A lazy sysadmin is a good sysadmin. Time spent in finding
more-efficient shortcuts is time saved later on for that ongoing project of "reading the whole of the internet", so try Juliet Kemp's 10 handy tips to make your admin life easier...
1. Cache your password with ssh-agent
Using ssh keys and ssh-agent allows you to enter your password once and
have it cached for the rest of the session. Edit your .xsession file to have
this as its final line:
ssh-agent gnome-session
(or substitute the window manager of your choice for gnome-session). This
starts gnome-session via ssh-agent, so that any programs run from that
gnome-session will have access to the keys stored in ssh-agent.
Next, type:
ssh-add ~/.ssh/id_rsa
and enter your passphrase for that key. That's it! From now until the end of
the session, you can log in with that key without entering the passphrase.
(Click for larger)
2. Speed up logins using Kerberos
Another option to reduce login hassle is to install Kerberos.
As well as being significantly more secure than
other setups, Kerberos also works as single-sign-on. So
once you've logged in in the morning, you can SSH anywhere else in
your Kerberised network without having to type your passphrase at all.
Unfortunately, this is only lazy once it's actually going: setting it up in
the first place does require a certain amount of effort. The documentation is
excellent, though, and an increasingly large number of applications (eg Apache)
are now Kerberised. This should also reduce your security workload, as it
reduces the chances of a break-in.
3. screen: detach to avoid repeat logins
Entering:
screen -d
(detach screen), or hitting Ctrl-A then D whilst in a screen
session, will detach your running screen, leaving it going in the
background. This can be useful for running background jobs; it can also
be useful for saving you frp, repeatedly logging into and out of
machines. Start a screen session, log in to another machine, and detach
the screen. Reattach the screen:
screen -r
and you'll be back connected to the ssh session, without having to type your password again.
4. screen: connect multiple users
When you're bugfixing, sometimes you really need to be able to see what the
user who brought the problem to you is doing, and what output they're getting.
You can use screen to avoid having to actually walk all the way over to their
desk. Log on to the user's machine as them, then type:
screen -S debug
Then get them to type:
screen -x debug
and they'll join your screen session. Then whatever they do will be
replicated on your screen (and vice versa). ('debug' is just the identifier;
you can use any name you want.)
5. Expand Bash's tab completion
Aliases and functions are great for minimising the typing on anything you
need to do on a regular basis. You can save both a few characters, and the
hassle of the occasional mistype.
Improve bash a bit by installing /etc/bash_completion from
this site. You'll get tab-completion of hostnames, and a
"fuzzy matching" feature that will try to guess mistyped directory names and
incomplete directory names.
There are also shell shortcuts that can be useful to speed you up. Learn a
few of them - for example:
- Ctrl-A, Ctrl-E to jump to the start/end of the line.
- Ctrl-r to search your command-line history backwards.
- Alt-. to paste the last argument to the previous command (this is
incredibly useful).
- $ then tab twice to show all the available system variables.
Check the bash documentation for more.
6. Automate your installations
Installation of new machines is boring, time-consuming, and can be
error-prone. What you want is to set up your systems so that install is as
close to a one-button process as possible. You can use TFTP and DHCP to boot
off a networked image, or use FAI, which is an automated installation tool (similar to Kickstart for Redhat or JumpStart for Solaris).
With FAI, you can install machines unattended, and even set up different
installs based on different hardware provision and different requirements.
At the very least, it's worth
creating your own install image with the packages and setup that you need.
This is fairly straightforward with most distributions. Having standardised
machines not only speeds install up, it makes it less likely that you'll spend
time later on installing the things you forgot initially.
7. Roll out changes to multiple systems
The one-button install concept should extend to other aspects of your
systems, for much the same reasons. Puppet enables you to manage your systems
centrally - you change files or settings in the repository on the central
Puppet server, and they're rolled out automatically to all your Puppet
clients. You will still have to change things twice (once on a test machine
to make sure what you're doing, then once in the central Puppet repository),
but it'll save a lot of time and reduce mistakes. (Remember that it really
is important to test - Puppet also makes it really fast to propagate an error
across all your systems.)
8. Automate Debian updates
The package cron-apt can be set up to automatically install as well as
download updates, but that's not really a good idea. Better is to get it to
send you emails, check over what'll happen, and then dump all the emails in a
specific folder. Then write a script which parses this local
mailbox, grabs the machine names from the subject lines, and does the rest
from there. (Probably using that root ssh key and ssh-add, again.)
Click here for one script
that will do this (there are other options).
9. Sanely reboot a locked-up box
Rather than hitting the power button when something is dramatically
crashed, thus meaning that you have to wait though a fsck disk check when the system
boots back up, hold down Alt + Sysreq/Print Screen, and press REISUB in turn
(whilst continually holding down those keys). Magic: your machine will
reboot, after stopping all the programs and syncing and unmounting the
drives.
There are some other magic things that you can do with SysReq, as well -
it's worth looking up the documentation.
10. Send commands to several PCs
Not everything that you want to do on all machines will work well with Puppet,
- you might for example want to temporarily mount a particular disk on
all machines. ClusterSSH is great for this - it enables you to log onto
a number of machines at once, and issue the same command on all of them
simultaneously. Usefully, you can also click on a particular machine's screen
and issue a command just on that machine, in case one machine is misbehaving.
You can set up groups of machines, as well, so that you can log in
immediately to all your servers, or all your desktops. Combine this with a
root ssh key and ssh-agent, and save yourself both typing and time.
