| Question |
Answer |
| What is the relationship between DECamds and the Availability Manager? |
The Availability
Manager is the follow-on product to DECamds. The Availability
Manager supports multiple platforms and has enhanced capabilities.
DECamds is in maintenance mode.
The Installation Instruction for OpenVMS Alpha Systems contains a
section that explains how to install and run DECamds and the
Availability Manager concurrently. |
| What memory does the Data Analyzer need to run optimally on OpenVMS? |
HP has found that the Availability Manager runs well using the default
configuration when you monitor approximately 20 nodes.
If you need to monitor more nodes or to
open several node displays simultaneously and find that the software runs
more slowly, you might need more memory to run
the Availability Manager more efficiently.
The performance section of the installation instructions explains how to change memory.
For example, setting the initial heap size to 40MB and the maximum heap size to
100MB is a good starting place. |
| How is this product related to other system management tools,
such as HP Systems Insight Manager, HP OpenView,
UNIX SYSMAN, and Windows NT tools such as Perfmon, Sysmon, and SMS? |
The Availability
Manager is a real-time monitoring and diagnostic
tool focused on monitoring OpenVMS nodes from a Windows 2000 or XP system
or an OpenVMS system. The Availability Manager provides unique capabilities to monitor and control OpenVMS
systems and, in addition,
is intended to complement server and enterprise management tools available from HP and other vendors.
|
| What licensing or fees are associated with the Availability Manager? |
This software is included with the OpenVMS license for
Alpha customers.
For Integrity servers, the Availability Manager is included with the Enterprise Operating Environment (EOE) license. For further
details, click here.
|
| Is there a way to provide product feedback, make suggestions, and request new features? |
You can use the standard IPMT case method of reporting problems. You can also send feedback
to make suggestions, to report problems with the web site, or
to ask general questions about the product.
HP is committed to protecting your privacy. Learn more.
|
| Can I get the Availability Manager documentation in
hardcopy format? |
Availability Manager
Version
3.1
documentation is distributed in various formats:
- The user manual is available in HTML
and PDF formats, both of which can be printed.
- The installation instructions for OpenVMS and
Windows systems
are available in HTML, TXT, and PDF formats.
- Release notes for current and prior versions are available in TXT
format.
For more information, refer to the Documentation page. |
| Does the Availability Manager run on Windows 2000? |
Version 3.1 of the Availability Manager
Data Analyzer supports Windows 2000 SP3 or later and Windows XP Professional SP2.
(Windows XP Home Edition is not supported.)
For more information, refer to the Product description
page. |
| Does the Availability Manager Java GUI depend on a particular Java Virtual Machine (JVM)? What happens if I load a different JVM on my system? |
No. The Availability Manager provides its own Java environment. The Availability Manager
will coexist with another installed JAVA kit. |
| Are there Data Collectors for Windows NT so that I can monitor NT nodes in the same way I monitor OpenVMS nodes? |
The Availability Manager Version 1.0, which
shipped with Compaq's Storage and Cluster Extensions (SCE) Version
1.0, contained a Windows NT Data Collector component for Windows
NT 4.0 (SP3). Availability Manager releases after V1.4 do not
provide the Windows NT Data Collector. |
| Do all the nodes I'm monitoring have to be on the same LAN or Ethernet segment? |
No. Now, with the Availability Manager
Version
3.1, the restriction that the Data Analyzer be on the same LAN or Ethernet segment as the OpenVMS systems has been removed. A new component, the Data Server, has been created to “tunnel” the data between the Data Analyzer and the OpenVMS nodes using the IP protocol. See Chapters 1 and 2 of the Availability Manager User's Guide for details.
|
| How many nodes can I monitor at a time? |
That depends on a number of factors: what
data is being collected, other network traffic, and
characteristics of the machine running the Data Analyzer
application. Running the Data Analyzer application on a 400 MHz
Windows 2000 machine with 192 MB of memory and collecting only node
summary data, 130-150 or more nodes can be monitored at a time.
Adding CPU power and memory increases the number of nodes
that can be monitored. Collecting additional detailed data from
nodes or groups of nodes reduces the number of nodes that can be
monitored at the default sample rate. With each new release of
the Availability Manager, these numbers increase and scale more
predictably. |
| Is there any way to dial in or VPN from home and use the Availability Manager to monitor my data center remotely? |
Yes. Now, with the Availability Manager Version
3.1, the restriction that the Data Analyzer be on the same LAN or Ethernet segment as the OpenVMS systems has been removed. A new component, the Data Server, has been created to “tunnel” the data between the Data Analyzer and Data Server over an IP network. The only limit is the bandwidth of the network connection.
|
| When I run the Availability Manager,
I can see the various OpenVMS nodes that I want to monitor, but they
remain in the yellow ("configuring") state. What is the problem?
|
One reason for this behavior might be
that your LAN is actually a VLAN, and the
port that your machine is running on is a read-only port -- where traffic
can flow to your machine but not from your machine out to the rest of the
VLAN. Make sure that the port that your machine is connected to allows
network traffic to flow in both directions.
|
| I installed the Availability Manager on OpenVMS,
but the only node I can see is the node on which I installed the application.
How can I fix this? |
On systems that have more than one network
interface card (NIC), the Data Collector (RMDRIVER) might not select the correct
network adapter. The solution is to edit SYS$MANAGER:AMDS$LOGICALS to modify the
entry AMDS$DEVICE, entering the name of the network adapter that is active on
the network (for example, EWA0).
|
| I installed the Availability Manager on my Windows PC. The
Availability Manager comes up on the screen, but I don't see any nodes
at all. I know I have active OpenVMS nodes because
when I install DECamds on one of the OpenVMS nodes, it can see the nodes.
All my other network connections work -- that is, web browser, email, and
so on.
|
This problem might be due to several
causes. One place to look is at any
firewall software installed on your PC. Another place to look is the way
the hub, router, or switch port you are connected to is configured.
The Availability Manager requires two-way communication through the port.
Windows systems with two Network interface Cards (NIC)
prompt you to enter which NIC you want to use. You might have
selected the wrong one when you started the Availability Manager.
|
| I am seeing all the nodes under the group DECAMDS. How do I put nodes into different groups?
|
Edit the SYS$MANAGER:AMDS$LOGICALS.COM file on the node you want to monitor. Change the AMDS$GROUP_NAME logical to the desired group name (e.g. “Test cluster”). Then restart the Data Collector by issuing the command $@SYS$STARTUP:AMDS$STARTUP RESTART. You should then see the node appearing under the group you have specified.
|
| I have created different groups for different sets of nodes. But I am not seeing any other group than the DECAMDS group. How can I configure the Data Analyzer to monitor the groups and nodes I want to see?
|
After starting the Data Analyzer, you can check the group and node filter settings by clicking on the “Customize” menu in the System Overview window, and the click on the Group/Node tab in the Customization dialog box. Make sure that the entries in the Group List and Node List are set to the groups and nodes you want to monitor. If you want to monitor all groups and nodes, uncheck the checkboxes for the Group List and Node List. More details on setting the Group List and Node List is given in the Explanation pane of the dialog box.
|
| How do I configure my Data Analyzers to monitor a particular set of nodes when I have two or more Data Analyzers? Also, how do I restrict some nodes from being monitored by a particular Data Analyzer?
|
You can configure a Data Collector to allow different types of access (monitor data only, allow fixes, etc.) for each Data Analyzer that connects to it. You can define a password for each kind of access in SYS$MANAGER:AMDS$DRIVER_ACCESS.DAT, and configure each Data Analyzer with the password that defines the access you want it to have.
Edit AMDS$DRIVER_ACCESS.DAT on the node and add or edit the security triplets for each type of access you desire to have. The password section is at the end of the file with the Data Analyzer info, password and access control parameter (R, W or C). After you edit the file, restart the Data Collector by issuing the command $@SYS$STARTUP:AMDS$STARTUP RESTART. Repeat the process for each system that you want to monitor by the various Data Analyzers at your site.
To configure the password on the Data Analyzer, refer to the Customization chapter of the Availability Manager Users Guide available from the Help menu on the System Overview window.
If you want one Data Analyzer to monitor one set of OpenVMS systems, and another Data Analyzer to monitor another set, you can do this by defining different passwords for the two sets of OpenVMS systems, and giving each Data Analyzer the password for the OpenVMS systems it will monitor. Using these combinations, you can configure many Data Analyzers to monitor any set of nodes you specify.
More details are available in the Security Triplet section of the AMDS$DRIVER_ACCESS.DAT file.
|
| Are there any restrictions on where I run the Data Server so that I can monitor my OpenVMS systems?
|
Yes, the Data Server must be on the same LAN as the OpenVMS systems you want to monitor. The network protocol used by the Data Collector is not routable, so if you have a router in between the Data Collector and the Data Server, the two can not communicate with each other.
|
| How do I configure my Data analyzer to monitor a selected set of process on particular node?
|
Right click on the node icon where you want to monitor the processes. Click on Customize from the pop-up menu to bring up the Customize dialog box. Click on the Watch Process tab and enter the process names in the text box on the left side of the dialog box. Note that the process names are case sensitive. When you are finished, click on Apply to save the process names, and click on OK to dismiss the dialog box.
|
| How do I correct DPGERR events with “Watch Process” program after setting some processes in watch list?
|
If you have a high process count on your system along with a large number of processes in the watch list, the watch process data collection program might run out of CPU time while examining your system. The higher the process count on your system and the longer the watch process list means more processing time is needed to complete the examination of your system.
To increase the amount of time that the data collection can use, edit SYS$MANAGER:AMDS$LOGICALS.COM for the system, uncomment AMDS$RM_MAX_BRANCHES and set the value higher, perhaps 50% higher. After saving AMDS$LOGICALS.COM, execute $ SYS$STARTUP:AMDS$STARTUP RESTART to load the new value into the Data Collector.
|
| DSKMNV events are displayed even I have the disk on my system. How do I correct it?
|
The DSKMNV event will be displayed for anyone of the following reasons:
- A removable disk on a local or remote node was removed.
- A disk on a local or remote node has gone offline due to errors.
- The node that serves the disk is down.
- The connection to a remote disk is down.
If the disk is on your system, check out whether the disk is removed or is whether properly inserted. If it is not installed properly, insert it properly into the system.
|
| If I stop the process using Fix option from Data Analyzer, the process is not stopped. Also, the fix options that I use on the process for a particular node is not happening. How do I solve this?
|
One possibility is that the system you are trying to fix has Read only permission. Edit AMDS$DRIVER_ACCESS.DAT on your Data Collector which you want to fix from Data Analyzer. Check the security triplet at the end of the file. The last part of the triplet is your privilege. If the privilege is set to R, then you have only read access to that node. Change it to W or C and restart the Data Collector with $ @ SYS$STARTUP:AMDS$STARTUP RESTART.
The stop process fix uses an AST to stop the process. If the process is not accepting ASTs for various reasons, the process cannot be stopped by the Availability Manager.
|
| I am not able to connect to the Data Server from the Data Analyzer, even though I have imported the public key from the Data Server. What could be going wrong?
|
First, try using the PING command to see if the Data Server system responds, using both the alphanumeric and numeric IP addresses or WINS address. If the numeric address succeeds, then one possibility is that the spelling of the alphanumeric address is incorrect, is not defined on your network, or you have no connection to your DNS servers. You can also use the NSLOOKUP command and the TRACEROUTE command (TRACERT on Windows systems) to verify the alphanumeric address and the route to the Data Server system. For Windows systems, start a Windows Console window (a.k.a. “DOS prompt”) to execute the commands. For OpenVMS systems, execute $@SYS$MANAGER:TCPIP$DEFINE_COMMANDS to define the commands.
If the Data Server system is running a firewall, make sure that the firewall is not blocking the port number used by the Data Server. Enter the port number you want to use for the Data Server into the firewall software (connection type used is TCP).
|
