Date: Fri, 26 Jan 2001 10:34:42 -0600 (CST)
From: James Annis
Subject: Re: 2.5-m Obslog 2001012 (MJD 51935)
The short version:
   the problem is resolved, the gang tape is being written,
   and the tapes will be taken to FedEx today, either through
   the normal routine or by an emergency drive down to Alamogordo.
The longer version:
The response of Eric, Steph, and endNight was appropriate at each step,
up until they were stopped by a problem related to the camera power supply
giving up earlier in the night.
When the power supply gave out, the DA wrote out zero-length
images to the pool. IOP noticed that they existed and tried
to archive them, but Astroda threw an error saying it could not archive
these zero-length frames.  If one looked at the murmur log, one would see:
   sdssid1    dArchSrv   TEXTONLY archCmd: archive 2072 42.11 0 1
   sdssid1    dArch      TEXTONLY archCmd: single frame archive 2072 42.11
   sdssid1    dArch      dsc_E_noFrameHdr Frame header does not exist
   sdssid1    dArch      TEXTONLY *** FATAL   *** dscArch: While double writing frames
5 times, one for each of the frames. Doing a
	poolDir 2072
would show that the files were zero length. Doing
	poolDelFrame 2072 42.11 42.11
deleted one of them from the pool. After deleting all 5, endNight
ran normally.
Just a related note:
> We try running endNight with all the switches except for framecheck,
> to get it to skip the framecheck. There was no change in the result.
When one does -doImager, there is no way to get past doing a
frameCheck. This was a design choice requested by the observing team.
The code is called twice; Eric (at the end of a long night) found only
one of the two places in the code where it is fired.
jim