Re: HPSS logs

[Henrik Nordberg <HNordberg@lbl.gov>]

I got clarification from Shigeki that the logs are snapshots of a
circular buffer and so they may contain the same data.
This might be a bad thing: if there is a lot going on the buffer might
complete a turn before we get a snapshot, and we'll loose information.
Or, as we see happened here, there could be less activity and we waste
space.

One would think that a system like HPSS, critical in many cases I am
sure, would ship with proper logging tools?

The files I looked at were `cmp log.010000 log.010030` and `cmp
log.010030 log.010100` (from memory) in the Star directory.

 - Henrik

dave zimmerman wrote:
> 
> Hi Henrik,
> Which files which were identical?  There are some problems with the
> records from the phenix runs.  I am looking into them still, but I had
> thought that the STAR logs were in good shape.
> david
> >
> > I had a look at the HPSS logfile that are on starsu00. I found that much
> > of the information is duplicated across the files. A cmp on the two
> > first files show that the first 14000 lines are the same. The same
> > number for the next two is 19000.
> >
> > Dave, did you know this?
> >
> > I wrote a nawk script (StarLogs/foo) that will parse out when each mount
> > happened. But since a lot of information in the logs is duplicated, so
> > is the output of the script. It would be very nice to somehow strip the
> > duplicate lines. Hints?
> >