Thursday Nov 20, 2014

Restoring an Enterprise Controller on a new system

An Ops Center user was recently planning a deployment, and they were wondering about disaster recovery. Apart from Ops Center's High Availability option, they asked, what recovery options are available if an EC system completely dies?

One other option that you can use is Ops Center's backup and restore capability. If you've made a backup of your Enterprise Controller, it's possible to restore that Enterprise Controller on a new system. There are a few tricks with doing so, though:

  • The new system has to have the same version of Ops Center, including all upgrades and IDRs.
  • If the IP address of the new system is different, you have to manually change the Proxy Controllers to point to the new IP.
  • Ops Center libraries are not part of the backup, so if they're on nfs storage, they need to be accessible to the new system as well.
  • The /var/opt/sun/xvm/images/os directory, containing OS images, needs to be backed up and restored separately, because it's big.

The Backup and Restore chapter explains how to restore the Enterprise Controller on a new system.

Thursday Nov 13, 2014

New Library Design

As you may have noticed, we've revamped the design for docs.oracle.com. While the new look and feel lets us present information for a lot of products in a clearer (and snazzier) format, it also means that the path to the Ops Center docs from the main page is different.

The new library front page has a set of icons for the various categories of documentation. To get to Ops Center, you'll click on the Enterprise Manager category.


This brings up the main page for Enterprise Manager software, which has a tab for Ops Center. You'll click that tab:

Then, once the tab is displayed, you'll click the version of Ops Center that you want:


It's worth noting that the urls for the Ops Center libraries haven't changed, so if you have them bookmarked you don't need to change anything. You can also still get to the library through the help button at the top of the Ops Center UI. This only affects how you get to those libraries through the landing page.

Thursday Nov 06, 2014

Incident Icons

Ops Center packs a lot of information into a browser window. The upside is that you have a lot of information available, but the downside is that some of the information is presented using icons that can be confusing at first look.

The incident icons are one example of this sort of concentrated information. If you select an asset in Ops Center, you'll see a set of icons in the top-right section of its display:


These icons indicate the number of critical, warning, and info-level incidents on the selected asset. You can click any of the icons to go to the incidents tab:


There's also another set of icons in the top left that give you a broader picture of the incidents in your environment:

The first two icons show you how many unassigned critical and warning-level incidents there are in your environment. The middle icon shows the number of relayed incidents. The right two icons show how many critical and warning-level incidents have been assigned to you (that is, whoever is currently logged in.) Clicking on any of the icons takes you to the relevant section of the message center.

The Managing Incidents how-to has more information about these icons and about how you can manage incidents in Ops Center.

Thursday Oct 30, 2014

Running Scripts Using an Operational Profile

I saw an interesting question recently about operational profiles. An Ops Center user wanted to run a script with several variables on several managed systems, with the variables having different values on each system, and they wanted to know the best way to do it.

This can all be done through an operational profile (and an operational plan, which runs the profile). First, you'll select Operational Profiles in the Plan Management section of the UI, then you'll click Create Profile.

You put your script in here, using whatever variables you need. You also specify a type - Remote Shell scripts can be run on any system with an agent, using root permissions, while EC Shell scripts are only run on the Enterprise Controller system with the logged-in user's credentials.

When you click next, there's a screen to define the variables used. You specify the variables, but leave the values blank. Then, when you run the operational plan on a system, you'll be prompted to supply the values for each variable.


Take a look at the Plans and Profiles chapter of the Feature Reference Guide for more information.

Thursday Oct 23, 2014

Managing user access in multiple sites

An Ops Center user who's setting up their environment sent in a question about their users:

"I'm looking to manage three different data centers from one Enterprise Controller instance. However, the three data centers have different administrators. How can I make sure that each administrator can see and manage only the resources that they're supposed to?"

The answer here is that you can use Ops Center's asset groups, combined with its fine-grained roles capabilities, to control which users can see and do what.

First, you create a new asset group in the Assets section of the UI. In this example, I'm creating a group for one of the three data centers:


Once you've created the group, you can add the correct assets to it, by selecting the assets and clicking Add Asset to Group:


Now that you have the assets for one of the datacenters grouped together, you add that admin to Ops Center:


Then you'll select that user and click the Manage User Roles icon. When the wizard comes up, you make sure they have the correct roles, then deselect the "Use the default role associations" checkbox:


When you click next, you select which groups the roles should apply to. So, for this user, we can apply their Asset Admin role only to the Data Center A group:


And there you have it. Rinse and repeat for other groups and users, and each user will be able to see and manage only the correct assets. For more information, check out the Asset Management and User and Role Management chapters.

Thursday Oct 16, 2014

How-To: Creating and Managing Network Domains

Efficient networking is crucial for a virtual datacenter. (I suppose it's crucial for any kind of datacenter, but I'm focusing here.) In Ops Center, you can keep your vDC networks running using network domains. Network domains manage public networks with defined resources as well as ad-hoc private networks.

To create a network domain through Ops Center, you need to identify at least one fabric to support it, then run the Create Network Domain wizard. This wizard lets you create new dynamic private networks for the new domain and associate existing networks with it.

Once you've created a network domain, you can associate it with a server pool by selecting the server pool and running the Associate Network Domain wizard. This wizard lets you pick a network domain, then pick a physical interface on each server in the server pool for the network domain to use. Once you've done this, the server pool can use public and private networks within the network domain.

The Creating and Managing Network Domains how-to provides more background information and walks you through both of these wizards.

Thursday Oct 09, 2014

How-To: Lifecycle Management for Zones

Oracle Solaris Zones are a useful too. As I've been discussing over the past few weeks, Ops Center lets you manage zones from a high level by creating them, grouping them into server pools, and migrating them. However, I haven't talked about how to use Ops Center for more day-to-day zone management.

When a zone is managed in Ops Center, you can see its current state - running, shut down, or unreachable - in the Assets section, shown by an icon over the asset. By selecting a specific zone, you can take a number of possible actions: you can reboot or shut down a running zone, halt a zone if it's not responding to a graceful shutdown, boot a stopped zone, or delete a zone if it's no longer needed.

The Lifecycle Management of Zones how-to explains how to take each of these actions. The zones chapter in the Feature Reference Guide also discusses them.

Thursday Oct 02, 2014

How-To: Migrating Oracle Solaris 11 Zones

I've talked a bit over the past couple of weeks about how to create Oracle Solaris 11 zones and how to group them into a server pool to enable automatic load balancing and high availability.

You can also use the server pools to manually migrate zones from one host to another. You can perform this migration to do your own load balancing, to reorganize your zones, or in preparation for system maintenance. When you use the Migrate Zone action to move the zone, the wizard lists all of the global zones in the pool, including the number of running zones, total CPU, and available dedicated CPUs. If you don't have enough resources in that server pool, you can migrate the zone to a global zone in a different server pool, as long as they're compatible and meet the storage and network requirements. 

The Migrating Zones how-to walks you through migrating a zone to a new global zone in the same server pool. There's also more information in the Oracle Solaris Zones chapter in the Feature Reference Guide.

Thursday Sep 25, 2014

How-To: Creating a Server Pool for Oracle Solaris 11 Zones

I talked a bit last week about creating Oracle Solaris 11 zones by following the zone creation how-to. If you're creating zones on at least two systems, it's a good idea to create a server pool for them. A zones server pool lets you group 2 or more global zones (hosts) that have the same processor architecture. The hosts must share storage resources and have access to the same virtual and physical networks.

There are some big advantages to pooling your zones, including load balancing and high availability. When you add hosts to the server pool, you can share resources between the zones, and create policies to manage a lot of the CPU utilization and resource balancing functions and automatically migrate hosts between servers to balance the load. You can also manually move zones between hosts, and restart zones on a new host if a pooled host shuts down.

If you want to know more, there's a Server Pools chapter in the Feature Reference Guide. There's also a Creating a Server Pool for Zones how-to which walks through the process, using a hypothetical environment with two Oracle Solaris 11 zones.

Thursday Sep 18, 2014

How-To: Creating Oracle Solaris 11 Zones

The odds are good that you use Oracle Solaris 11 Zones in your environment or are interested in doing so - it's an effective and secure virtualization technology that lets you get more use out of each piece of hardware.

If your environment uses Ops Center, you can use a profile to quickly create new zones with a specific setup. The profile specifies the OS version, zone configuration, storage, and file system for the zone. Once you've created a profile, it automatically generates a deployment plan, which you can run to create a new zone. At runtime you specify the target system, and the name, IP address, and network resources for the new zone, edit any of the info from the profile if you want, and schedule the zone creation.

We've put together a how-to that walks you through the whole process, so take a look.

Thursday Sep 11, 2014

OCDoctor Updated

The OCDoctor is a tool that you can use to gather prerequisite and troubleshooting information about your systems. It's included in Ops Center, but you can also download it from here.

Update 4.38 was just released a couple of days ago. It has some new checks, including checking the remote database tablespace sizes and looking for lockfiles left over from prior upgrades. Also, if you run the doctor with the --update option from an Agent managed system, it'll pull the latest version from the Enterprise Controller. This could be handy if you're in a site with limited connectivity.

The OCDoctor chapter explains some of the things that you can do with the OC Doctor, so take a look and get the latest version. (Note that if you're using Ops Center in connected mode, the latest version is downloaded automatically.)

Thursday Sep 04, 2014

Communication Between OC and EM

I've seen a few questions about connecting Ops Center and EM Cloud Control monitoring - what permissions are needed to make it work.

Ops Center and EM can share data using the Infrastructure Stack Plug-In. This plug-in lets you view EM processes in the Ops Center UI and Ops Center monitoring and annotations in the EM console, including monitoring for service processors, chassis, server domains, zones, and Oracle VM Server for SPARC domains and guests.

To share data between Ops Center and EM, you create an Ops Center user to interact with the plugin. This user can have any roles you choose, including the read-only role.

Each product has its own set of users and roles. The SYSMAN user can see asset information collected by Ops Center software in the EM console and can launch the Ops Center console, but doesn't have  SYSMAN permissions in Ops Center - they only have the permissions of the user you created for the connection. Similarly, an Ops Center Admin can launch the EM console from Ops Center to view information, but they don't have Admin permissions in the EM console. 

For more information about the plugin, see the Plugin Guide.

Thursday Aug 28, 2014

Oracle Solaris 11.2 Support

Oracle Solaris 11.2 was released just after Ops Center 12.2.1, and I've seen a few questions about when we'll support it, and to what degree.

The first answer is that Oracle Solaris 11.2 is now supported for most features in Ops Center. You can manage it, install it, update it, create zones on it - basically, anything you can do with Oracle Solaris 11.1, you can also do with Oracle Solaris 11.2.

The caveat to that is that the features introduced with Oracle Solaris 11.2, such as Kernel Zones, are not yet officially supported through Ops Center. We're working hard on adding that support right now.

Thursday Aug 21, 2014

Using Asset Groups

I got a question about putting assets in groups:

"I'm planning on installing some agents manually on existing systems, and I want to have them put in a specific asset group once they're discovered. I don't see any way to tell the install script to put the asset in a group. How can I add the assets to a group, either through the UI or the CLI?"

There are a few ways.

In the CLI, you can use groups mode, and use this command to add an asset to a group:

attach -n| --gear <asset name> -g| --group <group>

You can also use -U| --uuid <UUID> to specify the asset if you have multiple assets with the same name.

In the UI, you have a couple of options. You can select an asset and click Add Asset to Group to add it to a group you select.

Alternatively, if you're trying to make a group for assets with a specific characteristic, you can specify rules that will automatically add assets to a group based on that characteristic.

Thursday Aug 14, 2014

CPU Architectures in 12.2.1

I've talked about a few of the enhancements that came in version 12.2.1 over the past few weeks. The last one is a new generic architecture class for newer systems that gives you new migration options for logical domains.

Some new systems - Oracle SPARC T4 servers, Oracle M5 and M6 servers, and Fujitsu M10 servers - have a class1 architecture. When you start a guest domain on one of these systems, Ops Center recognizes the architecture, and you can migrate the guest to systems with other CPU architectures without losing any LDOM capabilities.

Take a look at the Oracle VM Server for SPARC chapter in the Feature Reference Guide for more information about CPU architectures and other Logical Domain configuration options.

About

This blog discusses issues encountered in Ops Center and highlights the ways in which the documentation can help you

Search

Archives
« March 2015
SunMonTueWedThuFriSat
1
2
3
4
6
7
8
9
10
11
13
14
15
16
17
18
20
21
22
23
24
25
26
27
28
29
30
31
    
       
Today