|
Instructions
|
|
- Check the INPUT PROCESS :
from the Web Browser Stacker window:
- LOGFILE button< --> $INPUT/logs/Stacker_$machine.log
- DATABASE button --> $INPUT/db/SpoolToDisk.txt
- DB button (for each file)
--> $INPUT/db/SpoolToDisk_$tapeName.txt
- LOG button (for each file)
--> $INPUT/logs/SpoolToDisk_$tapeName.logs
Write down which input tapes were spooled to disk,
and note if any of those had files with unusually small or big
size (i.e. differnt from about 150Mb), or if the tape had an unusual number
of files (i.e. different from 28 or 29).
- Check the ANALYSIS PROCESS :
- From the Web Browser ColoFarm window:
- LOG button --> $INPUT/run/ColoFarm.log
- JOBS button
--> jobs currently submitted to the DNS batch system
- CRASH/IMPORT/PREBATCH buttons
--> listing of corresponding directory
Make sure the analysis process is still running (i.e. submitting jobs
to the batch system). Note which files have been stored in the $crash directory
and mark them down on your sheet close to the corresponding input tape.
- From the Web Browser ColoSkim window, Tapes List frame:
- $runNumber button --> listing of all
files processed for that input tape
Make sure that all input tapes have been succesfully analized
- Check the OUTPUT PROCESS :
- from the Web Browser ColoSkim window, Skim Status frame:
- (right) LOG button
--> $OUTPUT/data/ColoSkimSpool.log
- Next Tape button
--> $OUTPUT/data/skm${i}db/$skmTape.log
- Tapes button
--> $OUTPUT/data/skm${i}db/TapesToBeSpooled_skm${i}.txt
- (left) LOG button
--> $OUTPUT/data/ColoSkimUpdate.log
Check that every output tape was written succesfully and that
no stream remained in a lock state (meaning some problem occurred).
- For each skim tape written out, use the script
$CGI/DecodeTapeLabel
to read the tape header and check it against the tape name
for that stacker slot.
- Label each skim tape.
- JOB SUBMISSION :
- Load the tape stacker with new input tapes.
- From the Web Browser Stacker window:
- Insert the input tape labels
- START button --> starts
the Tape to Disk spooling process
- Wait untill the stacker loads correctly the first slot
and the first input file is spooled to the $IMPORT
directory
- Make sure the analysis process submits the first file
in the batch queue system
- STACKER RESET :
- From the Web Browser ColoSkim window, Skim Status frame:
- Make sure the OUTPUT process is not running
- Enable Lock Spooling checkbox
- Insert new blank tapes in the stacker
- RESET button --> reset the stacker to start
back from slot=1
- Disable Lock Spooling checkbox