Subject: Re: About NGC001/NGC002 copy jobs
From: Erotokritos Katsavounidis (kats@hep213.cithep.caltech.edu)
Date: Thu Jun 22 2000 - 19:37:04 EDT
Nicola&Massimo:
Do not use the SLOW queues. It will be probably faster for me
to ftp them at Caltech and put them on tape here ...
In order to effectively address the problem you need to do ONE
of the following:
1) modify the GZIP+TAR TAPEMAKER in order to launch a SEPARATE BATCH
JOB gzip-ing a file before copying it one tape *or*
2) Add in the CLEAN scipt of the NGC processing the GZIPing
of the RAW GC file successfully processed. Any FAST queue
should be able to handle it within its limit. Then modify
the GZIP+TAR TAPEMAKER and SKIP the GZIP part.
You shouldn't spend more than 10 minutes inserting this modification
in the scripts. I do want to see the disks cleaned and new NGC jobs
launched today; we are running significantly late in this project.
--Erik
> Hi Erik,
>
> yesterday we launched the first copies of NGC001/NGC002 Gzipped DSTs on
> axpgs0/3_batch queues but both jobs crashed because of a CPU time limit
> expired.
> Today we relaunched them again on the SLOW queues but they seem to get
> a lot of time to Gzip/TAR these DSTs on tape.
> Have you any suggestion how to reduce the time ??
>
>A presto,
> Massimo&Nicola
This archive was generated by hypermail 2a24 : Thu Jun 22 2000 - 19:37:06 EDT