APG40 I - C Course Notes
APG40 I - C Course Notes
GO THROUGH SLIDE
External Connectivity
GO THROUGH SLIDE
The IPN connects to the IPNX board in the RPHM A and B side.
There are 64 external alarms in the APG. The external alarms are not redundant so
there are 32 defined on each node. 0 to 31 for Node A and 32 – 63 for Node B
The locally defined alarm displays are addressed using dipswitches on the alarm
board. These are configured using binary addresses.
Note: The crossover cable between the Public LAN ports is very important for cluster
communication. This has to be done during the Startup and Test Procedure.
Configuration
Configuration Slide
• Power Up APG40
• Log-on
• Site Parameter Change
Power Up APG40
During power up of the APG40 it is essential that both power cables are inserted
simultaneously to maintain the integrity of the cluster and subsequently the mirroring
of the data disks.
Log-on
When the nodes have powered up, log on using the administrator account. There
should be no password with the user name administrator. Note: this is not always the
case depending on the clone that was used on the APG, sometimes the password is set
to administrator1. Wait for 3 mins for all the resources to come online and then start
the cluster administrator.
Spchange Slide
SPChange is the command used to change the site specific parameters on the APG.
The command is first run on the PDC with the BDC in a shutdown state. To run the
command on the BDC the PDC must be up and running. The command spchange
should only be run on a fully functional APG
spchange -n [PDC Node Name] -ip [PDC IP-address] -s [Subnet mask] -g [Default
gateway] -d [Domain name] -cn [Cluster name] -cip [Cluster IP-address]
spchange -n [BDC Node Name] -ip [BDC IP-address] -s [Subnet mask] -g [Default
gateway] -d [Domain name]
The Server Manager on Win NT is used to verify the parameters entered during
spchange are correct. If other computers appear on the list they can be removed.
Node A should be the primary and Node B the Backup. If both nodes appear as
backups then the command spchange may need to be run again as this suggests that
the Node Names were entered in lowercase during spchange
From this GUI it is possible to promote the BDC to PDC and also to synchronise the
domains.
The MMC is the Microsoft Management Console which provides an interface for
various Microsoft and 3rd party snap-ins. These application snap-ins provide the actual
functionality of the applications but the MMC is used to assist in the administration of
windows based systems.
For the APG 2 snap-ins that need to be added are the IIS (Internet Information
Server), used for setting up the default FTP site and administrating virtual directories.
The second snap-in is the MTS (Microsoft Transaction Server) is used to assist in
remote administration of the APG.
The Guest account is not as important as the Administrator as the Guest account is
disabled by default.
The Administrator account user name and password need to be changed as the
Administrator account is a special type of account that cannot be locked out by
entering an incorrect password. This therefore has security implications
IPN
Before starting the test procedures the crossover cable needs to be removed from the
LAN port and the LAN connection inserted. The nodes then need to be rebooted and
when they come back up the cluster administrator needs to be checked to make sure
all resources have come back online
Print the RP Bus Branch Index and physical and logical addresses
<DPRBP;
RP BUS BRANCH INDEX
LRPBI PRPBI ADDRESS SUIT RESERVED BW ALLOCATED BWTIME SLOTS
0 0 RP-0&&-31 0 10647 13
1 1 RP-32&&-63 0 10647 13
2 2 RP-64&&-95 0 10647 13
3 3 RP-96&&-127 0 10647 13
4 31 RP-128&&-159 0 - -
. . . . . .
. . . . . .
30 30 RP-960&&-991 0 - -
31 4 IPN-0&-1 102400 103194 126
END
<DPRBC:LRPBI=31,PRPBI=XX;
Defines the link between the IPN on physical index XX to the logical index 31
<DPRBC:LRPBI=31,RESBWM=100;
Sets the reserve bandwidth in Mb/s for the link.
A DHCP (Dynamic Host Control Protocol) Server issues the IP addresses to the
IPNA. These are allocated using a unique identifier. The hex number for the ASCII
string eg. 49 50 4e 41 30 30 = IPNA00. This means that if the IPNA’s are replaced the
addressing remains the same for the new card. IPNA identities of an even number are
connected to CP-A and the odd ones are connected to CP-B.
Configuration of the DHCP server is done through the Local Console. Start,
Programs, Admin Tools, DHCP Manager. It must be configured on both nodes and the
requires a failover or manual start to configure the server on the passive node.
1. Creating a Scope
2. Defining a Reservation
3. Selecting a Boot File
4. Defining Host Name
5. Repeat for CP-B (odd) IPN’s
The first step is to create the scope of the addresses. This defines the operating range
of IP addresses that will be used.
Step 3 is to select a boot file from the AP that is used by the IPNA.
Click Scope Address, Click on IPN 0 IP address, Properties, Option, from Unused
Options Select 067 Bootfile Name, Click Add, Value Type = boot.ipn0
Step 4 is to define the hostname. The Host Name is specified as the IP address of the
Node
Click Scope Address, DHCP Options, Scope, from Unused Options select 066 Boot
Serve Host Name, Click Add, Value = 192.168.169.1
Repeat for CP-B (odd) IPN’s Slide
Repeat the steps above for the BDC. To do this a failover or manual startup of the
service will need to take place
GO THROUGH SLIDE
Test of APG40
dsdls -a
Node Domain Application Conn-type Pid IP Address Port
cp0 JTP CLOCKSYNC OCP 0 0 85
cp0sb JTP ALCOSB OCP 0 0 86
apla MTAP CHS TCP/IP 292 127.0.0.1 1177
TCP/IP 127.0.0.1 1178
OCP 192.168.169.1 1179
OCP 192.168.170.1 1180
C:\> mml
<APAMP;
AP MAINTENANCE DATA
AP MAINTENANCE TABLE
AP IO ACTIVENODE LOCALIP1 LOCALIP2
1 YES A 192.168.169.128 192.168.170.128
END
<
The first step is to do the tests described above in the APM Only test, using the
commands
dsdls –a
mml apamp;
In addition to these tests there are a couple more that need to be performed.
cpdtest -c IO Device in CP, Test: Used to test and list the attributes of the AP-CP
comms
(-c This option specifies that the attributes of all connected CP-AP channels
are to be listed)
cpdtest -c
CHAN STATE CPSIDE NETWORK HANDLE KEY
5 Connected EX Lan 1 25 29
sybfp SYSTEM FUNCTIONS, BACKUP FILES PRINT: Lists all the reload
files in the CPF
The APG40 allows access to both the APG, CP and various commands through the
use of defined users.
A different user can be setup for each operator accessing the APG40, and an
individual password can be assigned to these users.
The users are made members of local groups which provide different functionality
and access to commands and resources in the APG domain.
Standard Groups
There are a number of different groups in the APG. If no applications are installed on
the APG then the standard groups defined are as follows
• Administrators
• Account Operators
• FTPUSRG
• EVENTVIEWERG
• SECUREADMG
• ACSUSRG
• ACSADMG
• Domain Users
Application Groups
If the APIO application is installed then the following localgroups will be defined
• CPUSRG
• CPADMG
• APLOCG
• CPSUSRG
• CPSADMG
• FMSUSRG
• FMSADMG
• MASUSRG
• MASADMG
• MCSUSRG
• MCSADMG
• Other groups
The Other Groups mentioned here is for groups relating to other applications that are
installed. For example if STS was installed then there would be an STSUSRG and
STSADMG
• Administrator
The administrator account is a special type of account as it cannot be locked out. For
this reason it is recommended that both the name of the administrator account and the
password be changed to something complex.
• Guest
The guest account is built in and therefore not possible to remove. It is disabled by
default but even so it should also be renamed and a complex password issued for it
• Service Accounts
There are 2 service accounts on the APG. These accounts have automatically
generated names which consist of a random array of numbers and characters in upper
and lower case. The password is also automatically generated.
One is responsible for each node and is required to automatically start selected
services on the APG, for example the cluster service, ACS_FCH_Server, LBB
services.
• IUSR_<Machine Name>
This account is used by the MS IIS for anonymous FTP logons. The APG does not
allow anonymous logons but this feature is used between the AP and CP
• IWAM_<Machine Name>
This account is used by the MS IIS for process isolation. However this feature is not
used on the APG and therefore this account should be disabled.
SKIP
Creating a new User
Creating a new local user on the APG can be done either through the PCAnywhere
GUI or by using the CLI.
PCAnywhere
Start, Programs, Admin Tools, User Manager for Domains, User Add
This will create a new user but to be able to do anything on the APG (even log on) the
user must be part of some local groups
CLI
There are 16 AP groups and 16 CP groups. These groups can be associated together so
that a user will only need 1 username and password to be able to log into both.
The association can also define the permissions that a user has and whether the
permissions in the AP and the same as those in the CP
A master user domain is used to administrate user and their associated groups across
multiple APG’s
A MUD can be any computer running Win NT 4 server as there are no additional
software requirements
User defined in a MUD will have access across all APG’s in the domain.
There are therefore 2 domains, the AP Domain and the MUD Domain. To be able to
communicate between them a trust Relationship needs to be setup
Start, Programs, Admin Tools, User Manager for Domains, Policies, Trust
Relationship
A trust relationship is setup between the APG and the MUD. It is enough that the APG
trusts the MUD but it is not necessary for the MUD to trust the APG
On MUD
Add the APG as a trusting domain, give it a name and password
On APG
Add the APG as a trusted domain, give it a name and enter the password from above
On the MUD
Once the trust relationship is set, Global groups are now added. Since the MUD is
trusted by the APG, Local Groups on the APG can be associated with the Global
Groups on the MUD
The Global and Local groups are mapped in a 1-1 relationship, it is possible to map 1-
Many but this only complicates things.
A global group is setup on the MUD and from there Local Groups are created of the
same name as the groups on the APG. Once done the mapping can be started
Users are then defined in the MUD and added to the different Local Groups on the
MUD
Chapter 3 – APG40 Recovery Procedures
AP Restore
There are a number of different choices when performing a burrestore. Note that the
command burrestore on its own swaps the drive letters so that the D becomes C and
the C becomes D. The system auto-boots in C so the restore is performed
Disaster Recovery
Follow the OPI : AP System Disaster Recovery – Used when 1 or both nodes can’t
boot from the hard disk.
The OPI must be followed and the following is only a brief description of the steps.
Disable Flashdisk
On the PDC
Flash disk disabled from the BIOS (newer systems) or by a dipswitch (older
systems)
AP is rebooted so now boots from Hard Disk
D:\ is now C:\
Steps Repeated for PDC
Convert Partition
Convert c:/fs:ntfs
Reinstall Burserver
Stop Cluster Server
cd “Program Files”\AP\ACS\bin
burserver.exe –service
copy c:\”program files”\pkzip\pkzip25.exe
BDC
Convert file system
Reinstall burserver
Recreate F:\ partition
Backup of the D: drive
Zap Data Disks by pressing CTRL-A at the blue “Adaptec I2O” symbol and
follow OPI
Node Reconnection
Set ACS_FCH_Server to manual startup in services
Shutdown node and reconnect all cables in the correct order
Set Cluster Server to automatic startup in services
Synchronise the PDC and BDC
Net accounts /sync
Intergrate mirrored Data Disks
fcc_integrate other
Set ACS_FCH_Server to automatic startup in services
Quorum Restore
Enforces Cluster Unity preventing split brain syndrome. Should nodes not be
able to communicate then only nodes that can communicate with the Quorum
resource can form a Cluster.
Procedure
Printouts
The printout has no sign of any disk corruption or inconsistencies
The printout suggested disk corruption or inconsistencies
The printout was either
o “The system cannot find the drive specified”
OR
o “Cannot open volume for direct access”
3: The printout was either “The system cannot find the drive specified” OR
“Cannot open volume for direct access”
This OPI cannot fix the problem
Procedure
Reboot BDC
Enter Raid Bios by pressing CTRL-A at the blue Adaptec Symbol
The BDC will be left in the state for most of the procedure
Individual boards on the AP are not normally changed. An entire node is replaced
Node changes are carried out when the following conditions are applicable
The alarm AP Fault is present, or a work order is received
The node is faulty
Procedure
APG’s are all delivered with values setup for a single AP. To add a second AP there
values therefore have to be changed
Before looking at the procedure to set these values we will look at the different
configuration options.
CP-A CP-B
RPHM -A RPHM -B
LAN-0
LAN-1
AP-1 AP-2
Here both APG’s share the same network, this is they use the same IPN bus to connect
to the CP.
Each RPHM will require ?? IPNX and 2 IPNA (the IPNA needs 1 for each LAN
connection)
CP-A CP -B
RPHM-A RPHM-B
LAN -0
LAN -1
LAN -2
LAN -3
AP -1 AP -2
Here both APG’s will have their own network, this is they will use different IPN buses
to connect to the CP.
Each RPHM will require ?? IPNX and 4 IPNA (the IPNA needs 1 for each LAN
connection)
Procedure
Procedure on AP 1
1: Edit the HOSTS file on Passive then Active Nodes
Used to ensure the correct Nodes names and IP Addresses
192.168.169.3 ap2a-l1 ap2a
192.168.170.3 ap2a-l2 ap2a
192.168.169.4 ap2b-l1 ap2b
192.168.170.4 ap2b-l2 ap2b
Procedure on AP 2
2: Change the APNodeNumber parameter using a Soft Function Change
Parameter File: c:\Program Files\AP\ACS\conf\CXC13791.par
Parameter: ACS_ALHEXECBIN_apNodeNumber: 2;
3: Disable DHCP
Start, Settings, Control Panel, Services, Microsoft DHCP Server, Startup
Type, Disable
Select from the Adapter Pull down menu, the first “Intel 21143 Based 10/100 mbps
Ethernet Controller”
The IP address should currently be set to 192.168.169.1 or 192.168.169.2.
If the IP address is 192.168.169.1 change it to 192.168.169.3. (Node A)
If the IP address is 192.168.169.2 change it to 192.168.169.4. (Node B)
Select from the Adapter Pull down menu, the second “Intel 21143 Based 10/100 mbps
Ethernet Controller”
The IP address should currently be set to 192.168.170.1 or 192.168.170.2.
If the IP address is 192.168.170.1 change it to 192.168.170.3. (Node A)
If the IP address is 192.168.170.2 change it to 192.168.170.4. (Node B)
8: Verify Changes
Phaprint ACS_ALHEXECCIN_apnodenumber
Type c:\winnt\system32\drivers\etc\hosts
Ipconfig /all
Procedure on CP
9: Define AP2 in CP
mml apami:AP=2,TYPE=0;
apamp;
Procedure
CP Fault Free
ALLIP
Backup IOG
SUBUP
Define IPN
OCINI:IPN=0;
OCIEI:IPN=0,EQM=OCITS-0;
OCINI:IPN=1;
OCIEI:IPN=1,EQM=OCITS-1;
Define AP in CP
APAMI:AP=1,TYPE=0;
APAMP;
Deblock IPN
OCIBE:IPN=0;
OCIBE:IPN=1;
Backup CP and AP
SYBUP
Burbackup
Chapter 5 – APG40 Antivirus Configuration
Objectives Slide
The APG will be delivered virus free and if an initial backup is delivered with the
APG it is up to the operator to keep this backup free from viruses.
GO THROUGH SLIDES
Network Configuration
Configure the network the th APG40 is connected to so that it is only
accessible to the O&M or Network Element traffic
GO THROUGH SLIDES
The following procedure shows to download the updates via the LAN using a simple
pass through proxy protocol.