Replacing the operating system todo list
From Knot Atlas
Things we should be including in the backup
Okay -- I have a local backup of these, as of --Scott 19:26, 10 Oct 2006 (EDT)
Make paranoid backup copies of the SQL databases and SVN repositories -- Scott's job for Tuesday evening.
- Are we going to have any issues with userids changing?! ie, will permissions get all screwed up when we recreate the various accounts?
- I guess we can gaurantee that userids don't change, by recreating users with the -u switch (see man useradd).
- Here's a list of our current users.
- Reconfigure apache
- Hopefully copying the appropriate sections from the old /etc/httpd/conf/httpd.conf is all that's needed.
- Reinstall SQL 4.0
- See instructions at the top of this page.
- Restoring the SQL databases
- I can't work out where SQL is installed, let alone where it keeps its database files. This is definitely the scary part! --Scott 19:56, 10 Oct 2006 (EDT)
- In addition to doing an sqldump and restore (What's the procedure for restoring? I don't know at all!) we may have to reconfigure the SQL users and passwords.
- Reconfigure the firewall
- See /etc/iptables-firewall.conf
- Reinstall SVN
- Requires finding the RHEL3 rpms.
- Check that up2date works
- We'd previously had problems because Scott had installed python2.4, without all the necessary addons. See above.
Not so critical
- Reinstalling a modern java VM
- Installing ant
- Get the omath mailing lists working again.
- This is probably unlikely; I never documented what I was doing, and I remember there being quite a few configuration files that needed changes...
- Everthing is in /usr/local/mailman, I think?
- Make sure jira and fisheye start automatically.
- Check mysqladmin and webalizer still work? They should.