student.cs upgrade

Trying to upgrade from ubuntu 9.10 to opensuse using a custom suse studio build...

Loaded it onto a usb, if everything goes well I'll use xen to manage virtual machines.  Ideally I would like to use separate virtual machines for the different services that we provide.

It is a little sluggish from the usb...

It is going to be a little learning curve using yast

hmm... I think I can just dd from the usb into a partition and then redo the boot loader.... let's see



so I reinstalled the boot loader, and it over wrote the usb boot loader, blech

still no luck, I think I'll just do a live cd install and then patch it up later, the usb stick is already corrupted, otherwise I could have copied the boot loader from it.

so far so good, updating the packages now....

lab status

112 hardware fail
119 upgrade fail

for some reason the network is down in the lab, but I'm still connected to 141.... weird
well, my guess is that it was a power fail because I did a wakeonlan and everything came back,
weird.

a little while later we loss network connectivity in feinberg

ppa firewall work around

from the internet

gpg –keyserver hkp://keyserver.ubuntu.com:80 –recv-keys 0A5174AF
gpg –armor -o keyfile –export “key name” // (use gpg –list-keys to look it up)
sudo apt-key add keyfile

moodle notes

perlbrew

perlbrew is a cool application.  It lets you install multiple instances of perl with their own libraries.  This let me do development work using 5.14 without messing with the system's 5.10  It took a while to install everything for the catalyst app to work, but it was mostly painless.

open suse

so, I've been trying out opensuse.
  - xen is working, although I don't have any guest os running right now.  I thought about using my ubuntu partition, but it complained about a missing mbr, and I'm not sure that it is worth fixing that
 - it does not have as many prepackaged catalyst bits, so I'm using cpan
 - I'm not sure I like yast, the gui manager, it may just take some getting used to, but
   it seems to be nonintuitive, make things more cumbersome than they need to be
 - I'm hoping to upgrade the kernel soon, because the video support is not great,
  I can't do full screen video on the amd fusion system yet.
 - At some point I want to see how well it runs windows xp as a guest os, it would be
   nice if I could migrate the kids games to it.

mediawiki sidebar

I end up editing this often, but there is no convenient link, so for future reference MediaWiki:Sidebar

mediawiki spam

well, there was a bunch of media wiki spam.  I attempted to get rid of it with
update revision left outer join user on `rev_user` = `user_id` set revision.rev_deleted = 1 where user_id is NULL and rev_page > 1

after I deleted the bogus users.  I did a backup before hand just in case I got a query wrong.

It seems to have mostly worked.

open id

checked on adding openid to catalyst, but it seems broken and not used much.

todo

Although it is nice to know about the schema, I think the next step should be working on the checklist server to prepare it for software engineering.  And I think I should keep that as a separate database since it is student accessible.

I want to improve the logging/communication of the 436 and 446 students.  Tomorrow I'll work on the syllabi for those courses and 336 and then look at the evaluation mechanisms and see what I need to implement to make it work well.

I need to start working on the videos for 336 also.

catalyst, postgresql and schemas

the idea here is to use multiple schemas instead of databases to reduce redundancy,

the basic catalyst create model doesn't seem to handle non default schemas, so I created one with the default schema,

oops, I found the option after working through the loader documentation and making my own script.  it is odd because I don't see where the selection shows up in the source files.

much later:
I still have not found where it is identifying the schema used.  I tried some different experiments though and it is distinguishing between the two versions of the same table/view.  My current guess is that since it is inaccessible from one of the roles that it is skipping it.  It would be nice though if it was explicitly stated somewhere.

postgresql config

actually the next step is to migrate the configuration files so I can connect from other places.

ppa forwarding

to upgrade the database, I need to add a ppa archive, but the outbound
port is blocked, the port number is 11371

on db:

on server:
sysctl net.netfilter.nf_conntrack_acct=1
sysctl net.ipv4.ip_forward=1

and then the rules
iptables --append FORWARD -j ACCEPT
iptables -t nat --append POSTROUTING  -j MASQUERADE

and then on the client I set the gw for that host to be server.

The above may be a bit overkill, but it worked.  I tried using iptables on the client to forward the packets without setting the gateway, but I didn't have any luck with that.

Now onto upgrading postgres.

First backup with pg_dumpall

Next aptitude dist-upgrade

Then use psql to reimport the data, and it looks ok.

the next task is to connect to a different schema than public using perl

attendance/usage tracking

I have a separate database for attendance tracking, I think I am going to merge it into the department one though to reduce redundancy.

First though, I'm going to back up the databases, and then try to upgrade them to version 9

ldap admin

worked on web app to help manage the ldap group that contains add-ons to the allowed user list.  It works well and should be useful.