openSolaris 2008 - Modifying Timeouts on SLP Discovery Requests - System Administration Guide: Network Services

Modifying Timeouts on SLP Discovery Requests

Two situations might require that you change the timeouts for SLP discovery requests:

If the SLP agents are separated by multiple subnets, dial-up lines, or other WANs, the network latency can be high enough that the default timeouts are insufficient for a request or registration to be completed. Conversely, if your network is low latency, you can improve performance by decreasing the timeouts.
If the network is subject to heavy traffic or a high collision rates, the maximum period that SAs and UAs need to wait before sending a message might be insufficient to assure collision-free transactions.

Changing Default Timeouts

High network latency can cause UAs and SAs to time out before a response returns for requests and registrations. Latency can be a problem if a UA is separated from an SA, or if both a UA and an SA are separated from a DA;either by multiple subnets, a dial-up line, or a WAN. You can determine if latency is a problem by checking whether SLP requests are failing because of timeouts on UA and SA requests and registrations. You can also use the ping command to measure the actual latency.

The following table lists configuration properties that control timeouts. You can use the procedures in this section to modify these properties.

Table 9-4 Time-out Properties

Property

Description

net.slp.multicastTimeouts

net.slp.DADiscoveryTimeouts

net.slp.datagramTimeouts

The properties that control timeouts for repeated multicast and unicast UDP message transmissions before the transmission is abandoned.

net.slp.multicastMaximumWait

The property that controls the maximum amount of time a multicast message is transmitted before it is abandoned.

net.slp.datagramTimeouts

The upper bound of a DA timeout that is specified by the sum of values that are listed for this property. A UDP datagram is repeatedly sent to a DA until a response is received or the time-out bound is reached.

If frequent timeouts are occurring during multicast service discovery or DA discovery, increase the net.slp.multicastMaximumWait property from the default value of 15000 milliseconds (15 seconds). Increasing the maximum wait period allows more time for requests on high latency networks to be completed. After you change the net.slp.multicastMaximumWait, you should also modify the net.slp.multicastTimeouts and net.slp.DADiscoveryTimeouts. The sum of the timeout values for these properties equals the net.slp.multicastMaximumWait value.

How to Change Default Timeouts

Use the following procedure to change the SLP properties that control timeouts.

Become superuser or assume an equivalent role.
Roles contain authorizations and privileged commands. For more information about roles, see Configuring RBAC (Task Map) in System Administration Guide: Security Services. To configure a role with the Primary Administrator profile, see Chapter 2, Working With the Solaris Management Console (Tasks), in System Administration Guide: Basic Administration.
Stop slpd and all SLP activity on the host.
```
# svcadm disable network/slp
```
Back up the default /etc/inet/slp.conf file before you change the configuration settings.
Change the net.slp.multicastMaximumWait property in the slpd.conf file:
```
net.slp.multicastMaximumWait=value
```
value

A 32–bit integer that lists the sum of the values that are set for net.slp.multicastTimeouts and net.slp.DADiscoveryTimeouts
Default Value=15000 milliseconds (15 seconds)
Range of Values=1000 to 60000 milliseconds

For example, if you determine that multicast requests require 20 seconds (20000 milliseconds), you would adjust the values that are listed for net.slp.multicastTimeouts and the net.slp.DADiscoveryTimeouts properties to equal 20000 milliseconds.
```
net.slp.multicastMaximumWait=20000
net.slp.multicastTimeouts=2000,5000,6000,7000
net.slp.DADiscoveryTimeouts=3000,3000,6000,8000
```
If necessary, change the net.slp.datagramTimeouts property in the slpd.conf file:
```
net.slp.datagramTimeouts=value
```
value

A list of 32–bit integers that specify timeouts, in milliseconds, to implement unicast datagram transmission to DAs
Default=3000,3000,3000

For example, you can increase the datagram timeout to 20000 milliseconds to avoid frequent timeouts.
```
net.slp.datagramTimeouts=2000,5000,6000,7000
```
In high-performance networks, you can reduce the time-out bound for multicast and unicast UDP datagram transmission. When you reduce the time-out bound, you decrease latency that is required to satisfy SLP requests.
Save your changes and close the file.
Restart slpd to activate your changes.
```
# svcadm enable network/slp
```

Configuring the Random-Wait Bound

In networks with heavy traffic or a high collision rate, communication with a DA might be affected. When collision rates are high, the sending agent must retransmit the UDP datagram. You can determine if retransmission is occurring by using snoop to monitor traffic on a network of hosts that are running slpd as an SA server and a host that is running slpd as a DA. If multiple service registration messages for the same service appear in the snoop trace from the host that is running slpd as an SA server, you might have notice collisions.

Collisions can be particularly troubling at boot time. When a DA first starts, it sends unsolicited advertisements and the SAs respond with registrations. SLP requires the SAs to wait for a random amount of time after receiving a DA advertisement before responding. The random-wait bound is uniformly distributed with a maximum value that is controlled by the net.slp.randomWaitBound. The default random-wait bound is 1000 milliseconds (1 second).

How to Configure the Random-Wait Bound

Use the following procedure to change the net.slp.RandomWaitBound property in the slp.conf file.

Become superuser or assume an equivalent role.
Roles contain authorizations and privileged commands. For more information about roles, see Configuring RBAC (Task Map) in System Administration Guide: Security Services. To configure a role with the Primary Administrator profile, see Chapter 2, Working With the Solaris Management Console (Tasks), in System Administration Guide: Basic Administration.
Stop slpd and all SLP activity on the host.
```
# svcadm disable network/slp
```
Back up the default /etc/inet/slp.conf file before you change the configuration settings.
Change the net.slp.RandomWaitBound property in the slpd.conf file:
```
net.slp.RandomWaitBound=value
```
value

The upper bound for calculating the random-wait time before attempting to contact a DA
Default Value=1000 milliseconds (1 second)
Range of Values=1000 to 3000 milliseconds

For example, you can lengthen the maximum wait to 2000 milliseconds (2 seconds).
```
net.slp.randomWaitBound=2000
```
When you lengthen the random-wait bound, a longer delay in registration occurs. SAs can complete registrations with newly discovered DAs more slowly to avoid collisions and timeouts.
If necessary, change the net.slp.datagramTimeouts property in the slpd.conf file:
```
net.slp.datgramTimeouts=value
```
value

A list of 32–bit integers that specify timeouts, in milliseconds, to implement unicast datagram transmission to DAs
Default=3000,3000,3000

For example, you can increase the datagram timeout to 20000 milliseconds to avoid frequent timeouts.
```
net.slp.datagramTimeouts=2000,5000,6000,7000
```
In high-performance networks, you can reduce the time-out bound for multicast and unicast UDP datagram transmission. This setting reduces the amount of latency in satisfying SLP requests.
Save your changes and close the file.
Restart slpd to activate your changes.
```
# svcadm enable network/slp
```