VMware vCenter CapacityIQ 1.5.1 Release Notes
VMware vCenter CapacityIQ Version 1.5.1 | 18 March 2011 | Build 377019
Last Document Update: 20 July 2011
These release notes describe the following topics:
CapacityIQ 1.5.1 is a maintenance release that resolves some known issues and improves outlier detection and data filtering.
For information about the removal or renaming of global settings, see the Documentation known issues.
For the latest product documentation and online help content, search the vCenter CapacityIQ Online Library.
Top of Page
Prior Releases of VMware vCenter CapacityIQ
Features from prior releases of CapacityIQ are described in the release notes for each release. To view release notes for prior releases, see VMware vCenter CapacityIQ 1.5 or
Top of Page
Before You Begin
Before you install, configure, or upgrade CapacityIQ, check the requirements in the vCenter CapacityIQ Installation Guide.
Support and Scalability Guidelines
Updated: CapacityIQ 1.5.1 supports the collection of data from vCenter Server 4.0 or later versions that manage hosts running ESX Server 3.0.2 or later versions.
You can deploy the CapacityIQ virtual appliance to a host running ESX Server 3.0.2 or later versions.
This release supports 1000 hosts, 6000 powered on virtual machines, and 8000 registered virtual machines.
For information about vCenter Server scalability limits, see the Performance and Scalability Enhancements section of the What's New in VMware vSphere 4.1 — VMware vCenter white paper.
CapacityIQ supports license management through vCenter Server or through the CapacityIQ Administration Portal depending on the
license type and version of vCenter Server that CapacityIQ is registered to. Virtual machine-based licenses for CapacityIQ require vCenter Server 4.1 or later versions.
For more information about CPU and virtual machine-based licensing, see the
vCenter CapacityIQ Installation Guide.
CapacityIQ supports upgrading version 1.0.x or 1.5 to version 1.5.1. If you have CapacityIQ 1.0 and vCenter Server 2.5, you must first update CapacityIQ from
1.0 to a 1.0.x version, and upgrade vCenter Server 2.5 to vCenter Server 4.0.x. You must adhere to the strict sequence of steps depending on your versions.
The vCenter CapacityIQ Installation Guide and the Update release notes text box that appears during the update process in the Administration
Portal describe the upgrade procedure.
Check the Upgrade section of Known Issues for late-breaking updates to the upgrade procedure.
CapacityIQ stops the collection of data during the upgrade process. This process can take up to two days depending on the amount of data. The progress indicator
can only display a rough estimate of the completion time.
To avoid the need to start the process over, review all of the prerequisites.
- Check that the root partition as at least 500MB of free space to upload the upgrade .pak file.
For more information about clearing the root partition for the upgrade process, see the VMware knowledge base article 1034269.
- Do not install a version of VMware Tools that is not included with CapacityIQ before the upgrade process.
- Check that the CapacityIQ appliance has at least 3600MB of configured memory.
- Back up the virtual appliance.
- Avoid all operations from the user interface and command line that are not part of the upgrade process.
- Make sure that the vCenter Server system that CapacityIQ is registered to runs during the upgrade process.
- Close any instances of the vSphere Client that access CapacityIQ.
- Check that Hard disk 2 has at least 20% of free space to support the upgrade.
If you do not have enough free space, the upgrade process fails and you must start the process again. The Update release notes text box that appears during the upgrade
process in the Administration Portal provides the steps to increase the disk space.
For more information about completing the upgrade process and logging back into the Administration Portal, check the vCenter CapacityIQ Installation Guide
and the Update release notes text box in the interface. When you upgrade CapacityIQ, the 1.5.1 release removes the timeconfig command-line utility from the CapacityIQ virtual appliance. If you use the timeconfig utility
before you upgrade to CapacityIQ 1.5.1, you can adjust time zone settings in 1.5.1 according to the instructions in the
vCenter CapacityIQ Installation Guide.
Top of Page
The following known issues are grouped as follows:
New: When you update CapacityIQ 1.0.4 or 1.5 to 1.5.1, you see a progress bar without any indication of the remaining time during the upgrade process, and you refresh the browser, the browser displays a message that the server is not available.
Workaround: Do not refresh the browser until the upgrade process from version 1.0.4 or 1.5 to 1.5.1 ends. The upgrade process involves a period of time that makes the server unavailable for five to six hours.
When you log in but cannot determine the upgrade progress, do not refresh the browser or reboot the virtual appliance. You can instead use the tail command to
check on the upgrade status.
To check the upgrade progress from the command line:
- Log in to the CapacityIQ virtual appliance as the ciqadmin user.
- Check for any updates to the log file.
tail -F /var/log/vmware/dbupgrade.log
Updates to the log file indicate that the upgrade is still in progress. When CapacityIQ stops updating the log file, log in to the Administration Portal to confirm that the
upgrade process is successful.
When you update CapacityIQ, data might not appear immediately in the views.
Workaround: Wait two to three hours for CapacityIQ to process the data for new metrics.
If you cancel the upgrade process when the EULA or the Update release notes text box appears, the status on the Update tab in the Administration Portal
Workaround: Log out of the Administration Portal and log back in to the portal to restart the upgrade process.
When you complete the CapacityIQ upgrade process and the Administration Portal login page appears, the page might display an incorrect message about data
migration in progress.
Workaround: Ignore the message and log in to the Administration Portal. The message might appear when you migrate a small amount of data.
If you apply an update using the ciq-admin command-line interface, the build number in the Administration Portal header might not match the build number on the Status tab.
Workaround: Press F5 to refresh the Administration Portal. When the Administration Portal logs you out, you can log in to the portal.
The CapacityIQ documentation shows outdated names of the Use last known capacity per interval and Use average capacity per interval settings
in the Usable Capacity global settings. CapacityIQ 1.5.1 updates these names to Use last known capacity and Use actual capacity.
The CapacityIQ documentation describes the Use current capacity and Use capacity trend forecast methods in the Capacity and Time Remaining global settings.
CapacityIQ 1.5.1 removes these settings.
New: The CapacityIQ documentation does not include these details on negative virtual machine capacity when the capacity is smaller than the usage or demand.
CapacityIQ calculates virtual machine count capacity as the sum of the remaining virtual machine count and the deployed virtual machine count. The remaining
virtual machine count is based on the minimum value of all of the resource dimensions, such as CPU, memory, disk space, and disk I/O in the Capacity and Time
Remaining global settings. For each dimension, the CapacityIQ calculation is
(Usable or physical capacity - physical usage or demand) / (average virtual machine
size derived from
total virtual machine demand or usage / deployed virtual machine count).
CapacityIQ might display negative remaining virtual machine values in these situations:
- You activate the Use Usable Capacity global setting and you reserve too much capacity for either buffer or for VMware HA.
- You set the Use last known capacity global setting and the last known value is smaller than the average capacity. For example, the removal of a host
might lower the last known capacity value.
- CapacityIQ has missing usage data during the period when capacity decreases. This makes capacity smaller than the usage for monthly, quarterly,
and yearly intervals that roll up data. For these intervals, CapacityIQ does not maintain samples for all hours in the historical data
but instead rolls up the hourly samples into 24 by 7 weekly patterns. Each data point in the pattern represents an hour of the day and a day of the week and is
based on all of the samples of that hour and day in the interval. For example, the data point for Monday at 8 a.m. in the
monthly interval is calculated by taking the average of the usage or demand samples at 8 a.m. of all of the Mondays in the month.
If the VMware VirtualCenter Management Webservices service that disk space usage calculations depend on fails from the beginning of the second
week of the month to the end of the month, the weekly pattern for the month accounts for only the samples from the first week. If the disk
space capacity decreases during the second week while the VMware VirtualCenter Management Webservices is down, the actual or last known capacity might be
smaller than the average usage and generate negative remaining capacity values.
A virtual machine-based evaluation license for CapacityIQ that expires fails to trigger an alarm in vCenter Server.
CapacityIQ takes an hour or two to display information associated with the following events in the user interface:
- CapacityIQ might take up to two hours after a new host is added to vCenter Server for the host to account for the addition in capacity estimates.
- After installation, it takes 40 minutes or more for inventory information to appear in the user interface.
If you make changes in a cluster, such as moving a host out of the cluster and moving it back, and you create a what-if scenario, it might take two or three hours for the changes to appear in the what-if scenario.
If a dashboard or view is open for more than an hour, the information might become outdated.
Workaround: Click the selected object in the inventory panel to get the latest data.
When you create, delete, or make other updates to inventory objects, the inventory tree does not show the changes even after five minutes.
Workaround: Wait for an hour to see if CapacityIQ syncs the inventory again with vCenter Server. If CapacityIQ cannot receive updates from vCenter Server,
CapacityIQ takes one hour to detect this problem and to initiate the synchronization process.
When the vCenter Server is down, the License or Status tabs in CapacityIQ might take some time to update information.
When you specify the profile of a new virtual machine in a what-if scenario, the analysis estimates the .vmdk
file size based on the specified virtual machine disk size and options such as thin provisioned disks or linked clones.
The analysis does not account for the .vmem file size when calculating disk space usage.
Workaround: Add the estimated .vmem file size manually to the total disk space usage.
When you rename an NFS datastore, CapacityIQ cannot collect data for that datastore because the hostd management agent in vSphere cannot collect data. The
agent does not detect the changed name and marks the renamed datastore as inaccessible.
Workaround: Restart the hostd management agent, revert the datastore name to the original name at the time of creation, or add a
temporary NFS datastore to force the agent to recognize a change and trigger a restart operation.
To restart the hostd management agent on the ESX host associated with the renamed datastore:
- Log in to the ESX host using the console or SSH.
- Restart the hostd service.
Restarting the agent does not affect the running virtual machines but does affect the current operations on that host.
For more options to restart management agents from the direct console, see the vSphere documentation.
Updated: On the dashboard and views other than the Datastores - List view, utilization metrics do not account for I/O from NFS datastores.
When you select extreme thresholds in the Powered-Off and Idle VMs global settings or the Oversized and Undersized VMs global settings, the monthly, quarterly, and
yearly data in the Virtual Machine Capacity Usage - Summary view might not match the data in the Oversized Virtual Machines - List view, the Undersized
Virtual Machines - List view, and the Idle Virtual Machines - List view.
This issue might occur when you select thresholds close to the top or the bottom of the drop-down menus for CPU, disk I/O, and network I/O under Powered-Off
and Idle VMs global settings or for CPU and memory under Oversized and Undersized VMs global settings.
When you select monthly, quarterly, or yearly intervals for Non-Trend Views global settings, select extreme thresholds in the Idle VMs global settings or
Oversized and Undersized VMs global settings, and upgrade CapacityIQ from 1.5 to 1.5.1, the number of virtual machines and the metric values in the
Idle Virtual Machines - List, Oversized Virtual Machines - List, and Undersized Virtual Machines - List views might not match the original data in 1.5.
When you select monthly, quarterly, or yearly intervals for Non-Trend Views global settings, and you upgrade CapacityIQ from 1.5 to 1.5.1, the number of hosts
in the Configured Host Capacity - Distribution and Host Utilization - Distribution views might not match the original data in 1.5.
When you select a virtual machine in the inventory panel, and you select Perspective > Remaining in the dashboard, the Total VM Disk Space Remaining metric
shows negative numbers instead of zero.
For general information about what negative values for remaining capacity counts represent and the resource dimensions that determine the remaining host, cluster, and
virtual machine counts, see the vCenter CapacityIQ Online Library.
In the Virtual Machine Capacity - Summary view, the value of the Host Usage % column for disk I/O rows might be greater than the value in the Total Usage % column
even though the total disk space and disk I/O are typically larger than or equal to the disk space and disk I/O of the host.
An exception for disk I/O can occur because of the statistics that CapacityIQ collects. The total usage is based on a datastore metric that accounts only
for I/O to the VMFS partitions. The host usage is based on a disk-based metric that accounts for both VMFS and non-VMFS partitions that often exist on local datastores.
In the Undersized Virtual Machines - List view, a virtual machine without a guest operating system might appear with 100 percent memory utilization.
CapacityIQ calculates memory demand to include an estimated buffer for memory that is reserved by operating systems and that cannot be swapped out regardless of access
frequency. If you do not install a guest operating system, CapacityIQ overestimates memory demand. The memory appears undersized in the Undersized
Virtual Machines - List view because of this calculation.
Workaround: For virtual machines without guest operating systems, ignore the undersized status generated by the memory demand
CapacityIQ metrics assume that orphaned virtual machines are powered off virtual machines.
The data in the pie chart on the Common Virtual Machine Configurations - Distribution view might not appear properly if you change the Capacity Calculation Rules setting in the Usable Capacity global settings.
The default Capacity Calculation Rules setting is Use last known capacity. If you change the setting to Use actual capacity, the pie chart might exclude some data.
The information that does appear includes the counts of the unique virtual machine configurations observed within the analysis period, rather than the average or last known configurations.
In the Virtual Machine Capacity Usage - Trend view, the CPU Limit trending line can show a greater value than the Configured CPU trending line.
This information might seem inaccurate, but the CPU limit can be higher than the configured CPU. CapacityIQ calculates configured CPU as vCPU count x host CPU speed per core.
- Numbers for memory consumed are slightly higher than the numbers reported by vCenter Server. On the dashboard and on the Virtual Machine Capacity Usage - Trend view of a virtual machine, the numbers of memory consumed are slightly higher than the numbers reported by vCenter Server. This occurs because CapacityIQ adds ESX memory overhead to the consumed memory number for the virtual machine.
This issue affects the following pages and metrics:
- VM Memory Consumed - This number is slightly higher because CapacityIQ adds ESX memory overhead.
- Total VM Memory Remaining - This number is slightly lower because CapacityIQ subtracts ESX memory overhead.
- Virtual Machine Capacity Usage - Trend view
- VM Memory Consumed - This number is slightly higher because CapacityIQ adds ESX memory overhead.
When you try to open CapacityIQ on a Windows Server operating system, you see an authorization error in the vSphere Client that blocks you from viewing the page.
Workaround: Open Add or Remove Programs in the control panel, click Add/Remove Windows Components, and remove Internet Explorer Enhanced Security Configuration.
This workaround is specific to Windows Server operating systems, such as Windows Server 2003, and does not apply to other Windows operating systems, such as Windows XP or Windows 7.
If you continue to see the authorization error, update the internet security settings.
- In the control panel, select Internet Options.
This opens the Internet Properties dialog box.
- On the Security tab, click Trusted sites, click Sites, and add the CapacityIQ URL to the trusted zone.
- For that zone, click Custom Level to update security settings.
- Under the Miscellaneous settings, select Enable for the Allow META REFRESH setting.
- Under the Scripting settings, select Enable for the Active scripting setting.
- Click OK.
- Restart the vSphere Client.
The CapacityIQ plug-in does not load properly in the vSphere Client.
Workaround: If you have high security settings in Internet Explorer, select Tools > Internet Options, select the Security tab, and lower the security settings for untrusted servers.
New: The Capacity Overview report does not show units for cpu, memory, I/O, and other resources in the Trend and Forecast graphs.
CapacityIQ reports can only render Latin characters even though the corresponding views can render other characters. For example, Chinese, Japanese, Korean, and
Russian characters that appear in views do not appear in reports.
When you move a host in to and out of a vCenter Server 4.1 environment, the CapacityIQ interface fails and cannot display
data for the host.
Workaround: Close the vSphere Client and open a new instance of the vSphere Client to view the data for
The dashboard and views display dashes.
Workaround: Because dashes might stem from a number of possible causes, check the
vCenter CapacityIQ Online Library for information about the conditions that trigger dashes and the affected views. Some storage metrics
are unavailable because of ESX or vCenter Server version limitations. Disk I/O usage and datastore disk I/O capacity metrics are available with only ESX 4.1 hosts.
Even if you have a vSphere 4.1 environment, you cannot view NFS and iSCSI datastore metrics for hosts with ESX 4.0 or earlier versions. Datastore space and
host disk I/O usage metrics are available with ESX 4.0 or earlier versions. The VM Disk Space Provisioned metric requires vCenter Server 4.0 Update 1 or later.
When you expand a cluster with thousands of child objects in the inventory panel, you might not see the child objects immediately.
Workaround: Because the environment requires more time to process the number of objects, wait for a few minutes and expand the node again.
When you register CapacityIQ to a vCenter Server system that becomes permanently inaccessible, unregistering CapacityIQ through the Administration Portal generates an error and
blocks the ability to register another vCenter Server system.
Workaround: Open the CapacityIQ virtual appliance console and type ciq-admin register
--vc-server <vCenter_Server_name> --user <vCenter_Server_user_name> --password <vCenter_Server_password> --force from the command line.
To see more options for the ciq-admin command, type ciq-admin register help.
When you enable Fault Tolerance (FT) on a virtual machine, the primary and secondary virtual machines appear with the same name in the navigation tree.
When you remove a host or virtual machine, the data that CapacityIQ collects before the removal cannot be reconciled with the data that CapacityIQ collects after you add the host or virtual machine back to the datacenter.
Workaround: Disconnect and reconnect the virtual machine or host.
On the Virtual Machine Capacity - Summary view, you might see values, such as Time Remaining > 1 year, even when the Total Capacity, Deployed, and Remaining values are null.
This occurs only within the first two hours after the addition of an object.
Trend views might mishandle the display of information on the view page.
If you use a computer with the screen resolution set smaller than 1024 x 768, the text string that appears below the Data Window bar might be truncated.
Workaround: Change the screen resolution to a higher resolution when working with trend views.
What-if analysis cannot occur on a new standalone host that the vSphere Client is still adding to the inventory.
Workaround: Links for applying what-if scenarios to the host, either in the dashboard or in views, do not appear until an hour or two after the host is added to the vSphere Client.
If CapacityIQ performs a shutdown that is not initiated through the ciq-admin command-line interface, the Administration Portal might not indicate that the connection status is disconnected.
Workaround: Restart CapacityIQ and check the status.
When you select objects in the inventory panel, the objects might inaccurately appear to have child objects.
The following object types might display this issue:
- Hosts and clusters with only folders
- A datacenter with only folders
- A cluster with only resource pools
- A host with only resource pools
If you create an empty folder under All Datacenters, and attempt to expand All Datacenters in Capacity IQ, you see a Loading... indicator.
Workaround: Press F5 to refresh CapacityIQ.
- On some CapacityIQ pages, the client shows an Error loading view message.
CapacityIQ uses an Internet Explorer browser control for rendering its pages. This message might appear after upgrading CapacityIQ and the Internet Explorer cache is full.
Workaround: Clear the Internet Explorer cache. See Microsoft knowledge base article 260897 for instructions. If the problem persists after clearing the cache, contact VMware technical support.
- The wrong start time appears in the slider if you select a different time zone for the interval range start date. If the current time is Standard Time (non-Daylight Savings Time), and you select Daylight Savings Time for the interval range start date, the slider displays the wrong start time (non-Daylight Savings Time) when CapacityIQ reloads the data. The correct requested interval range appears in the view, but not in the slider. The start time appears as one interval before the actual requested start date.
Workaround: Set the start date to the correct date before updating the interval range for the view.
For more information about known issues with CapacityIQ, see the VMware Knowledge Base.
Top of Page
The following resolved issues are grouped as follows:
If you select the CapacityIQ virtual appliance in the inventory and navigate to the Console tab to configure time zone settings,
you must restart CapacityIQ at the end of the process. To restart CapacityIQ, log in to the Status tab of the
Administration Portal and click Restart.
The CapacityIQ documentation will include this information in the next release.
Rendering time for certain views might be slow.
This might happen when you select a cluster in the inventory panel, or when the selected object contains over 500 virtual machines. In these cases, the following views can take up to a minute to render in the user interface:
- Powered-Off Virtual Machines - List
- Idle Virtual Machines - List
- Oversized Virtual Machines - List
These views are available in report form on the Reports tab. CapacityIQ can start and process these reports in the background.
The Update release notes text box that appears during the update process in the Administration Portal is missing a step in the procedure to increase
the partition size.
Workaround: In the upgrade procedure to increase the partition size, add the (parted) rm 1 command.
# export http_proxy=<your_company_proxy:port>
# yum install parted
# service ciq stop
# service ciqadmin stop
# service postgresql stop
# umount /data
# parted /dev/sdb
# (parted) print
# (parted) rm 1
# (parted) mkpart primary 512B 268435456000B
# (parted) quit
# resize2fs -f /dev/sdb1
# mount /dev/sdb1 /data
# service postgresql start
# service ciqadmin start
# service ciq start
The prerequisites in the Deploy the Virtual Appliance and Deploy the Virtual Appliance with Thin Disks in vSphere 4.0 sections of the
vCenter CapacityIQ Installation Guide require clarification of the supported ESX host versions.
The prerequisites of the Deploy the Virtual Appliance procedure do not clarify that deploying CapacityIQ with thin disks involves either using vCenter Server 4.1 with ESX hosts version 3.5 or later or
seeing the Deploy the Virtual Appliance with Thin Disks in vSphere 4.0 section of the installation guide.
The prerequisites of the Deploy the Virtual Appliance with Thin Disks in vSphere 4.0 include an inaccurate requirement of ESX host version 4.0. You can use an ESX host version 3.5 with
VMFS 3 or a later version of the datastore.
The CapacityIQ documentation does not clarify that CapacityIQ does not account for RDM disks in
disk space capacity metrics and provisioned disk space metrics.
The CapacityIQ online help and online library describe conditions, such as
dependencies on ESX and vCenter Server versions, that trigger dashes in views. The documentation does not clarify that the VM Disk Space Provisioned metric requires vCenter Server 4.0 Update 1 or later. The online library notes that this metric
affects the Virtual Machine Capacity Usage - Trend view and the Cluster or Host Capacity Usage - Trend view.
The CapacityIQ online help does not clarify that outliers appear as hollow circles in the views. The vCenter CapacityIQ Online Library includes this information.
Top of Page