The task list follows
KSD 153
with tasks added at the end which are more ESI-specific.
Comments on specific implementation are added in italics.
2. Initial Instrument Computer Setup
 |
2.1 Set up a new name for the instrument computer.
Instrument computer is "kanaha," CCD computer is "koki".
|
 |
2.2 Set up summit user accounts with home directories rooted on the
instrument machine. Accounts will be "esi," "esieng," and "esi1" through
"esi20". HQ accounts will be rooted in a similar place to other instruments (e.g. papahaku:/home/instrhome1/...).
Currently both kanaha and koki have separate home directories. This
may be revisited at some point after the first commissioning run.
|
 |
2.3 Set up passwords appropriately.
|
 |
2.4 Make "esi" account available to other machines via yp.
Kalama/kawa will not have these accounts.
|
 |
2.5 Modify yp (kalama auto.home, etc.) and kalama /etc/hosts for kanaha
and koki.
Will not be done for ESI, to minimize crossmounts.
|
 |
2.6 Check that kanaha and koki can access Keck printers (summit and HQ).
JChock. Also, ask JSimmons to create a bits queue.
|
 |
2.7 A link to summit /nightly2.
We are not going to do this for ESI to avoid crossmounts. For
MAlign data, will create an malign subdirectory.
|
 |
2.8 Is a disk mirroring system in place?
Need to implement a cron job. (JChock)
|
 |
2.9 Do we have enough disk space for science?
Yes. We now have ~64 GB, since Jon got six more 9 GB disks for us.
|
 |
2.10 Assign IP addresses.
|
3. Licenses
 |
3.1 Check if IDL, Dataviews, etc. licenses are in place.
|
4. Software (Libraries, etc.)
 |
4.1a Populate instrument computer with latest versions of SKY, Facsum, show,
xshow, modify, DCS keyword library, DCS simulator and simulator support
package.
FACSUM needs to have "-noacs"; use Solaris HQ machine for SKY
b. Bring over ACS keyword library if needed (e.g. for MAlign). Bring over
libkidl, etc. Maybe just bring over "core" /kroot.
Don't need libkidl on kanaha yet.
c. Make dcs keyword library available via "dcs" as well.
|
 |
4.2 Populate DCS computer with the instrument keyword library if the
operating systems are compatible.
OSs are not compatible.
|
 |
4.3 Make software /kroot tree owned by kics.
|
 |
4.4 See if realtime data reduction support is needed which requires
instrument keyword libraries to be available on non-instrument
machines.
Not needed at this point.
|
 |
4.5 Bring over to instrument machine: /usr/local/multiple_screens, os, etc.
Not needed.
|
5. Instrument Account on Summit
 |
5.1 Try and retain most of original stuff.
|
 |
5.2 Test account in Keck's whole networked environment. (E.g. getting to
other summit machines, other HQ machines, etc.)
|
 |
5.3 In the main INSTRUMENT account (on summit and HQ), create a ~/config
subdirectory that represents the template for numbered instrument
accounts. Set this up under CVS.
Not under CVS yet (EChock, low priority).
|
6. Other User Account Details
 |
6.1 Modify .login file to support multiple_screens.
|
 |
6.2 Update /usr/local/etc/multiple_screens file with new instrument machine,
if it is multiheaded.
Not necessary.
|
 |
6.3 Update .openwin-menu per generic .openwin-menu.
For HQ accounts only, add keck_menu (not completely available on the
summit).
|
 |
6.4 May need to set up IDL environment variables to ensure enough IDL licenses
are available--use the generic setup.
Done at HQ.
|
 |
6.5 Determine if we will initially run instrument from the instrument account,
the engineering account, or one of the numbered accounts. Set up pull-down
options appropriately.
Instrument will be run from esi1 or esi2 during commissioning.
|
7. Testing at HQ and on the Summit
 |
7.1 Test each type of account on summit and at HQ.
|
 |
7.2 Make sure tcsh works.
|
 |
7.3 Ensure we do not have auto-logout. (Disable it by adding "unset
autologout" in .cshrc file.)
|
 |
7.4 Confirm that xgterm works.
|
 |
7.5 Confirm that Netscape works, including bookmarks.
|
 |
7.6 Check that IDL, IRAF (cl, ncl, saoimage, saotng, ximtool) work from
appropriate machines. May need to do lmkiraf beforehand, although this should
be done in "conk."
|
 |
7.7 Test show/xshow/modify of instrument keywords, as well as DCS keywords.
|
 |
7.8 Test guider eavesdrop, Facsum, SKY (plan mode), XMET, tkroses (instrument
and guider), and chopper display if applicable.
|
 |
7.9 Check that you can open remote windows from the pull-down menu.
|
 |
7.10 From the engineering and numbered accounts, and the instrument
master account if desired, make sure that you can run the instrument.
|
 |
7.11 Check the FITS headers for keyword completeness and accuracy.
BobG to go through FITS headers for completeness.
|
 |
7.12 From HQ only, check access to "keck_menu".
|
8. Accessing DCS (Real and Simulated Modes)
 |
8.1 There is no 8.1.
|
 |
8.2 Test ability to switch between real and simulated modes, and
that the correct keywords are being read.
|
9. Science Data Disks
 |
9.1 Assign science data disk sequence (e.g. /sdata700, /sdata701, etc.)
BobK needs to rename the ESI disks /data1 and /data2
to /sdata700 and /sdata701.
|
 |
9.2 Make data disks available via "/s" from both summit and HQ.
JSimmons, after BobK does 9.1
|
10. Other Tasks
 |
10.1 Add instrument to Remedy's Instrument list.
|
 |
10.2 Add instrument to dcsgui "Select Instrument" and create ~kNtel/qfix
parameter files.
|
 |
10.3 Add instrument's guider(s) to xguide pull-down menu if appropriate.
To be released Wednesday, August 18.
|
 |
10.4 Update MAlign files if appropriate, and create an MAlign control
script.
Tcl/tk GUI for MAlign control has been produced and debugged.
|
 |
10.5 Update SKY to support new instrument.
AConrad and TStickel
|
 |
10.6 Update "conk" script for numbered accounts, both HQ and summit.
Currently must be run from manuka.
|
 |
10.7 Update data disk scrubbing program.
NIRC-style "wiper" and "trickle" are in place, using new paradigm
which copies data to an HQ scratch disk before sending to a bits queue.
Still undergoing testing. (JSimmons/BobG)
|
 |
10.8 Update archiving procedures (STB).
In place, but see above for problems with the kanaha bits queue.
|
 |
10.9 Update the daily chortl script to check instrument's science disk
available space.
|
 |
10.10 Update automated archive monitor (archmon). Add to instrument startup
script.
Waiting for archiver issues to settle. (AConrad)
|
11. Other ESI Tasks
 |
11.1 Add new accounts to the appropriate group(s) or create those
groups. Also add to starlist group.
No action at this time except to add to starlist group.
|
 |
11.2 Check time source for xntp if appropriate.
|
 |
11.3 Add instrument computers and accounts to tape backup schedule.
|
 |
11.4 Change hostnames on private net to new
hostnames.
BobK, JChock, summit
|
 |
11.5 Create a DCS oneshot command for instrument, if appropriate.
AConrad, low priority.
|
 |
11.6 Create MAlign files for ESI guider.
|
 |
11.7 Move Dataviews from /u/alt-u/kics/dv to a more appropriate place.
AConrad has installed Dataviews 9.5.
|
 |
11.8 Diagnose "packet storm" problem during readout display.
BobK and JChock
|
 |
11.9 Test DCS v3.4 and v3.2h on Keck II.
DCS v3.4 has problems on both telescopes; we have used v3.2h for ESI
engineering.
Plan is to test v3.4 during November ESI run.
|
 |
11.10 Provide OAs a guider control window.
DClarke
|
 |
11.11 Add tip line from terminal server to CCD crate.
JChin/RMatsuda, summit, high priority
|
 |
11.12 Move tkrose into the DCS release structure on kanaha, and get running.
EChock
Currently we can use tkrose, but it is not within the DCS release structure.
|
 |
11.13 Move XMET into the DCS release structure and get running.
EChock
|
 |
11.14 Get disk.space and swapwatch running as cron jobs.
EChock; BobG to provide disk.space trigger value.
|
 |
11.15 Write "dump_watch" to check backup dates.
Al has initial version; needs to be integrated into cron job, etc.
|
 |
11.16 Put a /kroot/starlists link on kanaha.
|
 |
11.17 Get tklogger working.
|
|