Sun Microsystems Logo
Products and Services
 
Support and Training
 
 

A  B  C  D  E  F  G  H  I  J  K  L  M  N  O  P  Q  R  S  T  U  V  W  X  Y  Z  
 
File Formatsnhfs.conf(4)


NAME

 nhfs.conf - Foundation Services configuration file

SYNOPSIS

 /etc/opt/SUNWcgha/nhfs.conf

DESCRIPTION

The nhfs.conf file contains configuration information for the Foundation Services such as the Cluster Membership Manager, Reliable NFS, and the Node State Manager. This file also provides cluster addressing and interface configuration information.

Configure this file on each node if you plan to install the Foundation Services manually. To manually configure the parameters in the nhfs.conf file, uncomment the parameters, that is, delete the comment mark (#) at the beginning of the line, and modify the value of the parameter. A description of each parameter is provided in the following sections.

When you have manually install the Foundation Services packages, copy the template file /etc/opt/SUNWcgha/nhfs.conf.template to the default location /etc/opt/SUNWcgha as nhfs.conf on each peer node. For each file, make the necessary modifications in a text editor.

Do not re-edit the nhfs.conf file when the cluster is running.

The nhfs.conf file format is ASCII. Parameters consist of a keyword followed by an equals (=) sign followed by the parameter value, of the form:

Keyword=Value

The following Keyword and Value parameters are supported.

COMMON PARAMETERS

Cluster.DataManagementPolicy

Define how the cluster behaves when the vice-master node starts up while the master node is down. You can select one of three data management policies. Values are Integrity, Availability, and Adaptability. The default value is Integrity.

Cluster.DataManagementPolicy=Integrity | Availability | Adaptability
Integrity

If there is a vice-master but no master in the cluster, the vice-master waits for the old master to rejoin the cluster before it takes the master role. This ensures that the cluster has the most up-to-date data. Choosing this value might reduce the availability of the cluster, but it prioritizes data integrity.

Availability

If there is a vice-master in the cluster but no master, the vice-master does not wait for the old master to rejoin the cluster before taking on the master role. The previous master is still off-line. Any data that is written to the new master while there is no vice-master will be lost. When the old master node comes back on line as the vice-master node, a full synchronization occurs between the two master-eligible nodes.

Adaptability

If there is a vice-master in the cluster but no master, the vice-master checks the disk synchronization state. If the state is not synchronized, that is the state returned by nhcmmstat is synchro:NEEDED, the vice-master waits for the master to come up. This is equivalent to the Integrity data management policy. If the state is synchronized, that is the state returned by nhcmmstat is synchro:READY, the vice-master is elected the new master. This is equivalent to the Availability data management policy.

Cluster.Master.ID

Specify the host part of the master node's floating IP address in decimal form. The floating address triplet for the node with the master role is calculated from the local network interface addresses, the netmask, and the value of this parameter. These addresses are IPv4 address.

The default value is 1.

Cluster.Master.ID=1

If the Cluster.Master.ID is 1 and the cluster network is set up as follows:

  • Netmask: 255.255.255.0

  • NIC0 subnet: 192.200.168.0

  • NIC1 subnet: 10.250.2.0

  • cgtp0 subnet: 192.200.175.0

The floating address triplet is as follows:

  • NIC0 floating address: 192.200.168.1

  • NIC1 floating address: 10.250.2.1

  • cgtp0 floating address: 192.200.175.1

This address is calculated similarly if you have IP addresses of any other class.

CMM.MasterLoss.Detection

Determine if the absense of a master node in the cluster must be detected by diskless and dataless nodes. Values are True and False. The default is False.

If CMM.MasterLoss.Detection is set to False, the diskless and dataless nodes are rebooted when they detect that there is no master node in the cluster for a period of more then 4-5 minutes. If CMM.MasterLoss.Detection is set to True, the diskless and dataless nodes do not reboot if there is no master node in the cluster. However, the diskless and dataless nodes cannot access directories that would be exported by the master node if such a node existed. You must ensure that diskless and dataless nodes can adapt to this situation.

Node.Domainid

Specify the domainid of the cluster. You must modify this parameter. There is no default value.

Node.Domainid=250
Node.NodeId

This parameter specifies the node ID of the current node. There is no default value.

Node.NIC0

The name of the first network interface, NIC0, used for CGTP. The default value is hme0.

Node.NIC0=hme0

This parameter could be a logical network interface, for example, hme0:2.

If you have not installed the CGTP patches and packages and you want to configure a single network link for your cluster, configure this parameter.

Node.NIC1

The name of the second network interface, NIC1, used for CGTP. The default value is hme1.

Node.NIC1=hme1

This parameter could be a logical network interface, for example, hme1:2.

If you have not installed the CGTP patches and packages and you want to configure a single network link for your cluster, do not configure this parameter.

Node.NICCGTP

The virtual interface used by CGTP. The default value is cgtp0.

Node.NICCGTP=cgtp0

If you have not installed the CGTP patches and packages and you want to configure a single network link for your cluster, do not configure this parameter.

Node.UseCGTP

Specify whether the CGTP is to be used or not. Values are True or False. The default value is True.

Node.UseCGTP=True

If you have not installed the CGTP patches and packages and you want to configure a single network link for your cluster, set this parameter to False.

Node.Type

Specify the type of node. Defining this parameter is mandatory. When you create the nhfs.conf file on each peer node, specify the type of node in this parameter. The Node.Type parameter can have one of the following values:

  • Diskfull--A master-eligible node

  • Dataless--A dataless node

  • Diskless--A diskless node

There is no default value. For more information on types of nodes, see the Netra High Availability Suite Foundation Services 2.1 6/03 Overview.

Node.RNFS.Installed

Specify whether Reliable NFS is installed on a node. Values are True or False. The default value is False.

Node.RNFS.Installed=True

Both master-eligible nodes must have the same value for this parameter.

DIRECT LINK PARAMETERS

To prevent a split brain situation, you can connect the serial ports of the master-eligible nodes and configure the following parameters. Split brain is a situation where the network fails and two master nodes are elected in the same cluster. Configure the following parameters if you have connected the serial ports of the master-eligible nodes.

Node.Direct-Link.Serial.Device

Specify the serial device, that is, the system's serial ports. There is no default value.

Node.Direct-Link.serial.Device=/dev/term/b 
Node.Direct-Link.Serial.Speed

Specify the serial line speed. Valid values are 38400, 57600, 76800, or 115200. The higher the value, the better the link.

There is no default value.

Node.Direct-Link.serial.Speed=115200
Cluster.Direct-Link.Backend

This parameter enables the direct link. If this parameter is not present in the nhfs.conf file, the direct link will not be used even if you have connected the serial ports of the master-eligible nodes.

The only value accepted by this parameter is serial:

Cluster.Direct-Link.Backend=serial
Cluster.Direct-Link.Heartbeat

Specify the number of seconds between two checks to detect a failure. Therefore, if one master-eligible node receives no heartbeat during the specified period of seconds, the node is alerted that there may be a problem.

There is no default value. For example, specify an interval of 20 seconds as follows:

Cluster.Direct-Link.Heartbeat=20

CMM PARAMETERS

CMM.IsEligible

Specify whether the node is master eligible. Values are True or False. The default value is False.

CMM.IsEligible=True
CMM.LocalConfig.Dir

Specify the directory where the configuration file, cluster_nodes_table, is located on each master-eligible node. There is no default value.

CMM.LocalConfig.Dir=/etc/opt/SUNWcgha
CMM.Miniconfig.Dir

Specify the directory where the configuration file, target.conf, is located on each master-eligible node. There is no default value.

CMM.Miniconfig.Dir=/etc/opt/SUNWcgha
CMM.Miniconfig.File

Specify the name of the local configuration file, target.conf. There is no default value.

CMM.Miniconfig.File=target.conf

For more information on the CMM parameters, see the nhcmmd(1m) man page.

RELIABLE NFS PARAMETERS

RNFS.StatdAlternatePath

Specify the alternate statd repository. For information about statd, see the statd(1m) man page.

You must modify this parameter. There is no default value.

RNFS.StatdAlternatePath=directory-path

where directory-path is the path to the statd directory. This directory must be on a replicated disk partition.

RNFS.StatdAlternatePath=/SUNWcgha/local
RNFS.Slice

Define the disk partitions that must be replicated by Reliable NFS. You must modify this parameter. The format for this parameter is:

RNFS.Slice.x=local-disk-partition local-bitmap-partition \
remote-disk-partition remote-bitmap-partition mount-flag
  • x is an integer (0, 1, 2) that distinguishes each partition to be replicated.

  • local-disk-partition and remote-disk-partition are the paths to the raw disk partitions on the local and remote nodes.

  • local-bitmap-partition and remote-bitmap-partition are the paths to the scoreboard bitmaps associated to local-disk-partition and remote-disk-partition respectively.

  • mount-flag is 0, 1 or 2:

    0

    Do not mount the partition.

    1

    Mandatory mount. If the partition is not mountable, stop with error.

    2

    Best effort mount. Try to mount the partition. If this is not possible, log an error and continue.

There is no default value.

RNFS.Slice.0=/dev/rdsk/c0t0d0s3 /dev/rdsk/c0t0d0s5 \
/dev/rdsk/c0t0d0s3 /dev/rdsk/c0t0d0s5 1
RNFS.Share

Describes the file systems to be shared. You must modify this parameter.

  • This parameter uses the same syntax as the share command. For more information, see the share(1m) man page.

  • Each share command must be on its own line. The first string of an RNFS.Share line must be either share or /usr/sbin/share.

  • Each RNFS.Share line must contain one share command line per partition to be exported by NFS.

  • Each RNFS.Share.x represents a partition to be shared and x is an integer (0, 1, 2) that distinguishes each partition to be replicated.

To ensure that the CMM behaves correctly, you must:

  • Grant superuser access for all the master-eligible nodes on the exported file system where the cluster_nodes_table file resides.

  • Use the addresses for the CGTP interface for the master-eligible nodes. If CGTP is not installed, use the NIC0 addresses.

There is no default value. In the following example, cgtp6 and cgtp7 represent the static CGTP addresses of the master-eligible nodes.

RNFS.Share.0=share -F nfs -o rw,root=cgtp6:cgtp7 \
-d "SUNWcgha" /SUNWcgha/local/export

RNFS.Share.1=share -F nfs -o rw -d "diskless1" -o \
rw,root=diskless1:diskless1-b:cgtp-diskless1 \
/export/root/netraDISKLESS1

If you have diskless nodes in the cluster, add the cgtp0 address of the diskless nodes.

RNFS.Share.0=share -F nfs \
-o rw,root=cgtp6:cgtp7:cgtp8 \
-d "SUNWcgha" /SUNWcgha/local/export
RNFS.CheckReplicatedSlices

Check the status of replicated slices by continuously scanning them. This property is disabled by default. To enable this property, set RNFS.CheckReplicatedSlices to True:

RNFS.CheckReplicatedSlices=True

This check is only performed when the cluster is in the synchro:READY state. To determine the synchronization state of the cluster, run the nhcmmstat command. If you enable this property and the cluster is not in the synchro:READY state the following action occurs:

  • An error message is displayed on the master node.

  • The cluster is put in the synchro:NEEDED state to prevent a switchover occurring. If a switchover occurs, the master might not have access to the most up to date data.

RNFS.CheckReplicatedSlicesInterval

Set the time between two successive reads of the replicated slices. Values are a number of milliseconds. The default value is 10 milliseconds. The RNFS.CheckReplicatedSlicesInterval property is ignored if the RNFS.CheckReplicatedSlices property is not set to True.

RNFS.CheckReplicatedSlicesInterval=time in milliseconds
RNFS.EnableSync

Determine when disk synchronization occurs. Values are True and False. The default is True.

If RNFS.EnableSync is set to True, disk synchronization is triggered at startup.

If RNFS.EnableSync is set to False, you delay the start of disk synchronization until you use the nhenablesync command. For more information on this command, see the nhenablesync(1m) man page.

RNFS.EnableSync=False
RNFS.SyncType

Specify the method used for synchronization between slices on master-eligible nodes. Values are FS or RAW. The default value is FS.

If RNFS.SyncType is set to FS, this feature is enabled. The time taken for a full slice synchronization is reduced because only the blocks used by the slice's file-system are replicated. If RNFS.SyncType is set to RAW, this property is disabled.

Both master-eligible nodes must have the same value for the RNFS.SyncType parameter. If you change the value of the RNFS.SyncType property, reboot the master-eligible nodes one at a time. For information on rebooting cluster nodes, see "Shutting Down and Restarting a Cluster" in the Netra High Availability Suite Foundation Services 2.1 6/03 Cluster Administration Guide. The RNFS.SyncType property is only valid for master-eligible nodes.

RNFS.SerializeSync

Determine how disk synchronization is performed. Values are TRUE and FALSE. The default value is FALSE.

Synchronization is necessary following a switchover or the vice-master booting, or when you request a full replication. In these circumstances, if RNFS.SerializeSync is set to FALSE, Reliable NFS starts the synchonization of all slices at the same time. If RNFS.SerializeSync is set to TRUE, slices are synchronized one slice at a time. This reduces the network and disk overhead but increases the time it takes for the vice-master to synchronize with the master. During this time, the vice-master is not eligible to take on the role of master.

For more information on the Reliable NFS parameters, see the nhcrfsd(1m) man page.

Note – Some parameters in this file require a single value while others require multiple values.
RNFS.StatdAlternatePath=/SUNWcgha/local

RNFS.Slice.0=/dev/rdsk/c0t0d0s3 /dev/rdsk/c0t0d0s5 /dev/rdsk/c0t0d0s3 \
/dev/rdsk/c0t0d0s5 1

RNFS.Slice.1=/dev/rdsk/c0t0d0s4 /dev/rdsk/c0t0d0s6 /dev/rdsk/c0t0d0s4 \
/dev/rdsk/c0t0d0s6 1

RNFS.Share.0=share -F nfs -o rw,root=b14-ct400-10-cgtp:b14-ct400-20-cgtp \
-d "SUNWcgha" /SUNWcgha/local/export

RNFS.Share.1=share -F nfs -o rw -d "Export" /export

NODE STATE MANAGER PARAMETERS

Configure these parameters to specify the external floating address assigned to the master node. The following parameters must be configured when you use the Node State Manager.

NSM.Exec.MasterDir

The directory containing the scripts for the master node. There is no default value.

NSM.Exec.MasterDir=/opt/SUNWcgha/actions/master
NSM.Exec.ViceMasterDir

The directory containing the scripts for the vice-master node. There is no default value.

NSM.Exec.ViceMasterDir=/opt/SUNWcgha/actions/vicemaster
NSM.Log.MasterDir

The log file directory for the master node. There is no default value.

NSM.Log.MasterDir=/var/run/SUNWcgha/actions/master
NSM.Log.ViceMasterDir

The log file directory for the vice-master node. There is no default value.

NSM.Log.ViceMasterDir=/var/run/SUNWcgha/actions/vicemaster
NSM.External.Master.Address

The external floating address that is always assigned to the master node.

NSM.External.Master.Address=IP-address/netmask-size

The netmask-size value is optional. There is no default value.

NSM.External.Master.Address=192.168.12.39
NSM.External.Master.Broadcast

The broadcast address. This parameter is optional. If not configured, the broadcast address is automatically determined based on the IP address and netmask. There is no default value.

NSM.External.Master.Broadcast=192.168.255.255
NSM.External.Master.NIC

The physical or logical network interface for the external floating address of the master node. There is no default value.

NSM.External.Master.NIC=hme0:3
NSM.External.Master.RouteAdd

Set the arguments to be passed to the route add command when a master-eligible node becomes a master node. This parameter is optional. If not defined, no routes are defined. There is no default value.

NSM.External.Master.RouteAdd=default 192.168.12.250

The preceding example generates the following command:

route add default 192.168.12.250
NSM.External.Master.RouteDelete

Set the arguments to be passed to the route delete command when a node loses the master role. This parameter is optional. There is no default value.

NSM.External.Master.RouteDelete=default 192.168.12.250

The preceding example generates the following command:

route delete default 192.168.12.250

WATCHDOG TIMER PARAMETERS

The following parameters are optional and must be configured only if you have installed the Watchdog Timer on the nodes of your cluster.

WATCHDOG.NhasWatchdog

If this parameter is set to true, the product enables the watchdog and initiates the patting of the watchdog.

Options are true or false. The default is false.

WATCHDOG.NhasWatchdog=false
WATCHDOG.ShutDownTimeout

The maximum number of milliseconds to wait for the system to reach init run level 6 during a shutdown. See the init(1m) man page.

The value must be provided in milliseconds. The default value is 5000.

WATCHDOG.ShutDownTimeout=6000
WATCHDOG.OsTimeout

The maximum number of milliseconds to wait between the system boot at init run level 2 and init run level 6. See the init(1m) man page.

The value must be provided in milliseconds. The default value is 3000.

WATCHDOG.OsTimeout=5000
WATCHDOG.PattingIntvl

The interval after which the hardware watchdog is re-enabled by the Watchdog Timer daemon, nhwdtd. The value must be provided in milliseconds and must be at least 1000 milliseconds less than the values of the WATCHDOG.OsTimeout and WATCHDOG.ShutDownTimeout parameters. The default value is 2000.

patting_intvl=5000

EXAMPLES

Example 1. Example of the nhfs.conf File on a Master-Eligible Node
### Common part
Node.DomainId=69
Node.NIC0=hme0:1
Node.NIC1=hme1
Node.NICCGTP=cgtp0
Node.UseCGTP=True
Node.type=Diskfull

### Cluster part
Cluster.Master.Id=1

### CMM part
CMM.IsEligible=True
CMM.LocalConfig.Dir=/etc/opt/SUNWcgha

### RNFS part
RNFS.StatdAlternatePath=/SUNWcgha/local

RNFS.Slice.0=/dev/rdsk/c0t0d0s3 /dev/rdsk/c0t0d0s5 /dev/rdsk/c0t0d0s3 \
/dev/rdsk/c0t0d0s5 1
RNFS.Slice.1=/dev/rdsk/c0t0d0s4 /dev/rdsk/c0t0d0s6 /dev/rdsk/c0t0d0s4 \
/dev/rdsk/c0t0d0s6 1

RNFS.Share.0=share -F nfs -o rw,root=b14-ct400-10-cgtp:b14-ct400-20-cgtp \
-d "SUNWcgha" /SUNWcgha/local/export
RNFS.Share.1=share -F nfs -o rw -d "Export" /export

RNFS.Installed=True

### NSM Part
NSM.Exec.MasterDir=/opt/SUNWcgha/actions/master
NSM.Exec.ViceMasterDir=/opt/SUNWcgha/actions/vicemaster

NSM.Log.MasterDir=/var/run/SUNWcgha/actions/master
NSM.Log.ViceMasterDir=/var/run/SUNWcgha/actions/vicemaster

NSM.External.Master.Address=192.168.12.39
NSM.External.Master.NIC=hme0:3
NSM.External.Master.Broadcast=192.168.255.255

NSM.External.Master.RouteAdd=default 192.168.12.250
NSM.External.Master.RouteDelete=default 192.168.12.250
Example 2. Example of the nhfs.conf File on a Diskless Node
### Common part
Node.DomainId=250
Node.NIC0=hme0:1
Node.NIC1=hme1
Node.NICCGTP=cgtp0
Node.UseCGTP=True
Node.type=Diskless

### CMM part
CMM.IsEligible=False
CMM.LocalConfig.Dir=/etc/opt/SUNWcgha
CMM.CurrentNodeName=netraDISKLESS1
Example 3. Example of the nhfs.conf File on a Dataless Node
Node.NICCGTP=cgtp0
Node.UseCGTP=True
Node.NIC1=eri1
Node.NIC0=eri0
Node.DomainId=250
Node.Type=Dataless

CMM.IsEligible=False
CMM.LocalConfig.Dir=/etc/opt/SUNWcgha
CMM.CurrentNodeName=netraDATALESS1

ATTRIBUTES

See attributes(5) for descriptions of the following attributes:

ATTRIBUTE TYPEATTRIBUTE VALUE
ArchitectureSPARC
AvailabilitySUNWnhcdt
Interface StabilityEvolving

SEE ALSO

init(1m), nhcrfsd(1M), nhcmmd(1M), share(1m), nhenablesync(1m), nhpmd(1M), nhnsmd(1M), cluster_nodes_table(4)


Netra HAS FS 2.1Go To TopLast Changed September 2004