List
of Commands
Ping “ip address/name”
Ipconfig
Machine
network details
·
/flushdns
= flush cache
·
/displaydns
= shows cache details
·
/registerdns
= refresh dhcp leases/dns related details
·
/release,
/renew = ip address renewal
Routing
·
Netstat
–r = show routing table
·
Pathping
ip address = route info to target (cross between ping and tracert)
·
Tracert
“ip address/name” = shows route to target
DNS
Previous
versions of DNS were separate from DHCP = static hosts file list. Can now
integrate DNS function with DHCP = Dynamic DNS.
Root
servers. Are used when local DNS cannot resolve query.
It starts a hierarchical search.
Forwarding.
Can specify a forwarding DNS server to use to solve DNS
queries. Saves on bandwidth and normally quicker.
Use Forwarding server FIRST, and if it cannot solve query, it will then perform
hierarchical search via root servers.
DNS sever
contains data for a domain. This is called a zone. If the DNS server has ALL the
domain data it is Authoritative for that zone. Secondary severs are used for
redundancy and load balancing. It performs a zone transfer. It reloads data
when zone data changes.
Recursion
= DNS server queries other servers on behalf of the client.
Iteration
= Client queries other servers directly.
If it is
not a domain on an intranet, but a domain on the internet, you need to register
the domain with an authority such as InterNIC.
·
A
Record = specifies IP address for a name/record.
·
Cname
= links an alias to actual domain name. ?
·
PTR
= used for reverse zones. If supply an IP address it will return the domain
name.
*Static/dynamic zones. If you need to
manually update/change a dynamic zone you MUST stop it first or there will be
replication errors. Might be worth having a static zone AND a dynamic zone in
such cases.
Nslookup
App for
examining DNS servers
·
Ls
–d domainname = zone transfer data = see if dns responds to client.
·
Server
ip address/name = direct queries to the specified server from this point.
·
(type
server to confirm default/new server)
If you
type ‘set type=any followed by domain = dns config data ie ttl, wins, # of dns
servers, etc.
By default
it looks at first dns server via its ip address, so reverse lookup zone is
used. If no reverse lookup zone is used = problems. Should at
least have rvuz contain dns server addresses.
Can often
get ‘query refused’ = lessen Zone Transfer security for testing.
Ntdsutil
AD cleanup
(config/seize FSMO roles, clean up metadata left behind by DC’s not removed
cleanly from network. Is a hierarchical app, use Help to see list of commands
at each step, and use Quit to go back a step.
ANOTHER
way to use NSLOOKUP
1. nslookup
2. Use current DNS server or change
server using ‘SERVER servername’ command.
3. 127.0.0.1
4. If working correctly it should return
the name/ip of server queried AND say Local Host ans 127.0.0.1
SIMPLE
METHOD to Use NSLOOKUP
In fact, using this second method lead me to a simple way of using
Nslookup. Invoke
Nslookup, either use default DNS server or choose another if you want to
examine a specific server (server ‘servername’). Then just type in a pc name
(forward lookup) or IP address (reverse lookup) and check to see if it is
resolved.
Can even check names not in local domain to see if they are cached.
Net
·
Netdiag
= Detailed network config data, inc: pc name, domain,
list of hotfixes, ipconfig data, afterwhich a series of tests are performed.
inc, netbt, wins, domain membership, loopback, default gateway, winsock, dns,
browser, dc discovery, ldap, kerberos, bindings, remote access sessions, etc
(Servers ONLY)
·
Netstat
= shows connections to/from pc (detailed)
·
Netstat
–e = shows packets sent/received/discarded and ERRORS!
·
Net
config workstation = shows basic info, inc: domain, pc
name, user name, os installed, etc.
·
Net
config server = similar to above, but gives option to HIDE pc from browse list
(not recommended).
·
Net
file = Shows list of files CURRENTLY open and SHARED/connected to over network.
Also shows user connected to relevant file.
·
Net
share = shows all SHARES on pc, whether currently connected to or not. They are
listed by SHARENAME, and also shows resource details
(filename/disk) and remarks (offline cache enabled? Default share? Etc)
·
Net
view = browse list.
·
Net
view ip address/pc name = shows shares on that pc.
·
Net
group = shows all groups in domain. Does not show OU’s. (on
DC’s only).
·
Net
help OR/AND Net help ‘cmd’ = net help shows all cmd’s in net app, and net help
shows specific details for each cmd.
·
Net
name = the name of the current pc and user
·
Net
send (ip address, user name, /DOMAIN) MESSAGE = send message to one of the
options. The /domain switch is useful as it sends a message to ALL pc’s/users
in domain, instantly. Sometimes the /DOMAIN switch doesn’t broadcast to all
pc’s on subnet – a surefire way to successfully broadcast to subnet is NET SEND
* MESSAGE HERE
·
Net
time = app to config clock. Ideally, PDC synchs with itself or internet time
source, BDC off the PDC (default), and clients off DC’s (default).
·
Net
use \\servername\sharename \ = maps
shared network drive. But I cannot get it to represent itself in os, so I’d
rather map via the os gui and not the cmd line. The os
gui mapping shows itself in ‘my computer’, and I can choose it to be
persistent, or disconnect via the pull down menu (can also see all mappings via
the pull down windows explorer menu too)
Got it to work! net use driveletteryouwantlocal: \\servername\sharename Use quotes
from \\ to end of sharename if any part of the path has a space. Use net use on it’s own to view network
connections on your pc. Use net view
pcname to see shares on that pc that you can use net view with. Use net use driveletterspecifiedforconnection: /d to
·
Net
user = shows all user accounts on that machine. Seems to show accounts that ORIGINATED
on that machine, and not domain acc’s. If I want to see domain acc’s, I should
run net user on a DC. Clients show local acc’s such as administrator acc for
that machine, and not the domain administrator account.
·
Netdom
query fsmo = shows which DCs hold which fsmo roles. Useful tool, but only runs
on DCs themselves.
System
Restore Point
XP lets
you go back to a restore point before significant changes were made, a bit like
‘undo’. To create a SR point: Help and Support Centre -> Undo Changes to your
Computer with System Restore. Can create/restore from here.
Note: SR does not uninstall a
program. You still need to manually uninstall troublesome application.
C$,
Admin$ and Print$ Shares
$ = indicates
a hidden share. If admin types location of this share, ie, \\servername\admin$ or \\servername\C$
he will be taken to relevant dir.
Only
admins can access these shares.
·
C$
= Root of C:
·
Admin$
= Gives access to system root hierarchy over nw.
·
Print$
= Used to remotely administer printers. When tested personally, I found this
share to be unuseful – it has dll’s etc, - not much that can be config easily.
There is another printer share for each machine, which is not hidden – much
more useful to use! Just drill down to machine over nw,
ot type \\servername and public ‘PRINTERS’ share is there. In it is each
printer attached to that machine, you can add new
ones, change props, etc.
These shares
can be deleted, but are recreated on next reboot.
Dcdiag
(Domain Controller Diagnostic)
Dcdiag is
a cmd line tool which analysis the state of a dc. It performs a battery of
useful tests which highlight any problems. Useful for tracking down AD replication,
sysvol replication, group policy, FSMO, etc problems.
Dcdiag.exe
is downloadable from MS. Also included with some SP’s.
Different
version available – some more exhaustive ethan others,
ie, 2003 has more tests than 2000.
Troubleshooting
·
Perfmon
= invokes gui app. Add counters etc to test hatdware
resources.
·
Msconfig
= invokes gui app. Can launch sys restore.
Select different start up modes, config BOOT.INI, and select different services
to start up at boot. When turning off services with msconfig, they are not
truly disabled, but rather it is a temp measure/test for safety.. To turn them off properly, use SERVICES. (XP/98 only)
·
Chkdsk
= check disk. /F (fixes errors), /R (find bad sectors and recover data). Some
error messages seem to not be present when pc is not booted into a domain
environment (handles?)
Policies
Gpedit.msc
Group policy editor. Enables the policy the current user/machine are subject to, to be changed. It will propagate to ou,
group etc. In a workgroup environment however, you need to implement group
policies on each machine individually. (It is for this reason we use AD!)
Secpol.msc
Local SECURITY policy editor. This enables the security settings to be set locally. The
settings are the same as those found in the group policy, but are only a subset
– the security settings only.
Secedit
Every
60-90 minutes machines query DC's for updates. To force this:
·
secedit /refreshpolicy machine_policy
·
secedit /refreshpolicy user_policy
(DC’s =
refresh pol every 5 mins. Clients = refresh pol every 60-90 mins)
When
updating the policy, it updates the LOCAL policy with the current policy held
on the DC.
Can use
/enforce switch to ensure machine updates all policies regardless if they are
different to updated version on DC.
The above
is for win2k machines. For win2k3/XP: Use gpupdate. It does both machine
and user polices. Can use /force with same effect as /enforce on w2k machines.
Gpupdate
Same as
the now obsolete secedit /refreshpolicy cmd, it refreshes policy settings that have
changed since last refresh. If you use the /force switch all policies (user and
machine) are refreshed, regardless if they have changed or not since last
refresh. XP ONLY. When Secedit is run on XP machines,
it starts an app that lets you config templates, security settings, etc – but
it does NOT refresh policy settings.
Policy
Replication and DC to DC Replication (Different)
The above
is to configure clients refreshing their policies from the DC. To configure
DC's replicating to EACH OTHER:
·
AD
Sites & Services - Sites - Your Site - Your Server - ntds - Properties of
Connection. Can configure replication to what you want.
BUT the
default value for DC's replicating within a site is 5 minutes. But the default
shown above is once every hour.
UPDATE. Replication takes place at
several levels:
1. Within a
site. (automatic, every 5 minutes)
2. Intrasite. (using AD, sites & services as outlined above)
3. Between
Sites
Gpresult
See what
policies the user and machine are subject to. Also shows which DC the policy
came from in a multi DC environment.
Shutdown.exe
Shutdown,
but it needs to be placed into sysvol to work (win 2k, XP seems to have it
there by default). Can be scheduled so pc’s shut down at a
given time.
F8
during boot
Recovery/boot
options
TEMP
Files
Emptying
IE Temp Folder
Temp Internet Files
Cache of IE.
(speed up browsing, allow offline browsing, but lack
of privacy).
Note: a file called index.dat CANNOT
be manually
Location:
·
In
XP/2K: C:\Documents and Settings\<username>\Local
Settings\Temporary Internet Files\
In
98/NT/ME:
·
C:\Windows\Temporary
Internet Files\
C:\Windows\Profiles\<username>\Temporary
Internet Files\ (if using PROFILES)
To
Cookies:
·
XP/2K:
C:\Documents and Settings\<username>\Cookies\
98/NT/ME:
·
C:\Windows\Cookies\
C:\Windows\Profiles\<username>\Cookies\
(if using PROFILES)
History:
There are
several historys, inc, sites visited, drop down
address bar, autocomplete, windows search, etc. Hard to manually
To
reliably
Windows
TEMP folder:
Windows
uses a TEMP folder to store tempt data (from proc such
as indexing, installations, etc). To empty this, type %TEMP% to
open folder, and
Scattered
Temp Files:
Windows
scatters temp files elsewhere too. Find them using search -> *.tmp,
*.chk, ~*.* Once all these files are found
Concept
Reminders
·
Workgroup
vs Ad
·
Search
AD or Browse entire network
·
Applications
Invoke
certain apps via cmd:
Explorer
Windows
explorer
Active
Directory Domains & Trusts
Change
mixed/native mode, and other domains that trust (to/from/transitive).
Active
Directory Sites & Services
Ntds,
global catalogue, see which servers current server replicates from, change
replication schedule, define sites in the site and
their replication link schedule.
Active
Directory Users & Computers = centralise dir of all domain resources. Can impose
policies to impose security, structure it via groups/ou’s,
etc. Printers/published folders are also in AD but not apparent via ADUC
application.
Publish
a Shared Folder
Firstly,
make and share a folder. Then go to ADUC and create new OU, the ‘new – shared
folder’ To get to it = drill down: entire nw –
entire contents – directory – choose ou and drill down. Search: entire nw – entire contents – rc domain – find – shared folders
Delete
explicitly from AD when you want to get rid of it – don’t just
Component
Services
Contains 3
parts: Component Services, Event Viewer, Services.
Event Viewer and Services are as per normal. (Event Viewer
and Services also in Admin Tools as separate apps.)
Computer
Management
Contains 3
parts: System Tools, Storage, Services &
Applications. Can also send console message via Services – rc
– all tasks.
Configure
your Server
Wizards
for AD, File Server, Print Server, Web Server, Networking, Application Server, etc.
DHCP
Configure
pool of ip addresses.
Distributed
File System
Logically looks
like 1 tree of folders – but can be scattered across different servers. Can
take server off line, point dfs to new server, and users need never know. Good
for load balancing.
2 types:
Stand Alone or Domain Based.
·
Dfs
host server is (if) AD integrated, therefore dfs topology is synchronised
across all host servers.
·
Dfs
links are added to root.
·
Dfs
links can refer to shared folders.
·
DFS
Root = Specify a shared folder as dfs root. Root is to be hosted on a member
server in a domain, and replicated via AD. For 2 replicated roots = 2 shared
(root) folders.
·
DFS
Link = Refers to element under the root. Can be on different server to root. If
it links to a folder it can have further folders within it. IT is this LINK
NAME that users see when they navigate the DFS hierarchy, and not the names of
the members of the replica set.
·
DFS
Shared Folders = First folder a link points to is linked via dfs. Further
folders within this set, from the same link, are added via the consoles New
Shared Folder dialogue box. To have folders replicate to each other: make new
empty shared folder, and in dfs add new replica. 2 or more folders which
repliate to each other are a REPLICA SET.
·
DFS
Client = Refers to dfs root/link. 9.x = download. 2k =
inc.
·
Check
Replication = To replicate root to other member servers: DFS Root – >New
Root Share
·
Can
publish the root into AD to make it viewable via ad.
·
Can
map to dfs share.
·
Can
cmd connect to dfs share ie \\servername\dfsshare
(explorer).
·
Can
view dfs tree via windows explorer: ‘explorer’ at cmd
The
resulting logical tree diagram is the logical dfs view of network shares. The
scattered and replicated nature is hidden.
DNS
Allows client to map a dns name to an ip address thus making network
naming more ‘human’.
·
A
record = maps host to ip. (A record is same as Host record)
·
Cname/canonical
= makes 1 domain an alias of another. It gets all the subdomains and dns
records of the original. Cname and Alias are same. Used when you want a host to
respond to more than 1 name. Record will show the alias name and the name of
the machine it refers to.
·
Ptr
record = maps the hostname to the canonical name for that host. Used in reverse
look up zones instead of Host/A records.
·
Name
server record = maps a domain name to a list of dns servers for that domain.
Delegations use this. Ip addresses aren’t used here, rather names are. You
enter the name of the domain down one level in the hierarchy. If same domain,
the default values of ‘same as parent folder’ is auto used.
·
Soa
record = Is the record of the PRIMARY dns server for
that domain, not secondary.
·
Authoritative
servers = are the servers referred TO by the dns servers up one level in the
hierarchy. Provides dns server that is authoritative for that domain.
·
Srv
record = Is a record of services on servers in the domain, ie, it is used to
refer to web servers, ftp servers, etc,
Using
NSLookup to query DNS
See above in CMD’s
Event
Viewer
Errors, warnings, etc. Use www to solve.
Internet
Services Manager
Inetpub
-> wwwroot -> publish as ‘Default’ = simple intranet site (indepth!)
Licensing
Licensing
details are not replicated among all servers, rather they replicate UPWARDS to
the enterprise server, which is the licensing server by default.
Any info
on any other server shows local info only – to get complete picture use the
licensing (enterprise) server.
Per seat
licenses can be entered/bought on any server running the licensing app, but
will only show up on the enterprise/licensing server. It is on this pc that the per seat
The per
server licenses can also be entered on any server running the licensing app,
but due to nature of server CAL’s you can choose which server to view/config
the licenses for from any server. Just drill down to relevant
server in licensing – server browser.
Gui is
confusing. Remember: Purchase History viewed on enterprise/licensing server.
Products View shows local Server/Seat/Backoffice, but view it on
enterprise/license server to see whole domain licensing. Server Browser can be
viewed on any server - just pick server in app to view and config. Can
add/remove licenses here and change mode (server or seat).
Local
Security Policy
Local
security shows the result of domain + local policies. Domain policies will over
ride local so there is a resulting ‘effective’ setting. Exact order: Local,
Domain, OU, OU’s inside OU’s.
Local
security is a subset of a policy (Computer Config -> Windows Settings ->
Security Settings).
Performance
Performance
in Admin Tools and Perfmon are the same app! Can add specific counters (there
are various entries for items such as Processor, Memory, Hard Disk, etc) Data
can be viewed in real time as a graph, histogram, etc. And can be saved as a
log and viewed later for comparison.
Logs of
pc/nw activity can be scheduled, used for baseline measurements, find
bottlenecks, etc.
To create
a log: Counter logs -> right hand pane -> New log settings
: name, etc.
To view a
precreated log: Sys monitor -> View log file data -> select relevant .blg
file -> Step2. Select counters. The log is created by preset counters, but
to view it you need to select these counters again. Why? You could create a log
with loads of counters, but might want to only view certain elements = filter
the activity. The list of counters to use are taken from the precreated log –
so there is no danger of you adding counters that were never initially used
when making the trace.
To match
the log with the precreated log settings the name of the Log given in ‘To
create a log’ will
also have a ‘Log file name’. This is what you look for when viewing a
precreated log.
Routing
and Remote Access
You config the server at several ‘levels’ in RRAS.
The Server
level: Props -> Router (Lan routing only, Lan and
Demand Dial routing), Remote Access Server. Can choose
authentication method. In IP you config how to use IP
addresses. If DHCP = the RRAS ‘takes’ IP add’s from the dhcp server and
allocates them. If Static = you config a range of IP add’s the RRAS ITSELF
allocates to clients, rather than the DHCP server. If you want a user to use 1
specific IP add you need to = 1. Config remote access policies to allow users
to request 1 IP add. 2. Config the dial in props on the user acc with 1 static
IP add (not if domain is mixed mode). PPP = compression
and multi link connections. Event Logging = config event viewer logging and
RRAS own logging in \systemroot\tracing folder.
Routing
Interfaces: Here you add/
Ports:
Config wanmini ports (pptp, l2tp) to allow vpn connections. Will also list dial
up connections if they’re present. Config info inc:
how many of each type of port there are, and what type of request they should
handle, ie, remote access inbound, inbound or outbound for demand dial, device
phone no, etc
Number pf ports corresponds to number of connections. Can in/decrease the number of
ports available, therefore the number of connections available.
IP Routing
= has several sections.
·
General
= Indepth - readup. Briefly…Add new router interfaces and protocols. For each
interface you can config routes, show routing tables, ip details, connections,
multicast details, etc. Default connections inc:
Loopback, Internal, Local Area Connection.
·
Static
Routes = Show IP Routing Table, New
·
DHCP
Relay Agent = new interface, add addresses of new dhcp servers. The dhcp relay
agent sends messages to the dhcp servers listed here.
·
IGMP
= ?
·
Remote
Access Policy = Configure what conditions need to be met to allow/refuse
connections (ie, phone no dialled by user, group user belongs to, phone no from
which call originated, etc). Also includes a PROFILE for each policy. Profile =
HOW the connection is handled, ie, disconnect if idle for, restrict to
following days/times, ip address assignment policy, multilink settings,
authentication, encryption, etc.
·
Remote
Access Logging = The third type of logging can be
config here. It is for authentication and accounting requests received by the
server. (The other 2 are errors reported in Event Viewer, and PPP details
reported in Sysroot/tracing).
RRAS Concepts:
·
Dial
Up Networking = Client makes a dial up, non permanent
connection to a physical port on the RRAS server, by using phone line, ISDN or
X.25. It is a DIRECT physical link, you can encrypt data but it is not
required.
·
VPN
= is the creation of a point to point connection across a private or public
network (internet). Is makes a virtual call to a virtual port to a VPN server with
tunnelling protocols. It is a INDIRECT, logical
connection that needs encryption.
Dial UP
Server uses a modem, bank of modems, ISDN connection, X.25, infra red, null
modem cable.
Dial UP
client must have modem and remote access software.
VPN can be
either a user connecting to a network, or a router connecting to a another router to form a router to router VPN connection.
VPN server use permanent WAN connections such as T1, etc. Clients can use the
same, or dial up to an ISP.
VPN server
must have a separate nic for connection to internal network. Remember, it is a
ROUTER and as such can be used without an off the shelf ‘black box’ router.
Accepted way is to have:
Internet---cable
modem---firewall---vpn server---hub/switch
Firewall
is to filter out everything except vpn traffic.
However,
it seems you can use a vpn server with only 1 nic if you configure the nat
router/firewall to do port forwarding (VPN PASSTHROUGH) to the vpn server. The
confusion here is that MS seem to want you to use the server as a router and
vpn and DNCP and everything. If you have a router on the lan
you don’t need a vpn server with 2 nics!
Protocols
= Dial up uses versions of Chap, Pap, etc. VPN uses L2TP, PPTP.
What I
did: 1 vpn server with 1 nic. Put it in DMZ, (not rec –
only testing). Only used PPTP. Gave account dial in
permission. Worked! Cannot browse (common fault). But
CAN map drives to network shares or access all shares on a pc by ‘run ->
\\servername’.
Services
Stop/start/auto config services. Can test stop services in msconfig, and if want to
permanently stop them – use services. Can also specify which services
start/stop in different hardware profiles! Some of the services inc: Auto
Update, DHCP Client/Server, FTP Publishing Service, Network Connections, etc.
Telnet
Server Administrator
If you run
a telnet server this is where you can disconnect users, List current users,
start/stop, etc.
Terminal
Services Client Creator
Lets you create
floppy disks (4) with the terminal services client app on them. Can also
install from nw share…
On Term
Server, share: /systemroot/system32/clients/tsclient/net. In here are 2
folders, for 16/32 bit systems. Drill to Disk1 and install from there. TO avoid
confusion for which to install, set perms, share individually, etc.
Terminal
Services Configuration = Connections and Server Settings
Connections
(for each protocol ?): Encryption, Logon Settings (always prompt for pw, etc),
Active Session Time Limit, End Inactive Session in Spec Time, Env (disable wall
paper), Remote Control Settings (view session, interact with session, require
users perm, Client Settings (spec printer, clipboard mapping, network adaptor
(match with spec protocol), perms.
Server
Settings: Terminal Server Mode (change Remote Admin or App Server. Changed via
Add/Remove Progs), Delete Temp Folders on Exit, Use Temp Folders per Session,
Active Desktop, etc.
Terminal
Services Manager = Admins can see all sessions, users and processes for each
terminal server. Can also Disconnect users, Send
Message, Reset, Logoff, etc. Can also Remote Control a session from here.
To Set
Up Remote Assistance
On User’s
Target PC: Need to enable pc user to ask for Remote Assistance/Control via
ADU&C ->choose individual user -> Props -> Remote Control: Enable,
Require Users Perm, View Session, Interact with the Session.
You can
config user perms on a policy/domain wide basis but you need to do it via an
extension/2003.
To
Initiate Session: User requests admin help via email, messenger, nw share folder containing requests.
Therefore
you need Remote Desktop/Assistance and Terminal Services.
Remote
Assistance uses Remote Desktop (Terminal Services) technology.
Remote
Assistance and Term Services are included in XP. But for 2000 and older you
will need to install RA from either XP cd or download from MS site. Term
Services is included in 2000 cd.
Remote
Desktop
Two types.
1: Remote Desktop via client connection (inc in xp, or
installed from 2k cd), OR Remote Desktop via browser.
Remote
Desktop does not need a connection license as it is designed for 1 user –
whether they are remote or local. Hence why the desktop locks
when the user connects remotely.
To
setup via client:
·
In
XP: My Computer -> Props -> Remote -> Remote Desktop.
·
In
2K: In
·
To
connect to a 2k system running TS, you can do it via www interface (/tsweb), RD
connection app in xp, or create TS client with client creator files and enable
a 2k system to connect. Can also install RD software from XP onto 2k system.
·
Browser
= ‘Remote Desktop Web Connection’, = Ipaddress/tsweb
To
setup via browser:
XP
·
Host
= Add/Remove Progs -> IIS _> Details -> WWW Service -> Details
-> Remote Desktop Connection.
·
Stop
service = net stop w3svc
·
Check
for updates
·
Start
service = net start w3svc
·
Configure
RD: RC My Computer -> Props -> Remote Tab -> Allow Users to
Connect Remotely to this Computer. Add Users.
·
Ensure
Users have correct permissions in their props in Remote Tab.
·
To
connect from client: ipaddress/tsweb Can also connect
over internet – can use ip address or register and use a domain name.
Win
2K
·
Need
to download app from MS (google ‘Set Up Remote Desktop
Web Connection with Windows 200’ to find it), works for NT too.
·
Install
it. When it asks where to install sample web pages, specify: C:\Windows\Web\TSWeb
·
Rest
of procedure: As above? Update, configure RD, Ensure
user permissions.
·
To
connect: Same as above.
Wins
Is a
dynamic replicated database that matches NetBios names to IP adds
on the network. Serves WINS enabled clients – described as NetBios over Tcp/ip. All OS’s before 2k need NetBios. Even modern OS’s
might be running services/apps that need WINS. WINS registrations are done auto
– whenever a client joins a network, or dhcp issues a new address, etc.
Domain
Controller Security Policy
Is the
security settings found in AD/U&C -> Domain Controllers -> Windows
Settings -> Security Settings.
Domain
Security Policy
As above, but for the domain level, ie, the very top of the AD/U&C
tree. AD/U&C
-> top of tree.
Network
Monitor
Analyze
network data traffic to find problems, ie, which pc makes most traffic, id
unauthorised users on network, see patterns, etc.
Need to
capture frames/packets, then display/filter/save them.
In-depth!
Read up.
Connection
Manager Administration Kit
Lets you
make a self installing exe which you distribute to users (cd, nw share, etc). You configure a set of settings to connect
to your network/servise for a users/group. It is this series of settings that
you configure and convert into a exe. Read up.
Internet
Authentication Service
Centralised
authentication, authorisation and accounting of users who connect to a nw using vpn or dial up. Uses Radius.
Needs to be registered in AD.
Not used
before. Read up.
Profiles
Profiles
consist of desktop settings, shortcuts, icons, application settings, outlook
express contents, etc.
1.
Local
Profiles
2.
Preconfigured
Local Profiles
3.
Preconfigured
Default Local Profiles
4.
Roaming
Profiles
5.
Preconfigured
Roaming Profiles
6.
Network
Default User Profiles
7.
Mandatory
Profiles
·
Documents
and Settings: create/contains profiles for every user.
·
All
Users: Contains shortcuts and icons available to all users on that machine.
·
Default
User: (Hidden by default) OS uses it as a template to create new profiles.
So a profile
is a mix of their own folder contents and All Users.
A new
profile is made using the Default User as a template, and the All Users for
common settings, and all this is copied into a folder named after the new
user’s account name.
Roaming
Profiles
Admin
creates a nw share, and specifies a path (\\servername\profiles
(sharename)\ %USERNAME% to it
in the users account in ADU&C. %USERNAME% means OS will create folder auto
according to username.
Local
(offline/existing) profiles are matched with nw
profiles using timestamps.
How to
Preconfigure a Default User Profile
Make a
bogus user account, and set it up as you wish. Copy it to the default user
profile. (lof off bogus acc and logon as local admin).
Must unhide default user acc first.
You can even
copy this acc to a server location to serve as a preconfigured roaming profile!
Use the
SYTEM APPLET to do this copying. It will do the above and assign appropriate
perms as well! It will also let the OS know we are dealing with profiles and
not ordinary folders/files. Remember to change the Permitted To Use value = to Everyone or the relevant group/user.
Note – Setup preconfig roaming
profiles by copying the profile to the SAME UNC path you specified in the
account properties, ie, \\machinename\profiles\user
acc name, and also use that path as the user prfile for user’s DOMAIN acc.
How to
Precreate a Default Roaming User Profile
To
precreate a roaming profile for a lot of users you don’t want to copy a
precreated profile into each users dir. Instead you
can copy the profile to the NETLOGON share on a DC to setup a domain wide
default user profile.
When
logging onto a domain the pc first looks for a Default User dir in the NETLOGON
share (\winnt\sysvol\sysvol\domainname\scripts\) of the authenticating dc. Only if no
Defaul User profile is NOT found in the NETLOGON dir does the machine
use the local Default User profile. If a Default User profile IS found in the
NETLOGON share (where logon scripts and sys policies are also stored) all new domain
user’s with NT pc’s will use this dir as the domain wide Default User template.
Specify the path as \\machinename\sysvol\domainname\scripts\Default.
User (must name the dir Default User), and grant perms to Everyone,
or what ever. Alt if you are part of admin group you can save it to \\machinename\netlogon\defaultuser
Restrictions
on Default User
If you
want to place RESTRICTIONS on the default user profile you will need a registry
editing tool (edit ntuser.dat hive file in nt4). You cannot log in as Default
User and change that and think that restrictions placed on it will be saved –
they won’t. The OS gets confused. In depth – read up.
If you
need to restrict what user’s can do – use system and group policies.
Problems
with Roaming Profiles (default or otherwise)
Video
settings on different machines, ALWAYS use default installation dirs when
installing apps – or shortcut will firstly try to resolve via Absolute Path
(local machine specific), then netwotk path (but uses hidden admin shares ($)
so asks for perms), then lastly attempts to resolve itself via ‘search’.
Each
machine a roaming users uses stores a cached local
copy. These can take space, so can delete via System app, or policy (computer
config – admin templates – sys – log on –
Lastly –
roaming profiles over a wan link not recommended. Slow link causes numerous
probs when synchronising profiles.
Cached
Profile Problem
If delete
an account from AD, the local cached copy could still let a user log on if not
connected to nw (no dc found, etc). To avoid this, use
System Policies (read up).
Mandatory
Profiles
Involves
changing users profile/dir name to include .man
extension. Also involves changing ntuser.dat to ntuser.man (found in the profile/dir)
so that this can be assigned to groups of users – they all share the same
mandatory profile.
Te set it
up configure your desktop etc, and then copy it to nw
share. Set appropriate perms. Change the copied ntuser.man to .dat. Change the
account to point to this profile/dir.
The .man
instructs the OS to not save changes made.
Not
in-depth –> on p729 & 730, but is isn’t the
recommended way to do it. Much better to use policies to lock down and control
desktops!
Setting
Up a Group Template Profile
To assign
a shared profile to users be aware that it should be a mandatory profile. If
not, when 2 users log off BOTH their changes will be saved!
If you
want a template profile for a group/user it will serve as a departure point
only.
Create a
bogus account and logon and configure it as you see fit.
Open the
system app, and copy the template profile to a shared dir, renaming it after
the user you are about to create.
Then
create the user, specifying his profile path as \\machine\profile
share name\user name.
This
profile isn’t complicated, but you do need to remember to copy a new profile to
the shared dir and assign the profile to that specific user each time you
create a new user acc.
Win 98
Isn’t compatible with 2k/xp, so better not to use roaming profiles.
RIP
Is the
protocol that routers use.Routers use it to exchange
information about routes so that they keep their routing tables up to date. It
is an alogithm, in this case it represents each
network (not router itself) as 1 hop. Therefore to send a datagram within the
same nw = 1 hop. To cross over 1 router = 2 hops. To
cross over 2 routers = 3 hops, etc. Highest number of hops = 16.
On a
regular basis each router will send out it’s routing
table entries to other routers about the networks and hosts it know how to
reach. The routing tables have their hop count adjusted (+1) as it goes through
routers. In this way if a router receives an update for a route the version
with the lowest hop count is used. The routing tables, as they propagate
through routers, have the current routers routing tables added, and hop counts
adjusted.
In this
way all routers are eventually updated with optimal routes.
In some
cases it is not advantageous for every nw in a large
interconnected nw to be fully specified. In this case a
V1 and V2
differences: V1 is a classful protocol, and v2 understands CIDR or VLSM
(variable length subnet masks), (when you subnet a subnet!), v1 does not
support route update authentication, it’s broadcast
packets generate sig overhead, and it has slow convergence (self healing when
there are nw outages).
Outlook
Addresses and Messages
To backup
outlook there are 2 parts you need to do this to: the Contacts, and the
Messages.
Contacts:
·
On
local machine: \Windows\application data\microsoft\outlook\outlook.pst
·
If
several users: I found addy book here:\Documents and Settings\user
name\application data\microsoft\address book
·
If
profiles: \profiles shared folder\user name\application data\microsoft\address
book
·
Most
of these are in ‘address book format’. (.wab)
·
In
Outlook Express: Addresses (top menu) –> file –> import/export
Messages:
·
In
Outlook Express: Import/Export
·
Location
of messages is NOT on nw, and I cannot find a way to do this (no browse nw
option, etc), but if you DO need to backup to a nw location, the folders are
here:
·
Documents
and settings\user name\local settings (hidden folder)\application
data\identities\sid no\microsoft\Outlook Express\ and here are different .dbx
files (inbox, outbox, deleted, etc)
How to
Make Groups on Outlook Express
Addresses on menu bar -> New -> New Group. Displays all contaqcts, from here
select the ones you want in the new group. Click on group -> Props = select
contacts to put in group. They are also in the main contacts list so be aware
they appear more than once. If group is
Import
Pol
3 elements
to be aware of:
·
1.Can IMPORT pol’s in ADUC.
·
2
& 3: Need to use MMC and 2 snap ins: Security Config & Analysis, and
Templates
In SCA you
need to create new database (Type name of a database when you want a NEW
database in ‘Open…’). The database contains the results of the template
comparison (Analyse Now) with existing security config.
The
database therefore contains a template, and any difference made to it. You need
a database because it is the process of compiling an ascii
file to a binary file containing security settings!
So you can
create a new security template from scratch, or import one, or import one and
change it.
You can
compare any changes you make, or a template, with your existing pc SCA and
Analyse Computer Now. The output can be text (log) or a security display
showing those parts which are the same/differ.
·
Default
= Setup Security.inf, is default on workstations. Should NEVER be applied using
gpo - rather it is used for disaster recovery.
·
Compatible
= Compatws.inf, takes everyone out of power users group, and relaxes users
group to give them more rights. By doing this users aren’t put into power users
group to give them elevated privileges.
Others
include different levels of security.
Templates
are incremental!
Templates
Contains all the included default security templates and any that you
make as well.
Templates
can be modified DIRECTLY if you use the Templates app instead of importing into
a database in SCA.
Thus far….
Need a database file to work in and cannot work on templates directly.
Templates can be imported, altered. Templates are then compared to existing pc
via Analyse Now. PC can be changed via Configure Computer. Templates can be
imported into policies via Import under Security Settings in relevant policy.
When
importing TP use ‘clear this database before importing’ to ensure you aren’t
merging templates.
Save TP
via Export Template.
What
is Sysvol, Netlogon, Ntds, Global Catalog, Winnt, etc?
Sysvol
(and Netlogon) (Profiles, Policies, etc)
In NT4 a
lot of important info (user config and control info, including system policy
files, default profiles, and login scripts) was stored in Netlogon on PCD. But
BDC's need access to this too. In Win2k all DC's have Sysvol replicated among
them automatically and in Sysvol is the Netlogon folder!
Therefore
AD and Sysvol are different. But both are replicated among all DC's in a
domain.
When
referring to System State data (ie when backing up), this includes a copy of
the AD and the Sysvol, in addition to other data such as System Boot files,
COM+ database, System Config info, etc. When restoring a DC, you need to start
the DC in 'Directory Services Restore' mode.
Ntds
(New Technology Dir Services) (Naming Context inc
Config and Schema, User Accounts, Links (member of), Groups, Application
Specific data, etc)
The active
directory itself is a database, and like most databases it has 2 parts, the
database itself and a transaction log. Both are normally stored in
c:\winnt\ntds\. You can put them on 2 separate hard disks for noticeable
performance gain, but need 2 scsi.
The
NTDS.DIT file is the 'main' part, or heart, of the AD database, and is located
in winnt\ntds\
NTDS
Settings in ADS&S is the replication to and from servers, etc, of the AD database.
Global
Catalog
All dc's know about their own domain, but not about the entire
forest. Only the DC with the GC knows about all the objects (but not all it's attributes. It DOES know about universal memberships,
and some other attribs, but not all) in the forest. Why only 1 DC? Replication issues. But you can manually change this.
Therefore GC's are used to deduce UNIVERSAL GROUP memberships in a forest.
Winnt
Used to install windows from a cmd/dos prompt. Can use Setup but if not in gui environment can use winnt. Go to i368 folder -> and
either winnt.exe or winnt32.exe.
·
winnt.exe = for when you are in dos or Win3.x or booting from a boot
disk. For a FULL CLEAN install.
·
winnt32.exe = for when you are in win9.x or NT (cmd prompt). It UPGRADES
or FULL CLEAN install.
Can add
the recovery console in winnt32 only = /cmdcons
So
broadly, winnt32 for higher level (windows cmd) and upgrades = winnt 32 for
UPGRADING in a 32 BIT environment. Winnt for low level (dos,
win3.x or boot disk).
Various
switches can be added, to invoke unattended installs, answer files, cmd's to be
followed, etc. Use internet for details.
RIS
Need DHCP,
DNS and AD running on nw.
RIS can do
3 types of install:
1. Simple i386 based install. Similar to
a cd install, but over a network and started automatically.
2. Scripted i386 install. Can add a
script to make the install an unattended installation.
3. Complete system image with minimal
setup interaction. You build an entire prototype machine with apps etc, and use
this to create an image.
Installing
RIS
Add/Remove
Remote Installation Services. Type risetup.exe = wizard. Where to install RIS
files = NOT system or boot drive. Must be NTFS.
Initial RIS file will be a copy of the cd so ensure at least 800mb.
Authorise
RIS server in AD = DHCP -> Manage Authorised Servers -> Add IP address
-> Authorise.
To
configure/turn on/off/ configure pc names etc the RIS server you need to get to
the Remote Install -> Props in AD,U&C. But you
can ONLY do this on the RIS server itself – you can’t do it on a DC unless that
DC is the RIS server. On RIS server type DSA.MSC to invoke the AD,U&C console and RC the relevant pc and drill down from
there.
Updates
for XP
RIS can
also image/store/riprep XP images, but you need to update/patch it first.
·
To RIS XP:
See Knowledge Base Article 308508 ‘Unable to Create a Windows Serer Image on a
RIS Server’.
·
To Riprep XP: See Knowledge Base article 313069 ‘Update for the Riprep tool’.
Setting
Required User Perms
AD,
U&C -> RC domain name -> Delegate Control -> Add (users who can
add computers) -> ‘Join a Computer to a Domain’ -> Next/Finish.
OR Create
Installers Group. (In depth, see page 129 for info)
Installing
Clients by Using RIS
Use boot
floppy or PXE. F12 = download client installation wizard = user name, password,
domain name. Product ID.
To make
boot floppy = on cd, \remoteinstall\admin\i386, use RBFG.exe
Can enter product ID in .SIF file to avoid this step.
RIS
Extras
If you
have a server with several different images, but you only wanted certain
users to see/access certain images, then the only way to do this is via
permissions.
Can only
image the C drive.
RIS Server
needs two drives. C (which contains system files and is unused by RIS), and D,
which RIS will use to store images.
When there
are multiple images of the same os, the SIS (Single Instance Store) will ensure
that duplicate files aren’t made and therefore 2nd image onwards
will use much less disk space.
The first
time of setting up RIS will invoke a process to make an image on that server –
therefore you need a cd. This will be the first image on the server and will be
a simple i386 installation.
Each image
gets a folder in the Remote-Install folder.
Setup
Manager
Used to create Installation scripts. Can be used to
create scripts to aid unattended installations, ie tailored cd rom. The
answer file can be used with RIS and Riprep images. The script can be further
enhanced manually. In depth script language for this.
Can also manually alter script to add SPs, hotfixes, even other apps. Also in-depth.
Need to know/use $OEM$ for this.
To INSTALL
Setup Manager: On W2K cd, drill down to \support\tools\deploy, and COPY/PASTE
all the files to a local folder that you make.
To RUN the
Setup Manager: Click on the setupmgr icon.
However,
the password for the local admin account is not encrypted!
Can create
answer files using Setup Manager, and associate different answer files with
different images in dsa.msc in Ris server properties. In this way you can have
1 image with several answer files, and when the client remote installs, they
will be presented with several choices, each of which is a different answer
file.
If you
make an answer file with Setup Manager and it is a .txt file, but you need to
turn it into a .sif file for it to be recognised by dsa.msc to associate it
with an image, just change the name of the ext from .txt to .sif. However in
Notepad, the default is to ALWAYS include the .txt AS WELL as the .sif, ie
filename.sif.txt. To get around this when you save put the name and the
extension you want in QUOTES, ie “filename.sif”.
You can
also get around this restriction of including the .txt extension in Windows
Explorer. Tools -> Folder Options ->View -> turn OFF ‘Hide File
Extensions for Known File Types’. Thenyou can RENAME the filename AND the
extension properly.
When you
rename the file to .sif and associate it with an image via dsa.msc, the file
will appear in the relevant ‘templates’ dir belonging to that image. .Sif files
in a given images’ templates dir are the ones that are presented to the client
upon booting via PXE.
RIPREP
Riprep is
used when you want to deploy an os and some apps. Obviously you can only image
c. Set up machine as you see fit, with all updates, apps, etc. This is the prototype
machine.
Run
Riprep, which strips all SIDs off the machine. To Riprep the prototype machine,
you nedd an active RIS server running (see above). Using domain admin account,
navigate to RIS server and the share called REMINST. Open it, and \admin\i386\
and a file named riprep.exe. This will start the riprep wizard. You can sent the result to any RIS server.
Once image
is on server it is available for remote installations.
After you
have run Riprep on the prototype machine you will go through a mini setup to
restore the original information.
Delivering
a Riprep Image to Target PC
Now that
there is more than 1 image on the RIS server, there will be multiple
options/images. If you choose the ripreped image (as opposed to the first,
‘simple i386’ cd based image) the entire install will be hands off.
Enabling
Users to Start RIS Transfers:
Users need
to be able to create a MACHINE ACCOUNT, so you need to give them the correct
perms. You could add them to ACCOUNT OPERATORS group, but that is a lot of power
just to kick off an image transfer.
So instead
we create an INSTALLERS GROUP. (Lengthy procedure, p129).
‘I think’
we could also delegate permissions. (See above for info).
Restricting
Image Choices
The RIS
server has a lot of dirs in \RemoteInstall\Setup\English\Images. So a simple
i386 os would be \RemoteInstall\Setup\English\ImagesWin2000.pro. Each Ris image
has a dir in the Images dir.
Each Image
contains a folder called TEMPLATES, and in that folder is a file with the
extension .sif. This is an answer file used to install without user
intervention.
The way to
control which groups see what os choices there are is to DENY READ ACCESS to
the relevant .sif file in Templates.
Scripts. Editing the Answer File to Get What you Want
Using the
above you will get an unattended install, except for the Product ID! To get
around this you can edit the .sif file.
In the [User Date] section, ProductKey – 1234-5678, (For XP), or
ProductID – 1233-45678 (For 2000).
Un[attend]
UnattendMode
= FullUnattend
The
dea=fault naming convention is ti use the server name as a guide, so Server
would produce Server1, Server2, etc.
You can
also change the naming convention FROM THE SERVER. AD,U&C
-> RC Ris server -> Props -> Remote Install -> Advanced -> New
Clients -> Options include:
·
%#username
= Username
·
#%first
= First name
·
%#last
= Last name
·
%#
= Increment by 1
However,
this is at RIS server level so I don’t know if it will affect both i386 cd
images and Riprep Images.
Note: You can use Setup Manager to
create multiple answer files for a given image!
Riprep
Extras
Target PC
disk size will be same as prototype machine.
Cannot easily associate different riprep .sif files with different
images (unlike scripted installs (i386)).
So just
change .sif file to configure product id, machine name, etc. (I think? Test
this)
NTOSKRNL
Issue: When applying kb835732 hotfix to prototype, Riprep won’t work, This is due to the NTOSKRNL being different between the
prototype and the existing image on the server. To fix ->
copy the NTOSKRNL from the client/prototype to the image on the server.
Ensure you REPLACE the original (in the I386 dir of the flat image).
Stopping Services. During Riprep you will be asked to stop services. Do this via Computer
Management (the computer management name will be presented during the Riprep
alert). Those services without a Computer Management alias can be stopped via
Ctrl, Alt,
Qmqr0.dat
Issue: When running Riprep, Qmqro.dat and Qmgr1.dat won’t copy to the image. To
fix: edit the .sif file to skip these:
Reminst\admin\i386\riprep.inf
(governs running of Riprep).
·
[Files
to Copy]
·
At
bottom type:
·
%16419%\microsoft\network\downloader\=2,dat
·
And
save.
Or it seems you can ignore error
messages. Seems to work for me!
Password
Encryption Issue: Password not always encrypted in Answer files with RIS
images. But with Riprep images password is always encrypted!
Remember,
i386 and scripted installs are cd based, Riprep is image based (but can use
some form of script/answer file with it). Hence why you can’t
associate different Riprep answer files with different images.
Riprep
images will NOT include 3rd party drivers for the NIC in the image!
So if you have 3rd party NIC drivers you need to use, you need to go
through the hassle of using the $OEM$ folder structure. Much better to ensure
the NIC on the prototype pc and the drivers in the os are compatible. AND, it
also seems that you might have to add the drivers to the first/primary flat RIS
image as well!
Image
Based Installs Using SYSPREP
Sysprep is
used to prepare a disk to be copied/cloned by a third party app such as Ghost,
etc. It strips off the SIDS from the original os.
Download
the latest sysprep from Microsoft.
Sysprep
cannot be used on DCs.
GPO
to Deliver Apps and Updates
Using GPO
to install is a major tool for managing software in an org.
User does
NOT need admin privileges to install software add/remove programs when they
have been added here by the admin!
The
components of delivering apps via GPO:
1. Software installation in GP.
You
use GP to control and manage apps, called PACKAGES.
2. The Windows Installer Service.
3. The Add/Remove applet.
2 and
3 are used to install/remove the apps based on what you have set up in GP.
Assign
Assign an
application when you want a group to have an application on his computer.
Software can be assigned to Users or Machines. The software is assigned on
first reboot or when user first logs on.
Publish
Publish
when you want to software available to install by users. They can install it if
they decide to. You can only publish software to users. Users install packages
via Add/Remove applet.
Only
publish or assign to EITHER users or machines, both = conflicts.
Packages
Packages
are .MSI apps. Some vendors have .MSI apps on their cd/site – these can be
used. Alternatively if you need to make an .MSI out of a program, you can
install and use Veritas’ WinInstall (more later).
To
Install Software Through a GPO
Using
adminpak.MSI as an example of Publishing a Package to
Users…
1. Create a nw
share and copy the .msi app to it.
2. Create a GPO.
RC the domain name -> Props
-> -> Policy tab ->New ->type Adminpak -> Enter
3. Filter the GPO.
Filtering
lets you choose users to which the GPO applies to. Props -> Security
->Ensure Apply Group Policy perm is UNCHECKED for all groups except the
one/s you want. As for further perms, it seems prudent to just have Read, so
total perms = Apply Group Policy and Read.
4. Add the package to the GPO.
Next
click the edit button. This will launch the GP snap in. Either Assign or
Publish. (Publish in this case). Packages are added to the GPO in Software
Settings\Software Installation in either Computer Config or User Config (in
this case User Config). New -> Package, drill down to .msi on nw (\\servername\sharename).
You can RC the package and choose/change Publish/Assign.
When
testing the Publish .MSI installation, LOG OFF/ON first and Add/Remove
Progs.
Filtering
lets you apply the .msi to anyone in the org – be it 1 person, or everyone, or
a group. Any group or user in the SECURITY tab of the GPO will get the policy
(and therefore the .MSI) ONLY if the Apply Group Policy option is checked.
The gist
is to create a separate policy just for the .msi installation, and filter it
via Security to let select people install the .msi.
We can
uninstall apps via the ‘Uninstall Applications When they Fall Out of the
Range…’ option in Software Installation.
Using
OUs
You can
also use OUs (and therefore ignore filtering) directly to roll out apps. It is
a good idea to have a Beta group of users in an OU to test the roll out
process. You can then ADD the GPO to other OUs in the org,
or to the domain directly. This is a useful, step by step process.
Assigning
a Package to Users or Computers
Assigning
a package means the app gets INSTALLED. If you assign the app to a computer it
gets installed at boot time, prior to logon. If you assign to a user it gets
installed at logon.
If a user tries
to uninstall it, it gets repaired/reinstalled next time!
In
actuality the app only gets partly installed (files are copied, shortcuts are
installed, file assoc etc are done. But the app is completely installed upon
FIRST USE, whether a file assoc with it is clicked, or the app is opened
directly.
In this
way assigned apps ‘roam’, much like roaming profiles.
To assign
an app, using Office as an example, the steps to follow are:
1. Run the Administrative Setup
Why
admin setup? Admin setup allows you to manage the app files centrally. It
allows you to create a CUSTOMISED version that users will install (see
paragraph below). Allows you to config options such as ‘install on first use’
OR ‘run from nw’. It also allows you to manage updates
by patching just the 1 Office image. To do admin install: Drive:\setup /a (on Win2K
you need SP3 >).
Need
to create a TRANSFORM and modify the Setup.ini file to customize the Office
installation for clients – indeed you can create multiple configurations for
the same CD based Office image. However, doing a plain admin install without
later customising it will let users have all the options normally present
during a cd install minus the need for a product key.
2. Create a GPO called Whatever
3. Add the package to the GPO
4. Customize the Package Props
Removing
a Package
Simply
check the ‘Uninstall Software From Users and
Computers’ in Software Settings -> Name (of installation).
Redeploying
a Package
Useful if
you have added modifications to a package, etc.
Creating
Your Own MSI
Two third
party progs on the Win2K Server cd are used when you need to create your own
msi files Drive:\Valueadd\3rdparty\Mgmt\Winstle\Swiadmle.msi (For some reason I
couldn’t install this locally, I had to share the cd drine on another pc and install
it to the machine I want to run Veritas from (and not ON, which would be the
clean pc)). So 3 pcs, 1 with shared drive, 1 to install it to and run from, and
a third which is the clean pc!
Veritas
Software Console allows
you to view and edit and msi, and WinSTALL Discover allows you to create
an msi package from an old style install/setup prog.
Broadly
speaking, you are taking a BEFORE and AFTER snapshot of a computer.
Might be
worth making another admin account and use this. If not you could get icons following the main admin account if using roaming profiles.
1. Create a Clean Computer
Only
has os and service packs installed. The differences between
the before and after snapshots are used to create the msi, so they all need to
be correct and accounted for. It should also NOT have the
Veritas/Winstall Discover software installed.
2. Take the Before Snapshot (using c$
and wininstall)
Go to
clean pc and run Winstall Discover. It is best to run it via the Run prompt,
and not by mapping a drive to it. Therefore the command will look something
like \\servername\c$\program
files\veritas software\winstall\DiscoZ.exe. When wizard asks what files to
exclude – accept the defaults. You can choose where to store the .msi, the
default is to store it on the server you’re running Discover FROM, in the
Winstall dir.
3. Install the Application and Reboot
Install app as you want the .msi to
install it. And REBOOT!
4. Test the Application
Keep in mind any changes made probably will not end up in
the .msi.
5. Take the After Snapshot and Compare
Log
on to what was the clean pc and run Winstall Discover again to take the after
snapshot. Again it is a good idea to run it across the nw
using the notation in Step 2. If you need to make another/the .msi again you
need to ensue the pc is clean again – which means you might need to reinstall
the os. Very often the wizard produces a list of warnings – check for red flags
and test the app.
6. Make any Customizations
Now
check the msi and possibly modify it. Use the Veritas Software Console:
\Winconsole\Seasw OR
from the start menu on the server that has the msi choose the Veritas Software
Console and Open the .msi. You can also check >miss from off the shelf
packages. Can edit what dlls to use, etc. In depth (p 884 for details).
7. Test the Application Installed by the
MSI
This
is the most time consuming part of the process. Need to install it on test pcs,
inc clean and unclean ones to ensure no conflicts with
other apps, etc. This is where conflict errors (dlls, etc)
crop up and can be sorted by using Veritas Software Console.
Zap
Files
Zap files
are simple text file which can be published to users (not assigned, and not to
machines). The text file is a direction
to the real/actual installation prog, which does not have to be an .msi file!
It will
appear in Add/Remove progs, or it can be installed by clicking on a file
associated with the prog – also spec in the text file.
Good way
of publishing setup/install progs on the nw for
selected users.
1. Create a Zap File
·
[application]
·
FriendlyName
= “Winzip Version 72”
·
SetupCommand
= \\Servername\packages\winzip\winzip70.exe
·
DisplayVersion
= 7.0
·
[ext]
·
ZIP
=
·
Application
= is the name seen for the package description in the GPO editor and the users
will see when they install the package. SetupCommand = This
is the actual installation prog run when the package is selected in Add/Remove
or double clicks the associated file. DisplayVersion = Is
displayed with the package in the GPO editor. Ext = is the next section and must
include any file ext you want associated with the prog. Ensure you name the
extension .ZAP.
2. Share the Zap file and Installation
Files
Create
a dir in the packages shared dir on the server. Both the Zap file and the
Installation prog should be in this dir. The path in both cases is: = \\Servername\packages\winzip\
but the installation path ends with winzip70.exe and the Zap file path ends
with, for example, Winzip.zap. When you copy the installation prog (ie
setup.exe) into the dir, ensure you copy ALL the files over – which could be a
cds worth of data.
3. Add a package to the GPO
You
can create a new GPO or add a new package to an existing GPO. Now we must add
(cannot assign) the package to UserConfig/Software Settings/Software
Installation. New Package and drill down using the path above as an example.
I
did this but when it cam to testing the installation failed. It found the
package, but when installing it couldn’t follow the nw
path. It included spaces and brackets (new folder (3)). I changed the location
to a path with normal chars and it worked perfectly.
Lastly,
test!
Customising
Off the Shelf Apps
Using
Veritas Software Console you can customise the .msi packages from 3rd
parties. Useful for dll conflicts, etc. In depth (p888).
You can
also provide TRANSFORMS to .msi packages. These don’t modify the .msi directly
but rather use the official .msi process and apply differences to that. But to
do this you need an MST tool provided by the software manufacturer.
Upgrading
Applications
The
upgrade package must be .msi format. And we treat it like other .msi packages.
We put the package in the correct shared folder. And we use AD,U&C to assign/publish it like any other package. In the
case of filtering you use the same GPO, and add the package to it, using the
perms to control who accesses it. Or we can use a GPO
directly if that is how the original package was deployed.
In either
case, once we have the correct GPO, we use the Software Settings -> Choose
the UPGRADE .msi -> Props, and choose Upgrades. Here you select the
current/other GPO and its associated .msi packages to which you want to apply
the upgrade. You can upgrade existing package, or uninstall current package and
replace with newer version.
To
summarise, you have upgrade package in .msi format, publish it via GPO, and
from here link it to an existing package which it will upgrade/replace.
In this
way you can have users choose to upgrade/replace software, or force it (ie,
when they click on the associated file it could invoke an upgrade if config in
Deployment for the upgrade .msi package).
Recovery
Chart
Slow
System?
Ctrl/alt/del
and/or Perfmon
Problems?
Msconfig (disable services, boot.ini, system restore. If disabled services, it is a
temp/test, do it properly via Services)
Need To
Recover AD?
Via backup (system state), or multiple DC’s replication.
F8
Options:
·
Safe
Mode = Use this to fix network problems if cannot get onto network.
·
Safe
Mode With Network Support = Use this to fix problems
if you know network is NOT the problem, and you need network services.
·
Safe
Mode With Command Prompt = Cmd shell, but can still
invoke gui apps.
·
Enable
VGA Mode = Can help sort display problems that prevent Windows from loading.
·
Last
Known Good Config = Do this FIRST! Any logons after problem has manifested
itself will LOSE setup data needed for successful boot. Unless you create more
than 1 hardware profile. LKGC will use drivers etc used in last successful
logon/alternative profile.
Emergency
Repair Disk
ERD is
made via Backup. It includes the option Repair Installation, which can be run
from Setup/boot floppies or Installation CD.
ERD
contains copy of setup data, services, drivers, ntldr, etc. these can be
replaced (repaired) when ERD mode is chosen.
ERD
modes/Choices:
Manual Repair ----- allows more control i.e,
choose:
·
Inspect
start-up environment (corrupted Windows files?)
·
Verify
Windows system files.
·
Inspect
boot sector.
Fast Repair ----- inexperienced admin doesn’t
need all these. However, can only fix REGISTRY with fat
repair, not manual repair.
Recovery
Console
RC can be
run from Setup/boot floppies, installation CD, or installed on pc (via
\i386\winnt32 /cmdcons typed into start – run. Available as
boot option).
Can’t
access non system partitions! But can do stuff if you’re familiar with it. Need
to log in as local admin.
Can
replace missing system files, can FIXMBR, FIXBOOT, (but check for viruses
first). Do boot first, THEN mbr, (use book for guidance when using RC)
Boot
Floppies
Made via
app cd, which can be run on ANY W2K machine = drive:\bootdisk\makeboot
on install cd. It will produce 4 floppies, which have the same initial process
as installing W2K from cd, but doesn’t install, rather it leads you to the
repair W2K via:
·
Recovery
Console
·
ERD
Therefore good when the PC won’t boot at all, either from hard disk or
cd.