Possible Computer Problems
							6/21/00, JMB

	Below are some possible problems that may be encountered while I
am gone.  On VAX log on as HEPoperator, or as root on Unix.  John can 
probably handle any Unix problem, if he is around.  If all else fails for 
VAX, contact Marc Cozzi at the Rad Lab, 1-5412.

Print problems:
	There is a note in the Console Reference on print queue problems.
If the LJ 4M+ or LJ 5M says toner low, the toner cartridge can usually be 
shaken up and used for a few hundred more copies, before print quality 
deteriorates. We have toner kits stored under the 4M+.  Note that there
are cartridges for the Laserwriters on the shelf over the 8mm tape cabinet.
After changing a toner kit, please print a test page and date it and put
it in my slot of the shelves.  To get a test page, take it offline, press
Menu until TEST MENU shows, then press Item for Test Page, then press Enter
(4M+) or Select (5M).  Be sure it is online after the test page.

Internet (smtp) mail problems:
	See the note in the Console Reference, or Helpnd web page.

Disk problems:
	See the note in the Console Reference, or Helpnd web page.

Network problems:
	See the note in the Console Reference, or Helpnd web page.

Xterm problems:
	See the note in the Console Reference, or Helpnd web page.

Power outages:
	UNDHD0 must be up to boot any of the VAXes.  If a node fails to boot, 
turn it off and try again after UNDHD0 is up.  UNDHE9 must have the "B"
entered by hand, as its diagnostics think there is a SCSI problem.
The monitor on UNDHE9 is dead, so you will just have to wait a while and
enter the B, and if UNDHD0 does not respond with activity, wait a little
while and try again.  If you do not succeed with UNDHE9, just leave it, as
all essential stuff is on UNDHD0.  UNDHE4 failed during the last power outage 
and has effectively been removed from the cluster.  The contents of the HE4 
disks have been moved to DU02 and DU90.  Once UNDHD0 is up and the 
BATSYSTARTUP job is done, the Xterminals can be rebooted, by cycling the 
power or typing "bd" at the prompt.

	Proton should be up first under controlled powerup, but all can come 
up in parallel.  You will probably have to do 
	mount -a
on all nodes after a total outage, especially if you do not see /u6 and/or 
/usr/local.  If an undersized /u6 or /usr/local appears in df, dismount it
and redo mount -a.  Undhe8 failed this spring and has not been investigated.

Hardware Problems:
	UNDHD0 and the fast IBMs and the SGI's are under maintenance.  See
the Helpnd web pages for contact numbers.  For UNDHD0 and the 4 fast IBMs, 
the Service Express notebook behind the monitor of UNDHD0 has the phone 
numbers and service records.  The SGIs are board-swap maintenance.  John 
should be able to handle this if he is around.