Tivoli System Automation

Getting Started Guide

The purpose of this guide is to introduce Tivoli System Automation for Multiplatforms and provide a quick-start, purpose-driven approach to users that need to use the software, but have little or no past experience with it.

Share:

Frank Goytisolo (fgoytiso@ca.ibm.com), IT Specialist, IBM

Photo of Frank GoytisoloFrank Goytisolo is a licensed Professional Engineer, and IT Specialist at IBM's North American Lab Services team. He is a DB2 expert and has extensive experience with Tivoli System Automation and IBM's Smart Analytics System clusters (BCUs). Frank has also developed High Availability and Disaster Recovery solutions for several clients using TSA.



17 May 2012

Introduction

The purpose of this guide is to introduce Tivoli® System Automation for Multiplatforms and provide a quick-start, purpose-driven approach to users that need to use the software, but have little or no past experience with it.

This guide describes the role that TSA plays within IBM’s Smart Analytics System solution and the commands that can be used to manipulate the application. Further, some basic problem diagnosis techniques will be discussed, which may help with minor issues that could be experienced during regular use.

When the Smart Analytics system is built with High Availability, TSA is automatically installed and configured by the ATK. Therefore, this guide will not describe how to install or configure a TSA cluster (domain) from scratch, but rather how to manipulate and work with an existing environment. To learn to define a cluster of servers, please refer to the References appendix for IBM courses that are available.


Terminology

It is advisable to become familiar with the following terms, since they are used throughout this guide. It will also help you become familiar with the scopes of the different components within TSA.

Table 1. Terminology
TermDefinition
Peer Domain: A cluster of servers, or nodes, for which TSA is responsible
Resource: Hardware or software that can be monitored or controlled. These can be fixed or floating. Floating resources can move between nodes.
Resource group: A virtual group or collection of resources
Relationships: Describe how resources work together. A start-stop relationship creates a dependency (see below) on another resource. A location relationship applies when resources should be started on the same or different nodes.
Dependency: A limitation on a resource that restricts operation. For example, if resource A depends on resource B, then resource B must be online for resource A to be started.
Equivalency: A set of fixed resources of the same resource class that provide the same functionality
Quorum: A cluster is said to have quorum when there it has the capability to form a majority within its nodes. The cluster can lose quorum when there is a communication failure, and sub-clusters form with an even number of nodes.
Nominal State: This can be online or offline. It is the desired state of a resource, and can be changed so that TSA will bring a resource online or shut it down.
Tie Breaker: Used to maintain quorum, even in a split-brain situation (as mentioned in the definition of quorum). A tie-breaker allows sub-clusters to determine which set of nodes will take control of the domain.
Failover: When a failure occurs (typically hardware), which causes resources to be moved from one machine to another machine, the resources are said to have “failed over”

Getting Started

The purpose of TSA in the Smart Analytics system is to manage software and hardware resources, so that in the event of a failure, they can be restarted or moved to a backup system. TSA uses background scripts to check the status of processes and ensure that everything is working ok. It also uses “heart-beating” between all the nodes in the domain to ensure that every server is reachable. Should a process fail the status check, or a node fails to respond to a heartbeat, appropriate action will be taken by TSA to bring the system back to its nominal state.

Let’s start with the basics. In a Smart Analytics System, the TSA domain includes the DB2 Admin node, the Data nodes, and any Standby/backup nodes. The management server is not part of the domain and TSA commands will not work there. Further, all TSA commands are run as the root user.

The first thing you want to do is check the status of the domain, and start it if required:

    # lsrpdomain
    Name      OpState RSCTActiveVersion MixedVersions TSPort GSPort 
    bcudomain Online  2.5.3.3           No            12347  12348

In this case it’s already started, but if OpState would show “Offline”, then the command to start the domain is,

startrpdomain bcudomain

Notice that the domain name is bcudomain, and it is required for the start command. Likewise, if you want to stop the domain, the command is,

stoprpdomain bcudomain

If TSA is in an unstable state, you can also forcefully shut down the domain using the -f parameter in the stoprpdomain command. However, this is typically not recommended:

stoprpdomain -f bcudomain

You should not stop a domain until all your resources have been properly shut down. If your system uses GPFS to manage the /db2home mount, then you need to manually unmount the GPFS filesystems before you can stop the TSA domain using the following command,

/usr/lpp/mmfs/bin/mmunmount /db2home

Next, you’ll want to check the status of the nodes in the domain. The following command will do this:

        # lsrpnode
        Name      OpState RSCTVersion 
        beluga006 Online  2.5.3.3     
        beluga008 Online  2.5.3.3     
        beluga007 Online  2.5.3.3

You can see that we have 3 nodes in this domain: beluga006, beluga007, and beluga008. This also shows their state. If they are Online, then TSA can work with them. If they are Offline, they are either turned off or TSA cannot communicate with them (and thus unavailable). Nodes don’t always appear in the order that you would expect, so be sure to scan the whole output (in this case, beluga008 shows up before beluga007).


Resource Groups

After you have verified that the Domain is started, and all your nodes are Online, you will want to check the status of your resources. TSA manages all resources through resource groups. You cannot start a resource individually through TSA. When you start a resource group however, it will start all resources that belong to that group.

To check the status of your DB2 resources, use the hals command. This gives you a summary of all nodes in the peer domain, including their primary and backup locations, current location, and failover state.

+===============+===============+===============+==================+==================+===========+
|  PARTITIONS   |    PRIMARY    |   SECONDARY   | CURRENT LOCATION | RESOURCE OPSTATE | HA STATUS |
+===============+===============+===============+==================+==================+===========+
| 0             | dwadmp1x      | dwhap1x       | dwadmp1x         | Online           | Normal    |
| 1,2,3,4       | dwdmp1x       | dwhap1x       | dwdmp1x          | Online           | Normal    |
| 5,6,7,8       | dwdmp2x       | dwhap1x       | dwdmp2x          | Online           | Normal    |
| 9,10,11,12    | dwdmp3x       | dwhap1x       | dwhap1x          | Online           | Failover  |
| 13,14,15,16   | dwdmp4x       | dwhap1x       | dwdmp4x          | Online           | Normal    |
+===============+===============+===============+==================+==================+===========+

In this example, we see that the admin node is dwadmp1x since it holds partition 0. There are 4 data nodes in this system, and all are in Normal state except for data node 3. We can see that data node 3 is in Failover state and its current location is dwhap1x, the backup server.

The hals command is actually a summary of the complete output. For more detailed information about each resource, use the lssam command. The following output is an example of a cluster with the following nodes:

Admin node:   beluga006
Data node: 	  beluga007
Standby node: beluga008
# lssam | grep Nominal
Online IBM.ResourceGroup:SA-nfsserver-rg Nominal=Online
Online IBM.ResourceGroup:db2_bculinux_NLG_beluga006-rg Nominal=Online
        '- Online IBM.ResourceGroup:db2_bculinux_0-rg Nominal=Online
Online IBM.ResourceGroup:db2_bculinux_NLG_beluga007-rg Nominal=Online
        |- Online IBM.ResourceGroup:db2_bculinux_1-rg Nominal=Online
        |- Online IBM.ResourceGroup:db2_bculinux_2-rg Nominal=Online
        |- Online IBM.ResourceGroup:db2_bculinux_3-rg Nominal=Online
        '- Online IBM.ResourceGroup:db2_bculinux_4-rg Nominal=Online

Notice that the full output was grepped to “Nominal”. This is a trick to shorten the output so that we only see the Nominal states, and soon you will see that it can get quite long otherwise.

Let’s step through the above output:

Online IBM.ResourceGroup:SA-nfsserver-rg Nominal=Online

This first line tells us that we have a resource group named SA-nfsserver-rg and it is Online. The Nominal state is also Online, so it is working as expected. By the name, we can tell that this resource group manages the NFS server resources. Typically, this should always be online.

Online IBM.ResourceGroup:db2_bculinux_NLG_beluga006-rg Nominal=Online
        '- Online IBM.ResourceGroup:db2_bculinux_0-rg Nominal=Online

Next we have a resource group called db2_bculinux_NLG_beluga006-rg. This is the resource group belonging to the Admin node. We know that because beluga006 is the hostname for the Admin node. Here, we have 1 DB2 partition (the coordinator partition). For every partition, we define a resource group. You’ll see why shortly. The resource group for the admin partition, partition 0, is called db2_bculinux_0-rg.

Online IBM.ResourceGroup:db2_bculinux_NLG_beluga007-rg Nominal=Online
        |- Online IBM.ResourceGroup:db2_bculinux_1-rg Nominal=Online
        |- Online IBM.ResourceGroup:db2_bculinux_2-rg Nominal=Online
        |- Online IBM.ResourceGroup:db2_bculinux_3-rg Nominal=Online
        '- Online IBM.ResourceGroup:db2_bculinux_4-rg Nominal=Online

Lastly, we have our data partition group, db2_bculinux_NLG_beluga007-rg. Every data partition in a Balanced Warehouse has 4 partitions, and they can be easily seen here.

Now, let us examine the full lssam output. Try to find each of the lines from the grepped output in the full output:

# lssam
Online IBM.ResourceGroup:SA-nfsserver-rg Nominal=Online
        |- Online IBM.AgFileSystem:shared_db2home
                |- Online IBM.AgFileSystem:shared_db2home:beluga006
                '- Offline IBM.AgFileSystem:shared_db2home:beluga008
        |- Online IBM.AgFileSystem:varlibnfs
                |- Online IBM.AgFileSystem:varlibnfs:beluga006
                '- Offline IBM.AgFileSystem:varlibnfs:beluga008
        |- Online IBM.Application:SA-nfsserver-server
                |- Online IBM.Application:SA-nfsserver-server:beluga006
                '- Offline IBM.Application:SA-nfsserver-server:beluga008
        '- Online IBM.ServiceIP:SA-nfsserver-ip-1
                |- Online IBM.ServiceIP:SA-nfsserver-ip-1:beluga006
                '- Offline IBM.ServiceIP:SA-nfsserver-ip-1:beluga008
Online IBM.ResourceGroup:db2_bculinux_NLG_beluga006-rg Nominal=Online
        '- Online IBM.ResourceGroup:db2_bculinux_0-rg Nominal=Online
                |- Online IBM.Application:db2_bculinux_0-rs
                   |- Online IBM.Application:db2_bculinux_0-rs:beluga006
                   '- Offline IBM.Application:db2_bculinux_0-rs:beluga008
                |- Online IBM.Application:db2mnt-db2fs_bculinux_NODE0000-rs
                    |- Online IBM.Application:db2mnt-db2fs_bculinux_NODE0000-rs:beluga006
                    '- Offline IBM.Application:db2mnt-db2fs_bculinux_NODE0000-rs:beluga008
                '- Online IBM.ServiceIP:db2ip_172_16_10_228-rs
                    |- Online IBM.ServiceIP:db2ip_172_16_10_228-rs:beluga006
                    '- Offline IBM.ServiceIP:db2ip_172_16_10_228-rs:beluga008
Online IBM.ResourceGroup:db2_bculinux_NLG_beluga007-rg Nominal=Online
        |- Online IBM.ResourceGroup:db2_bculinux_1-rg Nominal=Online
                |- Online IBM.Application:db2_bculinux_1-rs
                    |- Online IBM.Application:db2_bculinux_1-rs:beluga007
                    '- Offline IBM.Application:db2_bculinux_1-rs:beluga008
                '- Online IBM.Application:db2mnt-db2fs_bculinux_NODE0001-rs
                    |- Online IBM.Application:db2mnt-db2fs_bculinux_NODE0001-rs:beluga007
                    '- Offline IBM.Application:db2mnt-db2fs_bculinux_NODE0001-rs:beluga008
        |- Online IBM.ResourceGroup:db2_bculinux_2-rg Nominal=Online
                |- Online IBM.Application:db2_bculinux_2-rs
                    |- Online IBM.Application:db2_bculinux_2-rs:beluga007
                    '- Offline IBM.Application:db2_bculinux_2-rs:beluga008
                '- Online IBM.Application:db2mnt-db2fs_bculinux_NODE0002-rs
                    |- Online IBM.Application:db2mnt-db2fs_bculinux_NODE0002-rs:beluga007
                    '- Offline IBM.Application:db2mnt-db2fs_bculinux_NODE0002-rs:beluga008
        |- Online IBM.ResourceGroup:db2_bculinux_3-rg Nominal=Online
                |- Online IBM.Application:db2_bculinux_3-rs
                    |- Online IBM.Application:db2_bculinux_3-rs:beluga007
                    '- Offline IBM.Application:db2_bculinux_3-rs:beluga008
                '- Online IBM.Application:db2mnt-db2fs_bculinux_NODE0003-rs
                    |- Online IBM.Application:db2mnt-db2fs_bculinux_NODE0003-rs:beluga007
                    '- Offline IBM.Application:db2mnt-db2fs_bculinux_NODE0003-rs:beluga008
        '- Online IBM.ResourceGroup:db2_bculinux_4-rg Nominal=Online
                |- Online IBM.Application:db2_bculinux_4-rs
                    |- Online IBM.Application:db2_bculinux_4-rs:beluga007
                    '- Offline IBM.Application:db2_bculinux_4-rs:beluga008
                '- Online IBM.Application:db2mnt-db2fs_bculinux_NODE0004-rs
                    |- Online IBM.Application:db2mnt-db2fs_bculinux_NODE0004-rs:beluga007
                    '- Offline IBM.Application:db2mnt-db2fs_bculinux_NODE0004-rs:beluga008

Let us take a look at the NFS resource group:

Online IBM.ResourceGroup:SA-nfsserver-rg Nominal=Online
        |- Online IBM.AgFileSystem:shared_db2home
             |- Online IBM.AgFileSystem:shared_db2home:beluga006
             '- Offline IBM.AgFileSystem:shared_db2home:beluga008

The first line was what we had seen before (lssam | grep Nom). Now, we can see what resources actually form the resource group. This first resource is of type AgFileSystem and represents the db2home mount. We can see that it can exist on beluga006 and beluga008, and that it is Online in beluga006 and Offline in beluga008.

Similarly, for the admin node, we can now see the individual resources:

Online IBM.ResourceGroup:db2_bculinux_NLG_beluga006-rg Nominal=Online
        '- Online IBM.ResourceGroup:db2_bculinux_0-rg Nominal=Online
                |- Online IBM.Application:db2_bculinux_0-rs
                   |- Online IBM.Application:db2_bculinux_0-rs:beluga006
                   '- Offline IBM.Application:db2_bculinux_0-rs:beluga008

The first two lines were part of the previous grepped output, but now we can see an Application resource. You can see similar results for the data node and each of its 4 data partitions. The reason that each of these resources exist on two nodes (beluga006 and beluga008) is for high availability. If beluga006 were to fail, TSA will move all those resources that are currently Online there to beluga008. Then, you would see that they are Offline in beluga006, and Online in beluga008. You can see how this output is useful to determine on which nodes the resources exist.

The lssam command also shows Equivalencies as part of the output. I will include it for the sake of completion, but we will discuss this later on:

Online IBM.Equivalency:SA-nfsserver-nieq-1
        |- Online IBM.NetworkInterface:bond0:beluga006
        '- Online IBM.NetworkInterface:bond0:beluga008
Online IBM.Equivalency:db2_FCM_network
        |- Online IBM.NetworkInterface:bond0:beluga006
        |- Online IBM.NetworkInterface:bond0:beluga007
        '- Online IBM.NetworkInterface:bond0:beluga008
Online IBM.Equivalency:db2_bculinux_0-rg_group-equ
        |- Online IBM.PeerNode:beluga006:beluga006
        '- Online IBM.PeerNode:beluga008:beluga008
Online IBM.Equivalency:db2_bculinux_1-rg_group-equ
        |- Online IBM.PeerNode:beluga007:beluga007
        '- Online IBM.PeerNode:beluga008:beluga008
Online IBM.Equivalency:db2_bculinux_2-rg_group-equ
        |- Online IBM.PeerNode:beluga007:beluga007
        '- Online IBM.PeerNode:beluga008:beluga008
Online IBM.Equivalency:db2_bculinux_3-rg_group-equ
        |- Online IBM.PeerNode:beluga007:beluga007
        '- Online IBM.PeerNode:beluga008:beluga008
Online IBM.Equivalency:db2_bculinux_4-rg_group-equ
        |- Online IBM.PeerNode:beluga007:beluga007
        '- Online IBM.PeerNode:beluga008:beluga008
Online IBM.Equivalency:db2_bculinux_NLG_beluga006-equ
        |- Online IBM.PeerNode:beluga006:beluga006
        '- Online IBM.PeerNode:beluga008:beluga008
Online IBM.Equivalency:db2_bculinux_NLG_beluga007-equ
        |- Online IBM.PeerNode:beluga007:beluga007
        '- Online IBM.PeerNode:beluga008:beluga008

The lssam command also lets you limit the output to a particular resource group, with the –g option:

# lssam –g SA-nfsserver-rg
Online IBM.ResourceGroup:SA-nfsserver-rg Nominal=Online
        |- Online IBM.AgFileSystem:shared_db2home
                |- Online IBM.AgFileSystem:shared_db2home:beluga006
                '- Offline IBM.AgFileSystem:shared_db2home:beluga008
        |- Online IBM.AgFileSystem:varlibnfs
                |- Online IBM.AgFileSystem:varlibnfs:beluga006
                '- Offline IBM.AgFileSystem:varlibnfs:beluga008
        |- Online IBM.Application:SA-nfsserver-server
                |- Online IBM.Application:SA-nfsserver-server:beluga006
                '- Offline IBM.Application:SA-nfsserver-server:beluga008
        '- Online IBM.ServiceIP:SA-nfsserver-ip-1
                |- Online IBM.ServiceIP:SA-nfsserver-ip-1:beluga006
                '- Offline IBM.ServiceIP:SA-nfsserver-ip-1:beluga008

With the Smart Analytics System, some new commands were introduced to make it easier to monitor and use TSA with DB2:

Table 2. Useful Commands
CommandDefinition
hals: shows HA status summary for all db2 partitions
hachknode shows the status of the node in the domain and details about the private and public networks
hastartdb2 start db2 partition resources
hastopdb2 stop db2 partition resources
hafailback moves partitions back to the primary machine specified in the primary_machine argument
Equivalency: A set of fixed resources of the same resource class that provide the same functionality
hafailover moves partitions off of the primary machine specified in the primary_machine argument to it is standby
hareset attempt to reset pending, failed, stuck resource states

Stopping and Starting Resources

If you want to stop or start the DB2 service, you need to stop the respective DB2 resource groups using TSA commands. TSA will then start or stop DB2.

The command to do this is chrg. To stop a resource group named db2_bculinux_NLG_beluga007, issue the command,

chrg –o offline –s “Name == ‘db2_bculinux_NLG_beluga007’”

Similarly, to start the resource group

chrg –o online –s “Name == ‘db2_bculinux_NLG_beluga007’”

You can also stop/start all resources at the same time:

chrg –o online –s “1=1”

The Smart Analytics System also has some pre-configured commands:

hastartdb2 and hastopdb2

These two commands, however, are specific to DB2 and if there has been customization to TSA, they may not stop/start all resources.

If TSA has pre-configured rules/dependencies, they will ensure that resources are stopped and started in the correct order. For example, DB2 resources that depend on NFS will not start if the NFS share is Offline.


TSA Components

Now that you understand the basics of Tivoli System Automation, we can discuss some of the other components that it can manage.

Service IP

A service IP is a virtual, floating resource attached to a network device. Essentially, it is an IP address that can move from one machine to another, in the event of a failover. Service IPs play a key role in a highly available environment. Because they move from a failed machine to a standby, they allow an application to reconnect to the new machine using the same IP address – as if the original server had simply restarted.

The following command will allow you to view what service IPs have been configured for your system.

# lsrsrc -Ab IBM.ServiceIP
    Resource Persistent and Dynamic Attributes for IBM.ServiceIP
    resource 1:
    	Name              = "db2ip_10_160_20_210-rs"
    	ResourceType      = 0
    	AggregateResource = "0x2029 0xffff 0x414c690c 0x7cc2abfa 0x919b42d5 0xbf62ab75"
    	IPAddress         = "10.160.20.210"
    	NetMask           = "255.255.255.0"
    	ProtectionMode    = 1
    	NetPrefix         = 0
    	ActivePeerDomain  = "bcudomain"
    	NodeNameList      = {"t6udb3a"}
    	OpState           = 2
    	ConfigChanged     = 0
    	ChangedAttributes = {}
    resource 2:
    	Name              = "db2ip_10_160_20_210-rs"
    	ResourceType      = 0
    	AggregateResource = "0x2029 0xffff 0x414c690c 0x7cc2abfa 0x919b42d5 0xbf62ab75"
    	IPAddress         = "10.160.20.210"
    	NetMask           = "255.255.255.0"
    	ProtectionMode    = 1
    	NetPrefix         = 0
    	ActivePeerDomain  = "bcudomain"
    	NodeNameList      = {"t6udb1a"}
    	OpState           = 1
    	ConfigChanged     = 0
    	ChangedAttributes = {}
    resource 3:
    	Name              = "db2ip_10_160_20_210-rs"
    	ResourceType      = 1
    	AggregateResource = "0x3fff 0xffff 0x00000000 0x00000000 0x00000000 0x00000000"
    	IPAddress         = "10.160.20.210"
    	NetMask           = "255.255.255.0"
    	ProtectionMode    = 1
    	NetPrefix         = 0
    	ActivePeerDomain  = "bcudomain"
    	NodeNameList      = {"t6udb1a","t6udb3a"}
    	OpState           = 1
    	ConfigChanged     = 0
    	ChangedAttributes = {}

The above example shows three resources with the same name, db2ip_10_160_20_210-rs. The NodeNameList parameter tells us which node(s) the resource is referring to. The first resource has Opstate set to 2, which tells us that this is where the service IP is currently pointing (it is also the primary location of the resource). The second resource has Opstate 1, which tells us that this is the backup/standby node. The third resource contains both nodes in its NodeNameList parameters, and this tells TSA that this is a floating resource between those two nodes.

Application Resources

TSA manages resources using scripts. Some scripts are built in (and part of TSA), such as those for controlling DB2. These scripts are responsible for starting, stopping and monitoring the application. Sometimes it can be useful to understand these scripts, or even edit them for problem diagnosis. To find out where they are located, we use the lsrsrc command, which provides us with the complete configuration of a particular resource.

Following is an example:

# lsrsrc -Ab IBM.Application
resource 12:
  Name                  = "db2_dbedw1da_8-rs"	
  ResourceType          = 1
  AggregateResource     = "0x3fff 0xffff 0x00000000 0x00000000 0x00000000 0x00000000"
  StartCommand          = "/usr/sbin/rsct/sapolicies/db2/db2V97_start.ksh dbedw1da 8"
  StopCommand           = "/usr/sbin/rsct/sapolicies/db2/db2V97_stop.ksh dbedw1da 8"
  MonitorCommand        = "/usr/sbin/rsct/sapolicies/db2/db2V97_monitor.ksh dbedw1da 8"
  MonitorCommandPeriod  = 60
  MonitorCommandTimeout = 180
  StartCommandTimeout   = 330
  StopCommandTimeout    = 140
  UserName              = "root"
  RunCommandsSync       = 1
  ProtectionMode        = 1
  HealthCommand         = ""
  HealthCommandPeriod   = 10
  HealthCommandTimeout  = 5
  InstanceName          = ""
  InstanceLocation      = ""
  SetHealthState        = 0
  MovePrepareCommand    = ""
  MoveCompleteCommand   = ""
  MoveCancelCommand     = ""
  CleanupList           = {}
  CleanupCommand        = ""
  CleanupCommandTimeout = 10
  ProcessCommandString  = ""
  ResetState            = 0
  ReRegistrationPeriod  = 0
  CleanupNodeList       = {}
  MonitorUserName       = ""
  ActivePeerDomain      = "bcudomain"
  NodeNameList          = {"d8udb11a","d8udb3a"}
  OpState               = 1
  ConfigChanged         = 0
  ChangedAttributes     = {}
  HealthState           = 0
  HealthMessage         = ""
  MoveState             = [32768,{}]
  RegisteredPID         = 0

Some of the more common and useful attributes are described in Table 3.

Table 3. Resource Attributes
AttributeDefinition
ResourceType: Indicates whether the resource is allowed to run on multiple nodes, or a single node. A fixed resource is identified with a ResouceType value of 0, and a floating resource has a value of 1.
StartCommand: Specifies the command to be run when the resources is started
StopCommand: Specifies the command to be run when the resource is stopped
MonitorCommand: Specifies the command to be run when the resource is being monitored. This happens on a regular interval, and you will likely see this command often when you run the “ps –ef” command.
UserName: The userid that TSA will use to start this resource
NodeNameList: Indicates on which nodes the resource is allowed to run. This is an attribute of an RSCT resource.
OpState: Specifies the operational state of a resource or a resource group. The valid states are,
0 - UNKNOWN
1 - ONLINE
2 - OFFLINE
3 - FAILED_OFFLINE
4 - STUCK_ONLINE
5 - PENDING_ONLINE
6 - PENDING_OFFLINE

Network Resources

Every machine typically has an Ethernet adaptor, with a configured network address. TSA is aware of this and you can see how they have been configured with the lsrsrc command. For example,

# lsrsrc -Ab IBM.NetworkInterface
    resource 1:
       	Name             = "en0"
       	DeviceName       = ""
       	IPAddress        = "172.22.1.217"
       	SubnetMask       = "255.255.252.0"
       	Subnet           = "172.22.0.0"
       	CommGroup        = "CG1"
       	HeartbeatActive  = 1
       	Aliases          = {}
       	DeviceSubType    = 6
       	LogicalID        = 0
       	NetworkID        = 0
       	NetworkID64      = 0
       	PortID           = 0
       	HardwareAddress  = "00:21:5e:a3:be:60"
       	DevicePathName   = ""
       	IPVersion        = 4
       	Role             = 0
       	ActivePeerDomain = "bcudomain"

Log Files

It is important to be aware of the log files that TSA actively writes to:

  1. History file – this logs the commands that were sent to TSA
    /var/ct/IBM.RecoveryRM.log2
  2. Error and monitor logs – these logs are simply the AIX and Linux system logs. They will show you the output of the start, stop, and monitor scripts as well as any diagnostic information coming from TSA. Although the system administrator can configure the location for these logs, they are typically located in the following locations,
    AIX:	/tmp/syslog.out
    Linux:	/var/log/messages

Command Reference

Table 4 describes the most common commands that a TSA administrator will use.

Table 4. Common TSA Commands
CommandDefinition
hals: Display HA configuration summary
hastopdb2: Stop DB2 using TSA
hastartdb2: Start DB2 using TSA
mkequ:Makes an equivalency resource
chequ:Changes a resource equivalency
lsequ: Lists equivalencies and their attributes
rmequ: Removes one or more resource equivalencies
mkrg: Makes a resource group
chrg: Changes persistent attribute values of a resource group (including starting and stopping a resource group)
lsrg: Lists persistent attribute values of a resource group or its resource group members
rmrg: Removes a resource group
mkrel: Makes a managed relationship between resources
chrel: Changes one or more managed relationships between resources
lsrel: Lists managed relationships
rmrel: Removes a managed relationship between resources
samcrl: Sets the IBM TSA control parameters
lssamctrl: Lists the IBM TSA controls
addrgmbr: Adds one ore more resources to a resource group
chrgmbr: Changes the persistent attribute value(s) of a managed resource in a resource group
rmrgmbr: Removes one or more resources from the resource group
lsrgreq: Lists outstanding requests applied against resource groups or managed resources
rgmbrreq: Requests a managed resource to be started or stopped, or cancels the request
rgreq: Requests a resource group to be started, stopped, or moved, or cancels the request
lssam: Lists the defined resource groups and their members in a tree format

Command Tips

Following are some useful commands with examples.

Show relationships/dependencies:

lsrel | sort

Show details for a specific relationship:

# lsrel -A b -s "Name = 'db2_bculinux_0-rs_DependOn_db2_bculinux_qp-rel'"
Managed Relationship 1:
        Class:Resource:Node[Source] = IBM.Application:db2_bculinux_qp
        Class:Resource:Node[Target] = {IBM.Application:db2_bculinux_0-rs}
        Relationship                = DependsOn
        Conditional                 = NoCondition
        Name                        = db2_bculinux_0-rs_DependOn_db2_bculinux_qp-rel
        ActivePeerDomain            = bcudomain
        ConfigValidity              =

Delete/remove a relationship

rmrel -s "Name like 'db2_bculinux_%-rs_DependsOn_db2_bculinux_0-rs-rel'"

Change a resource attribute:

chrsrc -s "Name=='<resource_name'>" <resource_type> attribute=value

Example:

chrsrc -s "Name=='db2ip_10_160_10_27-rs'" IBM.ServiceIP NetMask='255.255.255.0'

To save current SAMP policy information:

sampolicy –s /tmp/sampolicy.current.xml

To check if the policy in the input file is valid:

sampolicy –c /tmp/sampolicy.current.xml

To activate it:

sampolicy –a /tmp/sampolicy.current.xml

Troubleshooting

This section describes methods that can be used to determine the cause of a particular problem or failure. Though techniques vary depending on the type of problem, the following should be a good starting point for most issues.

Cheat Sheet for Resolving Common Problems

Courtesy of Larry Pay lpay@ca.ibm.com

Resolving FAILED OFFLINE status
A failed offline status will prevent you from setting the nominal status to ONLINE, so these must be resolved first and changed to OFFLINE before turning it back to ONLINE. Make sure that the Nominal status is showing OFFLINE before resolving it.

To resolve the Failed offline messages, use the resetrsrc command.

resetrsrc -s ‘Name = "db2whse_appinstance_01.abxplatform_server1"‘ IBM.Application
resetrsrc -s 'Name = "db2whse_appinstance_01.adminconsole_server1"' IBM.Application

Recovery from a failed failover attempt
Take all TSA resources offline. The lssam output should reflect “Offline” for all resources before you attempt to bring them back online. To reset NFS resources, use:

resetrsrc -s "Name like 'SA-nfsserver-%'" IBM.Application (if necessary)
resetrsrc -s "Name like 'SA-nfsserver-%'" IBM.ServiceIP (if necessary)

When testing goes wrong, you are often left with resources in various states such as online, offline, and unknown. When the state of a resource is unknown, before attempting to restart it, you must issue resetrsrc for that particular resource.

When you are restarting DB2, you must verify that all the resources are offline before attempting to bring them online again. You must also correct the db2nodes.cfg file. Make sure you have backup copies of db2nodes.cg and db2ha.sys.

NFS mounts stop functioning
In testing the NFS failover, we were able to move the server over successfully, but the existing NFS client mounts stopped functioning. We solved this problem by unmounting and remounting the NFS volume.

Resolving Binding=Sacrificed
To resolve this problem you have to look at the overall cluster and how its setup/defined. Issues that can and will cause this are types that will have a cluster-wide impact but not specifically affect one resource.

  1. Check for failed relationships by listing the relationships with the following command "lsrel -Ab", and then determine if one or more of the relationships relating to the failed resource group have not been satisfied.
  2. Check for failed equivalencies by listing them with the following command "lsequ -Ab" and then determine if one re more of the equivalencies have not been satisfied.
  3. Check your resource group attributes and look for anything that maybe set incorrectly, some of the commands to use are listed as follows:
    lsrg -Ab -g <resource_group_name>
    lsrsrc -s 'Name="failed_resource"' –Ab IBM.<class_of_failed_resource>
    lsrg -m -g <resource_group_name>
    samdiag -g <resource_group_name>
  4. Check for anything specific to your configuration that all of the sacrificed resources share in common, like a mount point, a database instance, a virtual IP.

Check hardware configuration:dmesg – check initialization errors
date – check server synchronization
ifconfig <device> – to check network adapters
netstat -I – to check network configuration
ps -ef | grep inetd – will provide a list of the running processes, including group and PID

Resource state is unknown
Try resetting the resource using the resetrsrc command:

resetrsrc -s "Name like 'db2_db2inst2_%'" IBM.Application
resetrsrc -s "Name like 'db2_db2inst2_%'" IBM.ServiceIP

Timeout values for resources
For the health query interval of each resource, use:

chrsrc -s 'Name like "db2_db2inst2%"' IBM.Application MonitorCommandPeriod=300

For the health query timeout, use:

chrsrc -s 'Name like "db2_db2inst2%"' IBM.Application MonitorCommandTimeout=290

For the resource startup script timeout, use

chrsrc -s 'Name like "db2_db2inst2%"' IBM.Application StartCommandTimeout=300

For the Resource Stop script timeout, use:

chrsrc -s 'Name like "db2_db2inst2%"' IBM.Application StopCommandTimeout=720

Recycling the automation manager
If the problem is most likely related to the automation manager, you should try recycling the automation manager (IBM.RecoveryRM) before contacting IBM support. This can be done using the following commands:

Find out on which node the RecoveryRM master daemon is running:

# lssrc -ls IBM.RecoveryRM | grep Master

On the node running the master, retrieve the PID and kill the automation manager:

# lssrc -ls IBM.RecoveryRM | grep PID
# kill -9 <PID>

As a result, an automation manager on another node in the domain will take over the master role, and proceeds with making automation decisions. The subsystem will restart the killed automation manager immediately.

Resolving lssam hangs
http://www-01.ibm.com/support/docview.wss?uid=swg21293701

Move to another node in the same HA group and see if you can run the lssam command. If you can, go back to the original node to see if you can now do the lssam command. If this still does not work, then run the following commands:

lssrc -ls IBM.RecoveryRM | grep -i master 
lssrc -ls IBM.GblResRM | grep -i leader

Make sure neither of the above command outputs return the “hanging” node and if so, then reboot just that node and see if the issue is resolved.

AVOID the following (DON’Ts)

  • Do not use rpower –a, or rpower on more than one node in the same HA group when SAMP HA is up and running.
  • Do not offline HA-NFS using a sudo command while logged in as the instance owner and while in the /db2home directory. HA-NFS will get stuck online, and the RecoveryRM daemon has to be killed on the master. If RecoveryRM will not start, reboot may be required.
  • Do not use ifdown to bring down a network interface. This will result in the eth (or en) device to be deleted from equivalency member and will require you to add the "eth" device (in Linux) or "en" device (in AIX) back into the network equivalency using chequ command
  • Do not manipulate any BW resources that are under active SAMP control.
    Turn automation off (samctrl –M T) before manipulating these BW resources.
  • Do not implement changes to the SA MP policy unless exhaustive testing of the HA test cases is completed.

Check the following frequently (DOs)

  • Ensure the /home and /db2home directories are always mounted before starting up a node.
  • Check for process ids that may be blocking stop, start and monitor commands.
  • Save backup copies of the db2nodes.cfg and db2ha.sys file.
  • Save the backup copies of the current SAMP policy before and after every SAMP change. Compare the current SAMP policy to the backup SAMP policy every time there is an HA incident.
  • Save backup copies of db2pd -ha output before and after every SAMP change. Compare the current db2pd outputs to the backup db2pd outputs every time there is an HA incident.
  • Save backup copies of the samdiag outputs.

Resources

Comments

developerWorks: Sign in

Required fields are indicated with an asterisk (*).


Need an IBM ID?
Forgot your IBM ID?


Forgot your password?
Change your password

By clicking Submit, you agree to the developerWorks terms of use.

 


The first time you sign into developerWorks, a profile is created for you. Information in your profile (your name, country/region, and company name) is displayed to the public and will accompany any content you post, unless you opt to hide your company name. You may update your IBM account at any time.

All information submitted is secure.

Choose your display name



The first time you sign in to developerWorks, a profile is created for you, so you need to choose a display name. Your display name accompanies the content you post on developerWorks.

Please choose a display name between 3-31 characters. Your display name must be unique in the developerWorks community and should not be your email address for privacy reasons.

Required fields are indicated with an asterisk (*).

(Must be between 3 – 31 characters.)

By clicking Submit, you agree to the developerWorks terms of use.

 


All information submitted is secure.

Dig deeper into Tivoli (service management) on developerWorks


static.content.url=http://www.ibm.com/developerworks/js/artrating/
SITE_ID=1
Zone=Tivoli, Tivoli (service management)
ArticleID=816299
ArticleTitle=Tivoli System Automation
publish-date=05172012