OSSForks/opsschool-curriculum
1
0
Fork 0
mirror of https://github.com/opsschool/curriculum.git synced 2025-12-06 00:19:45 +01:00
Code Issues Actions Packages Projects Releases Wiki Activity

Mega trailing whitespace commit

Browse Source
This commit is contained in:
Mike Fiedler 2012-12-26 16:07:13 -05:00
parent 4696176936
commit 15ffe051d7
11 changed files with 58 additions and 58 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
capacity_planning.rst
View File

@ -7,7 +7,7 @@ this course)
Fundamentals of capacity planning
=================================
Resource usage investigation and exploration
Resource usage investigation and exploration
---------------------------------------------
* Examples: CPU:req/sec ratio, memory footprint:req/sec ratio, disk consumption
@ -19,7 +19,7 @@ Resource usage investigation and exploration
Finding ceilings
----------------
* Discovering resource limits
* Discovering resource limits
* Comparing different hardware/instance profiles - production load versus
synthetic
@ -30,7 +30,7 @@ Finding ceilings
uploads, caching storage+processing, etc.
* Architecture analysis (anticipating the next bottleneck)
Forecasting
Forecasting
============
Linear and nonlinear trending and forecasting ("steering by your wake")
@ -44,8 +44,8 @@ Seasonality and future events
* Organic growth approaches (bottom-up infra driven, top-down app driven)
* inorganic growth events (new feature launch, holiday effects, "going viral",
major public announcement)
* Provisioning effects on timelines, financial tradeoffs
major public announcement)
* Provisioning effects on timelines, financial tradeoffs
Diagonal scaling
================

2
careers.rst
View File

@ -121,7 +121,7 @@ servers, operating systems, databases, middleware and applications.
Systems Administrators can range from "jack of all trades" with knowledge of
multiple systems and platforms to specialists who focus on one system or
platform, for example Microsoft Windows or Linux.
platform, for example Microsoft Windows or Linux.
Whilst perhaps more "general" than some of the other specialist roles, Systems
Administrators tend to focus on managing individual hosts, usually desktops or

2
http_101.rst
View File

@ -5,7 +5,7 @@ Tools: Speaking http with telnet/netcat/curl
============================================
Understanding how a browser/tool parses a url and server receives it for a
request
request
Apache, nginx
=============

2
http_201.rst
View File

@ -31,7 +31,7 @@ Rails
Sinatra
-------
Python
Python
======
Django

2
labs.rst
View File

@ -240,5 +240,5 @@ Install Apache
Configure a virtual host
------------------------
Display a simple web page
Display a simple web page
--------------------------

62
loadbalancing_101.rst
View File

@ -9,14 +9,14 @@ Why do we use load balancers?
* They keep application servers busy by buffering responses and serving them to
slow clients (or keepalives). We want app servers to do real work, not waste
time waiting on the network.
* Load balancers provide a mechanism to verify the health of backend servers to
* Load balancers provide a mechanism to verify the health of backend servers to
ensure traffic is only routed to backends available to service the request.
These health checks can be simplistic ICMP pings to ensure host availability
to advanced HTTP (layer 7) health checks that use HTTP response codes or
to advanced HTTP (layer 7) health checks that use HTTP response codes or
content matching.
* Can provide a layer of abstraction such that end user endpoints remain
consistent (i.e URLs) as backend application infrastructure expands
(or contracts)
* Can provide a layer of abstraction such that end user endpoints remain
consistent (i.e URLs) as backend application infrastructure expands
(or contracts)
* Can be either software based (aka reverse proxies) or hardware based (physical
devices optimized for network throughput and ISO layer 2,3 and 4 routing).
* Can be tuned by itself for max conns?
@ -34,7 +34,7 @@ An application server behind a load balancer is typically referred to as a
Pools
-----
A collection of backends is referred to as a pool.
A collection of backends is referred to as a pool.
Balancing algorithms
--------------------
@ -42,7 +42,7 @@ Balancing algorithms
Session Affinity
----------------
Session affinity or session stickiness, is when the load balancer applies an
Session affinity or session stickiness, is when the load balancer applies an
algorithm to the incoming connection to direct it to a single server. This
is typically done for HTTP applications by setting cookies. In general TCP
applications often use IP addresses to determine the server to direct
@ -54,34 +54,34 @@ Local & ISP caching
SSL Termination and Offloading
------------------------------
SSL Termination is when the load balancer established an SSL connection
SSL Termination is when the load balancer established an SSL connection
between the client and the load balancer and a non-encrypted connection between
the load balancer and backend server.
Terminating SSL on the load balancer eliminates the need to distribute your
Terminating SSL on the load balancer eliminates the need to distribute your
certificate and key amongst all the servers. When using hardware load balancers
they typically have special hardware acceleration which is much more performant
compared to terminating connections on the backend server.
they typically have special hardware acceleration which is much more performant
compared to terminating connections on the backend server.
Balancing vs. Failover (Active / Passive)
-----------------------------------------
Often load balancing is used as a high-availability technique, by allowing
multiple backends to service a request if one node should become unavailable. It
differs from failover configuration because all nodes generally participate in
servicing clients. In failover configurations a single (active) node handles all
requests until an issue arises, and the secondary (passive) node takes over all
differs from failover configuration because all nodes generally participate in
servicing clients. In failover configurations a single (active) node handles all
requests until an issue arises, and the secondary (passive) node takes over all
of the incoming traffic. Failover configurations are usually not configured for
scaling purposes.
scaling purposes.
Health Checks
Health Checks
---------------
Most load balancers have some ability to test backend servers for availability,
these are called health checks. They can be as simple as whether hosts are
Most load balancers have some ability to test backend servers for availability,
these are called health checks. They can be as simple as whether hosts are
responding on the port the application is bound to, or complex configurations
that test special URIs, or response times. If a server fails a health check the
load balancer will temporarily remove the node from the pool, until it
load balancer will temporarily remove the node from the pool, until it
successfully passes the health check.
Non-HTTP use cases
@ -95,11 +95,11 @@ Software
Apache
------
Apache has the ability to load balance using ``mod_proxy_balancer``, and ``mod_jk``.
Apache has the ability to load balance using ``mod_proxy_balancer``, and ``mod_jk``.
Mod_poxy_balancer is purpose-built load balancing solution. It supports the HTTP, FTP,
and AJP protocols. There is basic support for sessions stickiness's, weighted round-robin,
and can remove unhealthy backends from the balancing pool. It lacks support for customizable
and AJP protocols. There is basic support for sessions stickiness's, weighted round-robin,
and can remove unhealthy backends from the balancing pool. It lacks support for customizable
health checks, other TCP protocols. Support for AJP is provided by ``mod_proxy_ajp`` and support
for FTP is provided by ``mod_proxy_ftp``.
@ -115,12 +115,12 @@ Nginx
HAProxy
-------
HAProxy is a general TCP load balancing server that is highly configurable. It
will generally support any TCP based protocol, and has special modes for HTTP,
RDP, MySQL, and Postgresql protocols. It has support for multiple types of
HAProxy is a general TCP load balancing server that is highly configurable. It
will generally support any TCP based protocol, and has special modes for HTTP,
RDP, MySQL, and Postgresql protocols. It has support for multiple types of
health check including URL based, traffic-based health, and external checks via
the ``httpchk`` options. It has several load balancing algorithms: round robin,
static round-robin, least connections, source hashing, URI hashing, URI
static round-robin, least connections, source hashing, URI hashing, URI
parameter, and RDP-cookie.
@ -130,10 +130,10 @@ Hardware
BIG-IP
------
BIG-IP has purpose-built hardware load balancers. They support protocols in layers
2, 4, and 7 of the OSI model. They allow for very complex configurations, and
BIG-IP has purpose-built hardware load balancers. They support protocols in layers
2, 4, and 7 of the OSI model. They allow for very complex configurations, and
support writing special TCL programs to modify the load balancing behavior. The
product supports SSL termination and offloading, with additional licensing.
product supports SSL termination and offloading, with additional licensing.
Netscaler
---------
@ -146,14 +146,14 @@ Anycast
DNS GSLB
--------
* A GSLB (Global Site Load Balancer) at the most simplistic level is a health
* A GSLB (Global Site Load Balancer) at the most simplistic level is a health
checking DNS server.
* Most often used to load balance between geographically dispersed data centers.
* Generally has health check mechanisms similar to load balancers which are used
to return an IP address (as part of the DNS lookup) of a host that is currently
available to service the request.
* Conceptually provides coarse-grained round robin and affinity balancing
algorithms by setting the time to live (TTL) of the DNS lookup for an
algorithms by setting the time to live (TTL) of the DNS lookup for an
appropriate duration.
CDN's

22
networking_101.rst
View File

@ -98,7 +98,7 @@ IPv4 uses a 32-bit address space most typically represented in 4 dotted decimal
each octet contains a value between 0-255, and is separated by a dot. An example
address is below:
10.199.0.5
10.199.0.5
There are several other representations, like dotted hexadecimal, dotted octal, hexadecimal,
decimal, and octal. These are infrequently used, and will be covered in later sections.
@ -120,7 +120,7 @@ Subnetting, netmasks and CIDR
A subnet is a logical division of an IP network, and allows the host system to identify which
other hosts can be reached on the local network. The host system determines
this by the application of a routing prefix. There are two typical representations of this
prefix: a netmask and CIDR.
prefix: a netmask and CIDR.
Netmasks typically appear in the dotted decimal notation, with values between 0-255 in each
octet. These are applied as bitmasks, and numbers at 255 mean that this host is not reachable.
@ -128,9 +128,9 @@ Netmask can also be referred to as a Subnet Mask and these terms are often used
example IP Address with a typical netmask is below:
============= ===============
IP Address Netmask
IP Address Netmask
============= ===============
192.168.1.1 255.255.255.0
192.168.1.1 255.255.255.0
============= ===============
CIDR notation is a two-digit representation of this routing prefix. Its value can range
@ -138,9 +138,9 @@ between 0 and 32. This representation is typically used for networking equipment
is the same example as above with CIDR notation:
============= ===============
IP Address CIDR
IP Address CIDR
============= ===============
192.168.1.1 /24
192.168.1.1 /24
============= ===============
Private address space (:rfc:`1918`)
@ -173,9 +173,9 @@ Practical networking
Cat5e, Cat6, Cat6a
------------------
Cat5e, Cat6, and Cat6a are all coper transport mediums. They use twisted pair
Cat5e, Cat6, and Cat6a are all coper transport mediums. They use twisted pair
wiring, relying on the twist with differential signaling to prevent noise. This is the most
common form of cabling for connecting computers in a network.
common form of cabling for connecting computers in a network.
Fiber
-----
@ -187,16 +187,16 @@ Multimode vs Single Mode vs OM{3,4}
Multimode fiber is a less expensive fiber optic cable, that is typically useable with lower
cost optical components. Depending on the application and bandwidth required, multimode fiber
can have a range up to 2000 meters, but as low as 33 meters. It is very common to see it
used for building backbones, and system to switch applications.
used for building backbones, and system to switch applications.
LC vs SC
^^^^^^^^
LC and SC connectors are the two most common type of fiber connectors.
LC and SC connectors are the two most common type of fiber connectors.
LC is also known as a Lucent Connector. They are typically used for high-density applications, and are
the type of connector used on SFPs or XFPs. Typically the connector is packaged in a duplex configuration
with each cable side by side.
with each cable side by side.
SC connectors are also know as Subscriber Connector, Square Connector, or Standard Connector. This is the type
of connector typically used in the telecom industry. They have a larger form factor than the LC connectors, and

2
programming_101.rst
View File

@ -37,7 +37,7 @@ I/O
Pipes
-----
stderr vs. stdout
stderr vs. stdout
------------------
/dev/null and /dev/zero

4
shells_101.rst
View File

@ -77,13 +77,13 @@ To list all of the shell's environment variables, use the ``env`` command: ::
PATH=/usr/local/sbin:/sbin:/bin:/usr/sbin:/usr/bin:/root/bin:/usr/local/bin
MAIL=/var/spool/mail/root
PWD=/root/curriculum
PS1=[\[\e[33;1m\]\t \[\e[31;1m\]\u\[\e[0m\]@\[\e[31;1m\]\h\[\e[0m\] \W\[\e[0m\]]#
PS1=[\[\e[33;1m\]\t \[\e[31;1m\]\u\[\e[0m\]@\[\e[31;1m\]\h\[\e[0m\] \W\[\e[0m\]]#
AWS_IAM_HOME=/opt/aws/apitools/iam
HISTCONTROL=ignoredups
SHLVL=1
SUDO_COMMAND=/bin/bash
HOME=/root
HISTTIMEFORMAT=[%Y-%m-%d %H:%M:%S]
HISTTIMEFORMAT=[%Y-%m-%d %H:%M:%S]
OLDPWD=/tmp
Global vs. User Profiles

2
soft_skills.rst
View File

@ -106,7 +106,7 @@ Business Savvy
Supporting business needs.
--------------------------
We exist professionally to support the needs of the organization.
We exist professionally to support the needs of the organization.
Developing the Trusted Advisor Relationship
-------------------------------------------

6
sysadmin_tools.rst
View File

@ -67,7 +67,7 @@ For example:
It is a reliable server which is close to the other servers Bob works on.
On ``server1``, but starts a multiplexer. The multiplexer gives Bob a regular
looking command prompt, and Bob continues his work.
If Bob's internet connection drops, he can reconnect to ``server1``, and then
re-attach to the multiplexer he started previously. His session is in the
same state he left it before being disconnected, and he can continue his
@ -188,7 +188,7 @@ If there is more than one session the output of ``tmux ls`` will look like this:
.. code-block:: bash
0: 3 windows (created Fri Nov 30 18:32:37 2012) [80x38]
4: 1 windows (created Sun Dec 2 17:44:15 2012) [150x39] (attached)
4: 1 windows (created Sun Dec 2 17:44:15 2012) [150x39] (attached)
You will then have to select the right session with the ``-t`` command line switch:
@ -202,7 +202,7 @@ you should only see one tmux process per user per system.
You should see the original session with the two shells again after
running this command.
tmux configuration
tmux configuration
~~~~~~~~~~~~~~~~~~~
``tmux`` is configured via a
config file which is usually called :file:`.tmux.conf` that should live in