condor queue has $jobs jobs left"; sleep 60; echo ""
|
| 31 |
jobs=`condor_q -global $USER | grep downloadFiles.sh | wc -l`
|
| 32 |
nowTime=$(date +%s); duration=$(($nowTime - $startTime))
|
| 33 |
done
|
| 34 |
echo " Queues are close to empty ($jobs) -- Let's get started."
|
| 35 |
echo ""
|
| 36 |
|
| 37 |
# all right, ready to download
|
| 38 |
nSamples=`cat $SAMPLE_LIST | wc -l | cut -d ' ' -f1`
|
| 39 |
i=1
|
| 40 |
while [ $i -le $nSamples ]
|
| 41 |
do
|
| 42 |
|
| 43 |
line=`head -$i $SAMPLE_LIST | tail -1`
|
| 44 |
if [ "`echo $line | grep ^#`" != "" ]
|
| 45 |
then
|
| 46 |
i=$(( $i+1 ))
|
| 47 |
continue
|
| 48 |
fi
|
| 49 |
|
| 50 |
# define some quantities
|
| 51 |
baseDir=`echo $line | tr -s ' ' | cut -d ' ' -f 1`
|
| 52 |
book=`echo $line | tr -s ' ' | cut -d ' ' -f 2`
|
| 53 |
version=`basename $book`
|
| 54 |
dataset=`echo $line | tr -s ' ' | cut -d ' ' -f 3`
|
| 55 |
|
| 56 |
# stagein the sample if it is at CERN
|
| 57 |
if [ "`echo $baseDir | grep /castor/cern.ch`" != "" ]
|
| 58 |
then
|
| 59 |
ssh paus@lxplus.cern.ch ./stageSample.py --dataDir=$baseDir/$book/$dataset
|
| 60 |
fi
|
| 61 |
|
| 62 |
# download the sample
|
| 63 |
echo " downloadSample.sh $line"
|
| 64 |
downloadSample.sh $line
|
| 65 |
|
| 66 |
# go into waiting loop
|
| 67 |
nowTime=$(date +%s); duration=$(($nowTime - $startTime))
|
| 68 |
jobs=`condor_q -global $USER | grep downloadFiles.sh | wc -l`
|
| 69 |
while [ $jobs -gt 15 ]
|
| 70 |
do
|
| 71 |
echo " waiting since $duration sec == condor queue has $jobs jobs left"; sleep 60; echo ""
|
| 72 |
jobs=`condor_q -global $USER | grep downloadFiles.sh | wc -l`
|
| 73 |
nowTime=$(date +%s); duration=$(($nowTime - $startTime))
|
| 74 |
done
|
| 75 |
echo " Queues are empty ($jobs) --> cleaning up and making catalogs."
|
| 76 |
echo ""
|
| 77 |
|
| 78 |
# remove zero length files
|
| 79 |
echo "removeZeroLengthFiles.sh /mnt/hadoop/cmsprod/$book/$dataset"
|
| 80 |
removeZeroLengthFiles.sh /mnt/hadoop/cmsprod/$book/$dataset
|
| 81 |
|
| 82 |
# finally make the corresponding catalog
|
| 83 |
echo "catalog.sh -ceg $version $dataset --retry /mnt/hadoop/cmsprod"
|
| 84 |
catalog.sh -ceg $version $dataset --retry /mnt/hadoop/cmsprod
|
| 85 |
|
| 86 |
i=$(( $i+1 ))
|
| 87 |
|
| 88 |
done
|
| 89 |
|
| 90 |
exit 0
|
