Old Geezil Install Log

Geezil install log - Chris Laws

Sep 14, 2009

basic geezil hardware in place and running at the office.

installed fedora 11, but was unable to get the scisoft package working (problem with 'libtermcap.so.2').

Sep 16, 2009

installed fedora core 6 with all default options, but also added the scientific and engineering packages, software development tools, and all available system tools.

had significant issues with 'yum' and updates, due to the old and largely unsupported nature of FC6. had to change the 'baseurl' entries in the /etc/yum.repos.d/fedora*.repo files to:

//archive.fedoraproject.org/pub/archive/fedora/linux/core/6/i386/ //archive.fedoraproject.org/pub/archive/fedora/linux/core/6/i386/debug/ //archive.fedoraproject.org/pub/archive/fedora/linux/core/6/source/SRPMS/

//archive.fedoraproject.org/pub/archive/fedora/linux/core/updates/6/i386/ //archive.fedoraproject.org/pub/archive/fedora/linux/core/updates/6/i386/debug/ //archive.fedoraproject.org/pub/archive/fedora/linux/core/updates/6/SRPMS/

//archive.fedoraproject.org/pub/archive/fedora/linux/extras/6/i386/ //archive.fedoraproject.org/pub/archive/fedora/linux/extras/6/i386/debug/ //archive.fedoraproject.org/pub/archive/fedora/linux/extras/6/SRPMS/

then everything updated fine except 'pidgin'. this was deselected in the updater, and everything went fine and i was able to install 'pidgin' via the updater the next time around without problems.

Sep 17, 2009

installed emacs ('21.4.17.3.fc6.rpm') using the 'add/remove software' menu.

installed Scisoft following the online guide from eso.org.

as root, created the text file /etc/yum.repos.d/scisoft.repo with the following contents:

[scisoft] name=Scisoft baseurl=ftp://ftp.eso.org/scisoft/scisoft7/linux/fedora6/yum-repo gpgcheck=0 enabled=1

then i ran the following in the same directory, as root:

yum clean all yum install scisoft-\*

many headers were downloaded, followed by a prompt to confirm package download ('y' + enter). 58 packages were installed.

added the following to /root/.bashrc and /home/MROuser/.bashrc:

. /scisoft/bin/Setup.bash set autolist

when i sourced this, i got an error message warning me that SELinux would prevent PyRAF and possible other packages from running properly (oliver saw this previously with asgard). more information on this is available at the Scisoft FAQs page.

i disabled SELinux, as we had done previously. on reboot i was able to source .bashrc without problems, and launched pyraf successfully.

i created /root/bin and /home/MROuser/bin/ -- i'm not sure if this is necessary, but oliver has it in his notes from asgard...

in /home/MROuser/ (and now logged in as MROuser), i created ~/iraf. i then ran 'mkiraf' and set terminal type to xgterm. i then edited the login.cl file with the following changes:

uncommented "set stdimage", set value to '= imt1100'; uncommented "set imtype", set value to '= "fits"'.

iraf then launched fine.

i tested the soundcard, but did not install an MP3 player.

on shutdown, the computer hung up after "System halted.". i powered geezil down and rebooted, and it came back up just fine. i "Restart"ed it, and it both shut down and restarted without problems. i then "Shut (it) Down", and it again got all the way through to "System halted" before hanging up. it turns out this is a known issue with power management in FC6 (why didn't it show up earlier? beats me!). i edited the /boot/grub/grub.conf so that the two lines that look like:

kernel /vmlinuz-2.6.22.14-72.fc6 ro root=...ETC

now read:

kernel /vmlinuz-2.6.22.14-72.fc6 acpi=off apm=on apm=power-off ro root=...ETC

i then rebooted, and now geezil completely powers off when asked. huzzah!

side note: the video card fan has been buzzing loudly throughout all of this. i just gave it a good cleaning, and it's better now, but i suspect we'll have to change the card out sooner over later.

jim davenport installed idl 7.1 - his notes follow...

downloaded the linux 32bit x86 version of IDL from ITT (the current owner of IDL).

in accordance with Oliver's install log from 2007, i downloaded and unpackaged the .tar.gz file into /scisoft/share/idl.

ran the command ./install to run IDL installation in /scisoft/share/idl.

allowed IDL to set symbolic links.

put standalone dept IDL license file on so IDL can be run w/o internet or network access.

created ~/idl folder in the MROuser home directory.

added the following line to .bashrc file after the scisoft Setup.bash line to automatically have IDL read in any programs the user downloads or writes and puts in the ~/idl folder:

IDL_PATH="$IDL_PATH:+/home/MROuser/idl:"

grabbed WHICH.pro from UW Astro servers, put it in ~/idl and tested, works!

bookmarked the IDL Astronomy Library Search in Firefox (which i find useful).

IDL is installed and (appears to! -cl) works perfectly!

Sep 18, 2009

as root, in /scisoft/share/iraf/iraf/noao/imred/ccdred/ccddb/kpno/ i created the file mro.dat, containing the following text:

subset filter

BIAS zero OBJECT object FLAT flat

in the same directory, i added the following text to instruments.men:

mro Manastash Ridge Observatory

in /scisoft/share/iraf/iraf/noao/lib/obsdb.dat, i added the following:

observatory = "mro" name = "Manastash Ridge Observatory" longitude = 120:43.668 latitude = 46:57.168 altitude = 1198 timezone = 8

i created the file /etc/fstab.formount by copying /etc/fstab and adding the following two lines:

chivens.mro.cwu.edu:/var/www/html/logfiles /mnt/gtcc/ nfs auto,ro 0 0 //popeye.mro.cwu.edu/My\040Documents /mnt/popeye/ cifs auto,ro,username=mro,password=mrouser 0 0

at MRO we setup geezil and plugged it directly into the network switch below popeye. Using the gnome GUI tool:

system-config-network

i assigned the proper fixed IP address, gateway,subnetmask,DNS, and search path (see wiki page on fixed IP's and networking @ MRO). This tool can be launched from the terminal by typing the command name above. after restarting geezil the internet and networking worked

as root, we created the /mnt/gtcc and /mnt/popeye directories, and then test mounted chivens and popeye manually using the following commands (from the asgard.install.log):

mount -t nfs chivens.mro.cwu.edu:/var/www/html/logfiles /mnt/gtcc/ mount -t cifs "//popeye.mro.cwu.edu/My Documents" /mnt/popeye/

the GTCC mount worked fine, but the popeye mount did not -- the name was not resolved by the network (?). we switched the full name out for popeye's static IP address as follows:

mount -t cifs "//72.233.250.80/My Documents" /mnt/popeye/

this worked just fine!

we then made similar changes to the popeye line in the /etc/fstab.formount file, then copied the original fstab file to fstab.old.butworks, and then copied fstab.formount to fstab. on reboot, everything mounted automatically, and we're able to see files directly on popeye and GTCC from Geezil. huzzah!

in /home/MROuser, i created the /data directory, and within this directory i created the following symbolic links (for use by add_headers.py):

ln -s /mnt/popeye ln -s /mnt/gtcc/ logs

we tested ds9 and encountered the following error message:

XPA unable to verify hostname

within the ds9 'Edit/Preferences' menu, we cleared the 'Initialize XPA' check box, and this seems to resolve the problem. what problems did we create? who knows. it seems to be working now.

we dowloaded a copy of add_headers.py (version 1.0) to the ~/data directory and made it executable with the following:

chmod u+x add_headers.py

we then created an image directory within ~/data (as per the MRO Observers Checklist instructions), and from within that directory we launched ds9, and then attempted to run add_headers. it failed multiple times due to bad directory paths (corrected in our brand-new version 1.1 of add_headers.py), but continued to run into a problem with launching add_headers.py directly from the command line. we invariably got the following error message:

bash: add_headers.py: command not found

this happened no matter where we launched add_headers from. however, by adding 'python ' to the command (see the documentation_for_add_headers) in the following manner (and from within the ~/data/RawDataDirectory on geezil - in this case, ~/data/091809):

python ../add_headers.py 091809

we were able to get add_headers successfully updating images taken on popeye. huzzah for us again!

April 2011

Geezil was rebuilt (again) mostly following the above notes:

Changes: SciLinux 5.5 instead of Fedora.:

Did not use yum to install scisoft (7.3), just installed from the tar.gz files. Needed to install a few other things first (installed version in parentheses):

No SELinux errors.

home directory is now /mro data directory is now /mro/data

Still need to install IDL...

Running add_headers.py (version 2.0) for spyder, and add_headers_flarecam.py (version 2.0) for flarecam. Notes for running it are on the add_headers page.

Computer is still named asgard, need to get in touch with CWU to change (I think).