CIT 470: Advanced Network and System Administration

CIT 470: Advanced Network and System Administration Slide #1

CIT 470: Advanced Network and System Administration

Administration Fundamentals


Topics

1. The Nature of System Administration

2. Organizations and Certifications

3. Change Management

4. Remote Administration

What is System Administration?

What is a system?

System: An organized collection of computers interacting with a group of users.


Servers

Services

PCs

Users

Network

run

on

run

on

help to accomplish work

System State

System policy: specification of a system’s configuration and its acceptable usage.

System state S(t): the current configuration (files, kernel, memory or CPU usage) of a system.

Ideal states S*(t): states of the system that match the system policy. Over time, the system state shifts away from the ideal state.

System administration: modifying the system to bring it closer to S*(t).



What do sysadmins do?

Small org: sysadmin can be entire IT staff– Phone support– Order and install software and hardware– Fix anything that breaks from phones to servers– Develop software

Large org: sysadmin is one of many IT staff– Specialists instead of “jack of all trades”– Database admin, Network admin, Fileserver admin, Help

desk worker, Programmers, Logistics


Common Activities

1. Add and remove users.

2. Add and remove hardware.

3. Perform backups.

4. Install new software systems.

5. Troubleshooting.

6. System monitoring.

7. Auditing security.

8. Help users.

9. Communicate.


User Management

Creating user accounts– Consistency requires automation– Startup (dot) files

Namespace management– Usernames and UIDs– Multiple namespaces or SSI?

Removing user accounts– Consistency requires automation– Many accounts across different systems


Hardware Management

Adding and removing hardware– Configuration, cabling, etc.

Purchase– Evaluate and purchase servers + other hardware

Capacity planning– How many servers? How much bandwidth, storage?

Data Center management– Power, racks, environment (cooling, fire alarm)

Virtualization– When can virtual servers be used vs. physical?


Backups

Backup strategy and policies– Scheduling: when and how often?– Capacity planning– Location: on-site vs. off-site.

Monitoring backups– Checking logs– Verifying media

Performing restores when requested


Software Installation

Automated consistent OS installs– Desktop vs. server OS image needs.

Installation of software– Purchase, find, or build custom software.

Managing software installations– Distributing software to multiple hosts.– Managing multiple versions of a software pkg.

Patching and updating software


Troubleshooting

Problem identification– By user notification– By log files or monitoring programs

Tracking and visibility– Ensure users know you’re working on problem– Provide an ETA if possible

Finding the root cause of problems– Provide temporary solution if necessary– Solve the root problem to permanently eliminate


System Monitoring

Automatically monitor systems for– Problems (disk full, error logs, security)– Performance (CPU, mem, disk, network)

Provides data for capacity planning– Determine need for resources– Establish case to bring to management


Helping Users

Request tracking system– Ensures that you don’t forget problems.– Ensures users know you’re working on their

problem; reduces interruptions, status queries.– Lets management know what you’ve done.

User documentation and training– Policies and procedures

Schedule and communicate downtimes

Communicate

Customers– Keep customer appraised of process.

• When you’ve started working on a request with ETA.• When you make progress, need feedback.• When you’re finished.

– Communicate system status.• Uptime, scheduled downtimes, failures.

– Meet regularly with customer managers.

Managers– Meet regularly with your manager.– Write weekly status reports.



Specialized SkillsHeterogeneous Environments

Integrating multiple-OSes, hardware types, or network protocols, distributed sites.

DatabasesSQL RDMS

NetworkingComplex routing, high speed networks, voice.

SecurityFirewalls, authentication, NIDS, cryptography.

StorageNAS, SANs, cloud storage.

Virtualization and Cloud ComputingVMware, cloud architectures.

Slide #17

Qualities of a Successful Sysadmin

Customer oriented– Ability to deal with interrupts, time pressure– Communication skills– Service provider, not system police

Technical knowledge– Hardware, network, and software knowledge– Proficiency with at least one scripting language– Debugging and troubleshooting skills

Time management– Automate everything possible.– Ability to prioritize tasks: urgency and importance.


First Steps to Better SA

Use a request system.– Customers know what you’re doing.– You know what you’re doing.

Manage quick requests right– Handle emergencies quickly.– Use request system to avoid interruptions.

Policies– How do people get help?– What is the scope of responsibility for SA team?– What is our definition of emergency?

Start every host in a known state.


Principles of SASimplicity

– Choose the simplest solution that solves the entire problem.– Work towards a predictable system.

Clarity– Choose a straightforward solution that’s easy to change, maintain,

debug, and explain to other SAs.Generality

– Choose reusable solutions that scale up; use open protocols.Automation

– Use software to replace human effort.Communication

– Be sure that you’re solving the right problems and that people know what you’re doing.

Basics First– Solve basic infrastructure problems before advanced ones.

Organizations and Certifications


Organizations

USENIX: Advanced Computing Systems Association

LISA: Large Installation System Administration

SAGE: System Administration Guild

LOPSA: League of Professional System Administrators


Types of Sites

Small2-10 computers, 1 OS, 2-20 users.

Small staff size requires outsourcing to obtain most specialized skills.

Midsized11-100 computers, 1-3 OSes, 21-100 users.

Large100+ computers, multiples OSes, 100+ users

Outsources to reduce costs, some specializations.


Certifications

• CCNA, CCNP, CCIE (Cisco)• cSAGE (SAGE)• MCSA (Microsoft)• RHCE (Red Hat)• SCSA (Sun)• VCP (VMware)


SAGE Job Descriptions

NoviceOS familiarity, help desk skills

JuniorCan use OS system administration tools (370)

IntermediateUnderstanding of distributed computing, common servers,

automate small tasks, independent actionSenior

Understanding of scaling issues, including capacity planning, solve problems by addressing root cause, higher level programming abilities, write proposals for purchasing, data center planning, etc.

SA Maturity Model (SAMM)1. Ad Hoc

Ad-hoc non-repeatable solutions, firefighting.

2. RepeatableSome repeatable processes.

3. DefinedDocumented standard processes

4. ManagedProcess effectiveness measured, adapted.

5. OptimizedCIT 470: Advanced Network and System Administration Slide #25

Maturity and Complexity


Constant firefighting, high downtime

Complexity: increasing numbers of systems and/or services

Mat

uri

ty

Scalable but time lost in process.

Low downtime, high efficiency

Works, but hardto scale up.

Tool Maturity Levels1. Ad Hoc

OS GUI, CLI, or web administration interfaces.

2. RepeatableVersion control (RCS, SVN, GIT), request tracker

3. DefinedAutomatic monitoring (Nagios, monit, god)

4. ManagedConfiguration management (AutomateIt, cfengine)

5. Optimized


SAGE Code of Ethics

• Professionalism• Personal Integrity• Privacy• Laws and Policies• Communication• System Integrity• Education• Social Responsibility


http://www.sage.org/ethics/

Terry Childs Case

Network administrator for San Francisco– CCIE who built city’s FiberWAN network

Terry was only person with router passwords– IT department acknowledges knowing that– He was on-call 24x7x365 to resolve issues

Terry refused to give passwords to boss– Cited fears that they would be misused by

management, outside contractors.

What was the right thing for Terry to do?CIT 470: Advanced Network and System Administration Slide #29

Change Management


Change Management

Effective planning and implementation of changes to systems.

Changes should be1. Well documented.

2. Have a backout plan.

3. Reproducible.


Why do we need Change Management?

March 26-29, 2006: BART trains halted to avoid running into each other when computer systems crashed.• Crashes on Monday/Tuesday resulted from

software maintenance upgrades.• Crash on Wednesday resulted from installing a

backup system to avoid future crashes.• Thousands of passengers stranded for several

hours each time.


Change Management

1. Plan change.

2. Test change on single system.

3. Test change on multiple systems.

4. File a change request.

5. Change committee approves request.

6. Schedule change.

7. Communication with users/admins.

8. Change systems at scheduled time.

9. Post-event analysis.


Testing Changes

• Automated checks.– Sanity checks like Samba testparm.– Reboot system.

• Test on one system first.• Then test on set of systems.

– Dedicated test systems.– System admin workstations.– Virtual machines.


When do you need a Change Proposal?

Does the change impact critical services?

Critical machines/services– Business critical: e-commerce server, etc.– Essential services: routers, DNS, NFS, auth.

Non-critical machines/services– Individual desktops– Internal news web server


Change Proposal

1. Description of the change.

2. Systems impacted by change.

3. Why the change is being made.

4. Risks presented by the change.

5. Test procedure.

6. Backout plans.

7. How long the change will require.


CommunicationCommunicate change to impacted people.

– What change is being made (nontechnical.)– Which services will be unavailable.– When and how long will they be unavailable.– What actions do they need to task (if any.)

Communication issues– If you send too many notes, they’ll be ignored.– Send notices only to those impacted.– Push critical notices; use pull for non-critical.


Scheduling

Scope When Notification Type

Routine Single host or user.

Anytime. Personal.

Major Many hosts or users.

Off-peak Push.

Sensitive None but major impact on failure.

Off-peak. Pull.


Change Freezes

Time when only minor updates can be done.– End of quarter or year.– “Crunch time” for projects.


Backing Out

Decide back-out conditions before downtime– Avoid the “just 5 more minutes” problem.– Be sure that someone is keeping track of time.

Questions:– How much time is required for back out?– When is the latest time you can successfully back out?– Will backing out this change prevent other changes from

being committed?


Backing Out: How to do it?

Service-level changesUse revision control system to revert config.

Machine-level changesSoft cutover: Old service is still running.

Hard cutover: Power up old server or restore from backups.

SnapshotsSnapshot VM before making change.

Revert to snapshot if need to backout.

IssuesData migration and compatibility


Automatic Checks

Check integrity of critical files before use.– Some services provide checks: LDAP, SMB.– Check startup files by rebooting machine.– Write your own checks for other files.

• Most people only do this after they have a problem.

Remote Administration

1. Network Access

2. SSH

3. Key-based Authentication

4. Console Access

5. X-Windows

6. VNC and NX

7. SSH tunneling


Network Access

Most tasks can be done from the shell.File management.

Disk/volume management.

Troubleshooting and viewing logs.

Installing/removing software.

Start/stop network services.

Reboot/shutdown.

All we need is a way to invoke a shell across the network.


telnet

Ubiquitous network terminal protocoltelnet hostname

Similar protocolsrlogin –l user hostnamersh –l user hostname command

InsecureData, including passwords, sent in the clear.

rlogin/rsh use ~/.rhosts for access w/o passwords.


ssh

Secure Shell

Replacestelnetftprloginrshrcp


SSH Security Features


SSH: Protocols and Products

• OpenSSH• SSH Tectia• F-secure SSH• Putty• WinSCP

• SSH v1– Insecure, obsolete.– Do not use.

• SSH v2– Current version.


SSH Features

Secure loginssh –l user host

Secure remote command executionssh –l user host command

Secure file transfersftp –l user hostscp file user@host:/tmp/myfile

Port forwardingssh –L 110:localhost:110 mailhost


The Problem of Passwords

1. Good passwords are hard to remember.

2. Password transferred to remote system.

3. Automating remote access with passwords is difficult.


Public Key Cryptography

Two keys– Private key known only to owner.– Public key available to anyone.

Applications– Confidentiality:

• Sender enciphers using recipient’s public key, • Receiver deciphers using their private key.

– Integrity/authentication: • Sender enciphers using own private key, • Recipient deciphers using sender’s public key.


Key-based Authentication

SSH uses public-key authenticationPrivate key stored in your machine.

Public key stored on remote machines.

Public-key login protocol1. Client sends server a login request.

2. Server issues a challenge.

3. Client responds with computation based on challenge and private key.

4. Server checks response with public key.


Using key-based authentication

1. Generate a public/private key pair.ssh-keygenEncrypted key files: id_rsa, id_rsa.pub

2. Copy public key to remote hostCopy to ~/.ssh/authorized_keys.

3. Login to remote hostssh –l user remote


Keys are more secure than Passwords

1. Need to have two items to login: key file and passphrase.

2. Neither key nor passphrase is sent to remote host.

3. Machine-generated cryptographic keys are infeasible to guess, unlike passwords.


SSH Agent

Problem: you have to enter passphrase to decrypt the key each time you use ssh.

Solution: SSH Agent> ssh-agent $SHELL > ssh-add Enter passphrase for /home/jw/.ssh/id_dsa: ********

Identity added: /home/you/.ssh/id_dsa (/home/jw/.ssh/id_dsa)

> ssh –l jw host


SSH Agent Features

Agent support for entire session.Start ssh-agent on initial shell.X: ~/.xsession (Often enabled by default.)

Multiple keysssh-add keyfilessh-add –l

Remove keysssh-add –d keyfilessh-add –D


Remote Access when Server is Down

Problem: No network access to host.

Solutions:– Go to computer room and bring host up.– Specialized hardware (network boot / power).– Virtual machines.– Console servers.


Console Servers

Console– Main I/O device for computer.– Historically: serial terminal.– Typically: keyboard/mouse/screen.

Server allows access to multiple consoles.– Console access: BIOS, Bootloader, Kernel– Eliminates need for keyboards, mice, monitors.– Serial line to each machine from server.– One user has R/W, other users have R access.

http://en.wikipedia.org/wiki/File:Terminal-dec-vt100.jpg


Console Hardware

Console servers solution– Commercial: Cisco, Cyclades,

Xyplex– Open source: Conserver +

serial expander card

Hardware issues– Connectors: DB-9, DB-25,

RJ-45– Encoding: 8N1, 7E1– Speeds: 9600 – 230k


X-Windows

Server– Handles user input and

graphical display.– Runs on the machine

with display unit.

Clients (applications)– Can run on a different

machine than server.• Set DISPLAY env var.• Use –display option.

http://en.wikipedia.org/wiki/File:Network_Computing_Devices_NCD-88k_X_terminal.jpg

http://upload.wikimedia.org/wikipedia/commons/b/b5/X_client_sever_example.svg


Window Manager

X client that provides features like:– Move, resize, iconify, and kill windows.– Window title bars.– Popup menus.

Example window managers– twm: Tab, primitive early window manager– mwm: Motif, found on commercial UNIXes– fvwm: Free, fast, very customizable.– WindowMaker: NeXT-like, see also AfterStep.


TWM Screenshot


FVWM Screenshot


WindowMaker


Desktops

CDECommon desktop env for commercial UNIXes.

GnomeStandard Linux desktop based on GTK+.

KDEWindows-like free desktop based on QT.

XfceLightweight desktop, also based on GTK+.


VNC: Virtual Network Computing


Why VNC?

1. Remote desktop access.

2. Helpdesk: control a remote desktop.

3. Persistent desktop.

4. Use same desktop from multiple clients.

5. Need Linux access from Windows.

6. Need Windows access from Linux.


What is VNC?

• Open remote desktop protocol.• Many implementations

– RealVNC: VNC from original researchers.– TightVNC: VNC with high compression.– VNCj: Java VNC, can run within web browser.– PalmVNC: VNC for Palm Pilots.– UltraVNC: enhanced VNC, only for Windows.


Using VNC

1. Start VNC server

UNIX: vncserver

Win: Start menu>Programs>RealVNC>VNCServer

2. Write down server name and display number.

It will look something like unix3:1

3. Start VNC client

UNIX: vncviewer

Win: Start menu>Programs>RealVNC>VNCViewer

4. Enter server and display to connect to (from step 2).

5. A VNC remote desktop should appear.


Configuring and Troubleshooting

• On UNIX, VNC stores files under ~/.vnc• Configuration: xstartup

– Indicates which X clients to start with server.– Typically includes vncconfig application.

• Configuration: passwd– Contains VNC server session password.

• Log files: host:display#.log– Any errors should appear in these logs.


Securing VNCVNC does not provide encryption.

Use ssh tunneling to encrypt login + data:ssh –L 5901:remotehost:5901 remotehostvncviewer localhost:1


Tunneling

Tunneling: Encapsulation of one network protocol in another protocol– Carrier Protocol: protocol used by network

through which the information is travelling– Encapsulating Protocol: protocol (GRE, IPsec,

L2TP) that is wrapped around original data– Passenger Protocol: protocol that carries original

data


ssh Tunneling

SSH can tunnel TCP connections– Carrier Protocol: IP– Encapsulating Protocol: ssh– Passenger Protocol: TCP on a specific port

POP-3 forwardingssh -L 110:pop3host:110 -l user pop3host

– Uses ssh to login to pop3host as user– Creates tunnel from port 110 (leftmost port #) on localhost to port 110 (rightmost post #)of pop3host

– User configures mail client to use localhost as POP3 server, then proceeds as normal


NX

Advantages over VNC:Speed: fast enough to use over dialup.

Built-in ssh encryption.

DisadvantagesImmature code; hard to install + set up.

GPL client/server for Linux only.

Free Windows client; commercial server.

Key PointsPrinciples of System Administration

– Simplicity– Clarity– Generality– Automation– Communication– Basics First

Changes should be– Documented– Reproducible (usually automated)– Have a backout plan (use VM snapshots and version control)– Approved with a formal process

Remote Administration includes– ssh (default)– console (for when network is down)– remote desktop (rarely needed)

Tunneling via ssh can secure insecure protocols.



References1. Daniel J. Barrett, Robert G. Byrnes, Richard E. Silverman, SSH, The Secure Shell, 2nd edition, O’Reilly, 2005.2. Mark Burgess, Principles of Network and System Administration, 2nd edition, Wiley, 2004.3. Conserver, http://www.conserver.com/4. John Fisher, “Secure X Windows,” CIAC 2316, http://www.ciac.org/ciac/documents/ciac2316.html, 1995.5. Aeleen Frisch, Essential System Administration, 3rd edition, O’Reilly, 2002.6. David K.Z. Harris, “Zonker’s Greater Scroll of Console Knowledge,” http://www.conserver.com/consoles/, 2005.7. Brian Hatch, “SSH Host Key Protection,” http://www.securityfocus.com/infocus/1806, 2004.8. Thomas A. Limoncelli and Christine Hogan, The Practice of System and Network Administration, Addison-Wesley,

2002.9. No Machine NX, http://www.nomachine.com/10. OpenSSH, http://www.openssh.com/11. Real VNC, http://www.realvnc.com/12. Daniel Robbins, “OpenSSH key management,” http://www-128.ibm.com/developerworks/library/l-keyc.html, 2001.13. SAGE, Job Descriptions, http://www.sage.org/field/jobs-descriptions.html.14. SAGE, SAGE Code of Ethics, http://www.sage.org/ethics/.15. Carla Schroeder, Linux Cookbook, O’Reilly, 2004.16. Carla Schroeder, “FreeNX ups the Remote Linux Desktop Ante,” Enterprise Networking Planet,

http://www.enterprisenetworkingplanet.com/netos/print.php/3508951, 2005.17. Glen Turner, “Linux Remote Serial Console HOWTO,”

http://www.tldp.org/HOWTO/Remote-Serial-Console-HOWTO/index.html, 2003.18. Paul Venezia, Why San Francisco's network admin went rogue,

http://www.infoworld.com/d/adventures-in-it/why-san-franciscos-network-admin-went-rogue-286?page=0,1, July 18, 2008.

19. Webmin, http://www.webmin.com/20. Window Managers for X, http://xwinman.org/21. Todd R. Weiss, “IT upgrades slow BART trains in San Francisco,”

http://www.computerworld.com/printthis/2006/0,4814,110107,00.html, ComputerWorld, March 31, 2006.

http://www.conserver.com/

http://www.ciac.org/ciac/documents/ciac2316.html

http://www.conserver.com/consoles/

http://www.securityfocus.com/infocus/1806

http://www.openssh.com/

http://www.realvnc.com/

http://www-128.ibm.com/developerworks/library/l-keyc.html

http://www.sage.org/ethics/

http://www.enterprisenetworkingplanet.com/netos/print.php/3508951

http://www.tldp.org/HOWTO/Remote-Serial-Console-HOWTO/index.html

http://www.infoworld.com/d/adventures-in-it/why-san-franciscos-network-admin-went-rogue-286?page=0,1

http://www.webmin.com/

http://www.computerworld.com/printthis/2006/0,4814,110107,00.html

CIT 470: Advanced Network and System Administration

Documents

system administrationslide

system monitoring

advanced network

system state st

system state shifts

network admin

software pkg

custom software