Subject: Tape shift report
Nicola.Zaccheo@lngs.infn.it
Date: Fri Jun 30 2000 - 11:58:14 EDT
This is the "tape shift" report from 26-JUN up to 30-JUN-2000
NGCXX processing:
I modified the CLGC_MON job in order to g-zip the files if their
processing went fine. I also mofified the TAPEMAKER in order to skip the
files g-zipping (only TAR now). Unfortunatly, this was not sufficient and
the GC*.DST-GZ TAR job crashed again for CPU time limit expired on
AXPGS0_BATCH. I re-launched the TAPEMAKER on the SLOW queues. Nazzareno
promised to give us a higher privilege in the AXPGS queues use.
Since the NGC processing lasts ~30 hrs, it's maybe better do not include the
TAR procedure in it (it's not convenient to allocate the tapes devices for
so much time).
A new SELECT_ALLGC has been implemented in order to select all the TRK/CAL/
MUDST/LIP DST files produced in the NGC processing..I also modified the
SELECT_DDM scripts in order to grab the TRK/CAL files from the DDM processing.
MUDST/LIP-DSTs for NGC01-NGC04 and TRK/CAL files for DDM17-DDM01/NGC01-NGC04
have been uploaded on tapes (MUDST2, NLIP01 and TKCL01 resp.).
Since the jobs are not yet completed, on Monday I'll give you the list of the
MUDST/TRK/CAL/LIP files backupped so that you will delete those files alredy
analysed.
NGC005 and NGC006 processing are still going on...
RAREMAKER:
I launched the RAREMAKER for :
- NRD009/NRD010 that gave no problems (see my diagnosis for more detailes).
- NRD011/NRD012: 3 runs (90354, 90355 and 90356) were missing from NRD012.
I recovered them using RDT1, I manually launched the RAREMAKER and
I copied their ZEB files at the beginning of the N11_12 DLT-IV tape.
Then the RAREMAKER_EMG started and the processing gave no errors.
I tried to copy those runs at the end of NRD012 but the job crashed for
parity error. I decided to repeat this in another moment.
I also downloaded on disk runs 13252, 13453 and 13607 missed from NRD014,
NRD015 and NRD016 resp.
- NRD013/14: is still running. As for NRD011/NRD012 I downloaded run 13252
(that was missing from NRD014) ZEB file at the beginning of the N13_14.
DDM production:
Massimo found out that the files restored from the last available TA90s are
actually ZEBRAs even if they are .DAT files....first of all he noticed that
all of them have a record fixed length of 21600 and, after having renamed
one file as .ZEB, we launched the ZEB2RAW which created the RAW file without
problems.
For this reason we renamed all of them as ZEBRA files and they were
included in the DDM37.
The TA90 tapes have been now fully processed.
The DDM38 is the last TA90s DLT tape. It contains ~ 11.5M blcks.
Massimo reports that 15 "PIZZAs" and corresponding TA90 tapes are
missing (M0035A - M0049A; runs 441 - 672). Note that the first MACRO
run is run000005...we have analysed it producing the relative GC-DST
file...in this manner we can see the first 'good(?)' MACRO event that is
now available on the main MACRO web page.
Regarding the CAL files, we're downloading on disk the full NCLM01 in
order to insert, in the right order, the remaining calibration files
(~ 5.5 Mblks).
REAL TIME ZEB files.
I launched the real time zeb copies/BCK for DD344. Note that now there are
problems during the BCK ('buffer overflow' in the BACKUP list) due to the
higher number of runs..This can be skipped using the "%" wild card instead
of listing the total number of files.
This archive was generated by hypermail 2a24 : Fri Jun 30 2000 - 11:58:17 EDT