CA NetVoyant 7.1 SP7 Release Notes

1.0 Welcome

CA NetVoyant (NetVoyant) provides SNMP-based performance metrics for managing network infrastructure, devices, and services. Using data from devices such as routers, switches, and servers, NetVoyant summarizes and condenses the data into easy-to-understand, web-based reports.

These Release Notes provide information about the enhancements and open issues in version 7.1 SP7. This information supplements and supersedes information in the Product Documentation.

2.0 Why Install this Release?

This release of NetVoyant resolves a memory-consumption problem in NetVoyant 7.1 SP5 and SP6. (88067)

Important: Beginning with release 7.1 SP6, the remote console installer (netvoyant-console-x.x.x-win32-i586.exe) is not included in the NetVoyant installation kit. You can download the remote console installer from the CA Support FTP site: ftp://ftp.ca.com/pub/netqos/products/NV/7.1/

3.0 Upgrade and Installation Information

Upgrades to NetVoyant 7.1 SP7 are supported from versions 6.1 (6.1.18) and later.

If an upgrade fails with a “not enough disk space” message, and there is sufficient disk space, contact CA technical support for assistance.

For more information about installing or upgrading the NetVoyant software, see the CA NetVoyant Installation and Upgrade Guide. This guide is available in the CA NetVoyant Bookshelf on the CA Support Online website.

4.0 Version Compatibility

CA NetVoyant 7.1 SP7 is compatible with CA NetQoS Performance Center 6.x.

CA NetVoyant is not compatible with any version of CA Performance Center.

5.0 NetVoyant Watchdog Service

The Watchdog service notifies you when NetVoyant services are down, and, when possible, automatically repairs the problem. When you install this service pack, the Watchdog service is installed in the <install path>/bin directory.

After you install NetVoyant, start and run nvwatchdog.exe. The resulting dialog prompts for a page title, login name, and password. Leave the Page Title field blank to disable the Login Name and Password fields.

The Watchdog process performs the following tasks:

• The Watchdog connects to the nms2 database and reads the email server settings for CA NetQoS Performance Center.
• The Watchdog requests an email administrative address and registers as a Windows service.

  Note: An SMTP failure can prevent the Watchdog service from registering.

• The Watchdog tests the parameters to ensure that email delivery can be accomplished.
• After a reboot, the service is visible in the Control Panel in the Services dialog with a status of “automatic startup.”
• Parameters are saved to \netvoyant\nvwatchdog.xml so that the Watchdog can raise an alarm even when the mysql-nms2 database fails.
• Testing is automatically performed every 2 minutes. Testing typically completes in five to 10 seconds, and then sits idle until repeated.
• Critical test results generate Sev-1 email notifications. Sev-1 emails are suppressed during the first 10 minutes after Watchdog starts.
• Major test results generate Sev-2 email notifications. Sev-2 emails are suppressed during the start-up period, for back-to-back failures (up to an hour), and during the interval following a Sev-1 notification.
• Warning test results generate Sev-3 email notifications, which provide information about poor collecting or reporting performance, such as overruns from software-only hardware limitations.

6.0 Performance and Scalability

A distributed NetVoyant 7.1 system supports ten pollers (each with 55,000 polled and 140,000 discovered poll instances) for a total of 550,000 polled and 1,400,000 discovered poll instances.

A standalone system supports a poller with 55,000 polled and 140,000 discovered poll instances.

6.1 Discovery Performance

In our testing, the average length of time for discovery is 2.5 hours per poller. The initial discovery, including synchronization, is four hours on a distributed system with ten pollers.

6.2 Manual (Right-Click) Synchronization Performance

In our testing, we saw the following results when manually forcing a synchronization:

6.3 Poll Instance Performance

In our testing, we spread the polling instances across multiple datasets. We focused on ifstats, qosclass, and qosred, all of which are known to cause performance and scalability issues. The results for each poller are as follows:

6.4 Console Performance

With 120 configured custom groups, we saw the following performance in our testing:

• Using group membership rules, devices were populated in all tiers in less than 5 seconds.
• Navigating the sub-groups took less than 3 seconds per group to load. The top groups of routers and switches from the Master took longer, up to 20 seconds.
• The Alarm Profiles Group tier had no problem with rendering or performance.
• During discovery:
  • After the initial load was completed, everything responded within 3 seconds.
  • The Routers and Switches groups took longer than 10 seconds, depending how many devices were already populated.
  • When custom groups were created, browsing during and after discovery took no more than 5 seconds.
  • With 7000+ routers in a group and 250+ switches in a group with many child polling instances, the load took about 20 seconds.
  • Browsing with few devices already in the groups while they were populating took no more than 2 seconds, with 25 percent of the devices discovered.

6.5 Performance Factors

Performance results can vary based on your environment. The following is a sampling of the factors that can affect performance:

• Line latency
• SNMP response delay
• Network loss
• Custom MIBs and datasets
• Number of discovered instances
• Number of enabled instances
• Non-default poll groups, such as one-minute
• Number of poll groups
• Custom software running on the system, such as anti-virus software
• SNMP device and agent mix:
  • The GetNext command requires more requests than the GetBulk command.
  • SNMPv3 adds encryption overhead.
• Number of time filters
• Data retention settings
• Custom hardware and virtual machine settings
• PollQueueLimit settings
• PollConcurrency settings
• Number of groups
• Distribution of poll instances across datasets

6.6 Test Configurations

Our test environment for a large distributed NetVoyant system included the following:

Master console

• Dell R610 with 12 GB of memory
• Microsoft Windows 2008 R2
• IIS, .NET, and SNMP

Ten pollers

• Dell R610 with 8-12 GB of memory
• Microsoft Windows 2008 R2
• IIS, .NET, and SNMP
• 55,000 polled and 140,000 discovered poll instances
• Seven pollers had added latency of 10-20 ms.
• Three pollers had added latency of 100-200 ms.

Our test environment for a small distributed NetVoyant system included the following:

Master console

• Dell R610 with 4 GB of memory
• Microsoft Windows 2008 R2
• IIS, .NET, and SNMP

Ten pollers

• One Dell R610 with 4 GB of memory
• One Dell R610 with 16 GB of memory
• Microsoft Windows 2008 R2
• IIS, .NET, and SNMP
• 55,000 polled and 140,000 discovered poll instances

Our test environment for a standalone NetVoyant system included the following:

One poller

• One Dell R610 with 2-12 GB of memory
• Microsoft Windows 2008 R2
• IIS, .NET, and SNMP
• 55,000 polled and 140,000 discovered poll instances

7.0 Known Issues and Workarounds

This section describes known issues and suggested workarounds.

Poll instances disappear on the poller after periodic discovery, or interfaces remap at random.

These two issues are accompanied by one or more of the following entries in the log:

1. Unable to compare poll instance (value)
2. Unable to copy (value)
3. Unable to copy pollinst: CORBA exception

Workaround: Ensure that the Master console does not run discovery while pollers are running discovery. NetVoyant versions 7.1 SP3 and later resolve 1 and 2, and reduce or eliminate occurrences of 3. (26259 and 26455)

Cannot create two auto-enable rules with the same SNMP requirement but different property requirements.

If you attempt to create an auto-enable rule using an SNMP requirement and property requirement, the SNMP requirement cannot be used in another rule. Workaround: You can vary the SNMP requirement by adding another item that is always true, or adding more characters. For example, use IFname = Eth0/0 for one rule and IFname like Eth0/0 in another. (21930)

The interface volume and interface rate views trend the same, but when switched to a quarterly view the volume trend takes a sharp downturn.

This problem occurs because the volume is a sum of data points in a weekly rollup, and the rate is an average. Therefore, when the rate is trending up, the volume is trending down. When viewing the graphs on a Saturday right before the weekly rollup, these rates would probably look similar, but on a Monday, when the volume only has one day’s worth of data, it looks as though there is a discrepancy. Workaround: Use a 30-day time period, which should provide the granularity that is needed and does not present the same issue. (20709)

SNMP profiles are not synced up when added to a poller, which is then connected to Master in a distributed system.

Back up the SNMP profiles and database before you migrate from a standalone system (all-in-one) to a distributed environment. (21443)

Adding a device in the Device Wizard does not assign the correct poller.

This problem occurs when you use the Device Wizard to add a device on the Master in a distributed system where the Master also polls. Typically, the Master server does not poll. However, it can support the polling of the pollers in some cases. Workaround: To ensure that the added device is assigned correctly, add a discovery scope in the NetVoyant console, and then start discovery of the single device by selecting File, New, Device. (22614)

NetVoyant online Help is not visible on a Windows Server 2008 computer.

The Help panel is blank, and no error message appears. This problem occurs when the operating system for the NetVoyant console computer is Microsoft Windows Server 2008 and the computer is not known by Internet Explorer as a trusted site. To add the local computer as a trusted site, click Tools, Internet Options in your browser window. On the Security tab, click the Trusted sites icon, and then click Sites. (24912)

Behavioral issues occur with the use of wild cards in auto-enable rules.

When you create an auto-enable rule with a wild card, always use the asterisk (*) wildcard. Do not use the percent (%) wild card. (10460)

8.0 Contact CA Technologies

Contact CA Support

For your convenience, CA Technologies provides one site where you can access the information that you need for your Home Office, Small Business, and Enterprise CA Technologies products. At http://ca.com/support, you can access the following resources:

• Online and telephone contact information for technical assistance and customer services
• Information about user communities and forums
• Product and documentation downloads
• CA Support policies and guidelines
• Other helpful resources appropriate for your product

Providing Feedback About Product Documentation

If you have comments or questions about CA Technologies product documentation, you can send a message to techpubs@ca.com.

To provide feedback about CA Technologies product documentation, complete our short customer survey which is available on the CA Support website at http://ca.com/docs.

Copyright © 2014 CA. All rights reserved. All trademarks, trade names, service marks, and logos referenced herein belong to their respective companies.