Unified Compute Platform 4.2 UCP Release Notes November 28, 2016 RN-92UCP016-18
Unified Compute Platform 4.2UCP Release Notes
November 28, 2016
RN-92UCP016-18
© 2014 - 2016 Hitachi Data Systems. All rights reserved.
No part of this publication may be reproduced or transmitted in any form or by any means, electronic ormechanical, including photocopying and recording, or stored in a database or retrieval system forcommercial purposes without the express written permission of Hitachi, Ltd., or Hitachi Data SystemsCorporation (collectively, “Hitachi”). Licensee may make copies of the Materials provided that any suchcopy is: (i) created as an essential step in utilization of the Software as licensed and is used in no othermanner; or (ii) used for archival purposes. Licensee may not make any other copies of the Materials."Materials" mean text, data, photographs, graphics, audio, video and documents.
Hitachi reserves the right to make changes to this Material at any time without notice and assumes noresponsibility for its use. The Materials contain the most current information available at the time ofpublication.
Some of the features described in the Materials might not be currently available. Refer to the most recentproduct announcement for information about feature and product availability, or contact Hitachi DataSystems Corporation at https://support.hds.com/en_us/contact-us.html.
Notice: Hitachi products and services can be ordered only under the terms and conditions of the applicableHitachi agreements. The use of Hitachi products is governed by the terms of your agreements with HitachiData Systems Corporation.
By using this software, you agree that you are responsible for:
1) Acquiring the relevant consents as may be required under local privacy laws or otherwise fromauthorized employees and other individuals to access relevant data; and
2) Verifying that data continues to be held, retrieved, deleted, or otherwise processed in accordance withrelevant laws.
Notice on Export Controls. The technical data and technology inherent in this Document may besubject to U.S. export control laws, including the U.S. Export Administration Act and its associatedregulations, andmay be subject to export or import regulations in other countries. Reader agrees tocomply strictly with all such regulations and acknowledges that Reader has the responsibility to obtainlicenses to export, re-export, or import the Document and any Compliant Products.
Hitachi is a registered trademark of Hitachi, Ltd., in the United States and other countries.
AIX, AS/400e, DB2, Domino, DS6000, DS8000, Enterprise Storage Server, eServer, FICON, FlashCopy,IBM, Lotus, MVS, OS/390, PowerPC, RS6000, S/390, System z9, System z10, Tivoli, z/OS, z9, z10, z13,z/VM, and z/VSE are registered trademarks or trademarks of International Business Machines Corporation.
Active Directory, ActiveX, Bing, Excel, Hyper-V, Internet Explorer, the Internet Explorer logo, Microsoft,the Microsoft Corporate Logo, MS-DOS, Outlook, PowerPoint, SharePoint, Silverlight, SmartScreen, SQLServer, Visual Basic, Visual C++, Visual Studio, Windows, the Windows logo, Windows Azure, WindowsPowerShell, Windows Server, the Windows start button, andWindows Vista are registered trademarks ortrademarks of Microsoft Corporation. Microsoft product screen shots are reprinted with permission fromMicrosoft Corporation.
All other trademarks, service marks, and company names in this document or web site are properties oftheir respective owners.
Contents
About this document 1Release highlights 1Related documents 3Supported and recommended limits 4Software requirements 6Included software 6
Hitachi 6VMware 6vSphere 5.x 7vSphere 6.0 7
Microsoft 7Linux 8
Image support 8SNMP support 8Issues with workarounds 9
Release issues 9Previous release issues 14
Issues with no workarounds 29Previous release issues 30
Getting help 34
Contents iiiUCP Release Notes
iv Contents
UCP Release Notes
About this documentThis document contains release notes for the 4.2 version of HitachiUnified Compute Platform (UCP). It describes supported limits,product documentation, and known issues.
Release highlightsThe following features and updates are new to UCP 4.2.
• Updates to the set server image and set cluster image operations:
¡ Added image as a required parameter for the API and CLI set serverimage and set cluster image operations.
¡ Introduced SetImageForServers, a new batch feature and API thatsets a single, selected image to a list of servers.
• Support for multiple UCPs in a single master vCenter:
¡ View and select multiple UCP Director plug-ins in the vCentervSphere Client.
¡ Right-click an object within the vCenter Inventory List and issue acommand to the appropriate UCP.
¡ For detailed information, see UCP Director 4.2 Administrator'sGuide
• Added new disaster recovery feature, Global Active Device (GAD):
¡ Supports DR between two storage arrays located at geographicallyseparate sites.
1UCP Release Notes
¡ GAD manages the protected data volumes using UCP DirectorOperations Center (DOC) for a single pane of glass user experience.
¡ Supports creating and maintaining stretch clusters across pairedsites.
¡ For detailed information, see UCP DOC 4.2 Administrator's Guide.
• Support for Enhanced Link Mode (ELM):
¡ Support for customers who want to scale out with multiple systemsby providing a single pane of glass experience.
¡ UCP automatically configures hosts, clusters, datastores andmanages resources without manual steps.
¡ Support for vMotion using L3 Networking, which allows for workloadsto be migrated between two UCP systems, or from a non UCP systemto a UCP system.
• Support for the new Hitachi VSP G1500 storage array.
• Support for the new Hitachi CB520XB2 compute blades:
¡ Full width compute blades to perform enterprise applications.
¡ Support for Symmetric Multiprocessing (SMP) configuration,allowing up to four blades to be combined as a single server.
¡ Support for storage port reservation.
• Updated parameter requirements for select APIs.
¡ For a detailed list, see UCP Director 4.2 API Reference Guide.
• Updated parameter requirements and examples for select CLIs.
¡ For a detailed list, see UCP Director 4.2 CLI Reference Guide
• Added VSP Gx00 unified support:
¡ Support for coexistence of NAS blade on UCP stack.
• Improved UCP Director UI to ensure an effortless user experience.
Release highlights
2UCP Release Notes
Related documentsThe following documents contain additional information about UCP:
• UCP Pre-Installation Requirements and Configuration Guide —Contains information and procedures you need to be aware of for asuccessful UCP installation.
• UCP Network Architecture and Configuration Guide — Containstechnical reference information about the networking architecture ofUCP systems and provides procedures to help you configure the systemto work in your network environment.
• UCP Director Administrator's Guide — Contains technical and usageinformation for UCP and UCP Director. Describes how to administer UCPDirector through UCP Director web interfaceUCP Director web interfacewith both VMware vCenter and Microsoft SCVMM.
• UCP Director API Reference Guide — Describes how to use the UCPDirector API.
• UCP Director CLI Reference Guide — Describes how to use the UCPDirector CLI.
• UCP Director Third-Party Copyrights and Licenses — Containscopyright and license information for the third-party softwaredistributed with or embedded in UCP Director.
• UCP DOC Administrator's Guide — Contains technical and usageinformation for Unified Compute Platform Director Operations Center(UCP DOC). Describes how to administer UCP DOC through UCP DOCConsole.
• UCP DOC API Reference Guide — Describes how to use the UCP DOCAPI.
• UCP DOC CLI Reference Guide — Describes how to use the UCP DOCCLI.
Related documents
3UCP Release Notes
Supported and recommended limitsTechnical LimitationsUCP supports the minimum and recommended/maximum values listed inthe following table. Maximum values can indicate either a technicallimitation or an acceptable performance threshold.
Item Minimum Recommended/Maximum
Fibre Channel zones 1000 recommended
Host storage domains 512 recommended
Servers (HB) 2 Brocade Ethernet: 64
Cisco Ethernet: 128 (2 port)Cisco Ethernet: 88 (4 port)
Cisco FCOE: 24
Servers (XB) 1 Brocade Ethernet: 32
Cisco Ethernet: 64 (2 port)Cisco Ethernet: 44 (4 port)
Cisco FCOE: Not supported
Storage Pools for VSP Fx00
VSP-Fx00 storage comes in three standard sizes for each model: 1.6 TB, 3.2TB, and 6.4 TB. The values listed in the table, reflect the three availabledrive sizes and their storage capacities in relation to the RAID configurationselected.
All configurations have a management capacity of 2 TB.
Supported and recommended limits
4UCP Release Notes
RAID Configuration Size
Available Compute Capacity
(Small, Medium, Large)
RAID 7+1
¡ 1.6 TB
¡ 3.2 TB
¡ 6.4 TB
11.2 TB
22.4 TB
44.8 TB
9.2 TB , 20.4 TB, 54 TB
20.4 TB, 42.8 TB, 110 TB
42.8 TB, 87.6 TB, 222 TB
RAID 6+2
¡ 1.6 TB
¡ 3.2 TB
¡ 6.4 TB
9.6 TB
19.2 TB
38.4 TB
7.6 TB , 17.2 TB, 46 TB
17.2 TB, 36.4 TB, 94 TB
36.4 TB, 74.8 TB, 190 TB
RAID 4+4
¡ 1.6 TB
¡ 3.2 TB
¡ 6.4 TB
6.4 TB
12.8 TB
25.6 TB
4.4 TB , 10.8 TB, 30 TB
10.8 TB, 23.6 TB, 62 TB
23.6 TB, 49.2 TB, 126 TB
RAID 3+1
¡ 1.6 TB
¡ 3.2 TB
¡ 6.4 TB
4.8 TB
9.6 TB
19.2 TB
7.6 TB , 17.2 TB, 46 TB
17.2 TB, 36.4 TB, 94 TB
36.4 TB, 74.8 TB, 190 TB
RAID 2+2
¡ 1.6 TB
¡ 3.2 TB
¡ 6.4 TB
3.2 TB
6.4 TB
12.8 TB
4.4 TB , 10.8 TB, 30 TB
10.8 TB, 23.6 TB, 62 TB
23.6 TB, 49.2 TB, 126 TB
Supported and recommended limits
5UCP Release Notes
Software requirementsThe following software and screen settings are required to use UCP Directorweb interface:
Software/Setting Version/Value
Java JRE 1.7 Update 21 or greater
1.8 Update 45 or greater
Internet Explorer 11
Screen resolution 1024x768 or higher
Included softwareThis section details the software that is included with UCP 4.2. The softwarelisted in the UCP Director Third-Party Copyrights and Licenses guide hasbeen omitted from the list.
Hitachi
The following Hitachi software is included with UCP:
Software Version
Command Control Interface (CCI) 01-38-03/00
Hitachi Compute Systems Manager (HCSM) 8.5
Hitachi Device Manager (HDvM) 8.5
Hitachi TuningManager (HTnM) 8.5
HTnM Agent for RAID 8.5
Hitachi vSphere APIs for Storage Awareness (VASA) N/A
Site Recovery Adapter (SRA) 2.01.04
UCP Director 4.2.0.5830
UCP Director Operations Center 4.2.0.5830
UCP Toolbox 4.2.0.5830
VMware
The following VMware software is included with UCP:
Software requirements
6UCP Release Notes
vSphere 5.x
Software Version
Enterprise Plus No longer distributed
ESXi No longer distributed
Site Recovery Manager (SRM) No longer distributed
vSphere 6.0
Software Version
Enterprise Plus 6.0u2a
ESXi 6.0u2
SRM 6.1
Microsoft
The following Microsoft software is included with UCP:
Software Version
AppFabric for Windows Server 1.1
ASP.NETMVC 3.0
Internet Information Services (IIS) 8
Internet Explorer (IE) 11
Microsoft SQL SQL 2014 w/ Downgrade rights to SQL2012
.NET 4.5
Windows Deployment Services (WDS) 6.2
Windows Server 1 2012 R2 Datacenter UR5
WindowsWebDeploy 3.0
1 Used for installing Management servers and Management VMs.
Included software
7UCP Release Notes
Linux
The following Linux software is included with UCP:
Software Version
CentOS for blade verification 7.1
CentOS for UCPUtility VM 7.1
CentOS shippedwith WDS 7.2
Image supportUCP supports images using the following OS versions:
Supported element Version
CentOS, Red Hat Enterprise Linux HB1: 6.61
HB2, HB3: 6.6, 7.2
HB4: 7.2
XB2: 7.2
Suse XB2: 12.1
ESXi 5.1 U3, 5.5 U3b, 6.0 U2
Windows Server Drivers 2008 R2 SP1, 2012 and 2012 R2 SP1
1 HB1 only supports CNA firmware 10.2 , so 10.2 Linux driver is required.
SNMP supportUCP supports SNMP monitoring of the hardware. With Ethernet and FibreChannel switches, UCP Director receives traps from the switches directly.When working with servers or storage, UCP Director receives traps from thecorresponding element manager, HCSM and HDvM respectively. UCP usesthe following SNMP versions:
Supported element Version
Ethernet switches 3
Image support
8UCP Release Notes
Supported element Version
Fibre Channel switches 3
HCSM 2c
Storage system 1
Issues with workarounds
Release issues
The table below lists new issues that are relevant to this release forwhich workarounds exist. The issues are listed in order by referencenumber.
# Description
469 Baremetal service template job reports completion before OSdeployment completes
In UCP Director, when a baremetal service template is applied (withoutadditional volumes provisioned), the service template job reports a successbefore the OS has completed deployment. As a result, if the user tries toperform any volume provisioning immediately, that job will fail.
Workaround: Monitor the OS deployment job in the server console andperform operations once the OS deployment successfully completes.
1409 Occasionally, UCP vROPS dashboard fails to show data
When the UCP Management Pack 2.0 for VMware vRealize Operations isdeployed, due to a bug in vSphere, the dashboard occasionally does not showany data.
Workaround: Edit and save the dashboard again. This enables the data to bedisplayed on the dashboard again.
Issues with workarounds
9UCP Release Notes
# Description
3922 ESXi deployment fails occasionally due to timeouts caused byautodeploy multicast requests overloading the network
Due to a bug in VMware's implementation of iPXE, the multicast requestsoverload the network and causes timeouts during ESXi deployment. Theservers subsequently fail to load ESXi..
Workaround: Change the ucptramp file to use the nomulticast file of iPXE.
1. SSH onto the Utility VM.
2. Edit the tftpboot/ucptrampby changing the current line,
set bootfile_autodeploy undionly.kpxe.vmw-hardwired
to the following line,
bootfile_autodeploy undionly.kpxe.vmw-hardwired-nomcast
4314 Service templates cannot be applied to blades with existing serverprofiles
Using the UCP Director UI, it is not possible to apply service templates to bladeswith existing server profiles. The list of available server profiles does not showthe currently applied profile for that blade.
Workaround: First remove the server profile from the desired blade, then beginthe service template apply.
4685 Applying service template sometimes fails with create deploy ruleserror
Apply service template sometimes fails because it cannot create new autodeploy rules in vCenter.
Workaround: Verify that there are no hosts attempting to boot. Restart the"VMware vCenter Auto Deploy Waiter" service on the vCenter VM.
5233 API may fail with 400: Bad Request error
When using an API, you may see 400: Bad Request error if the parameters inthe API request body are not correct. The API request body is case-sensitive soparameters in API request body should be written exactly as shown in theUCP Director API Reference Guide.
Workaround: Carefully check the parameters, such as "ID" or "VLAN," in theAPI request body. Correct the parameter name and issue the request again.
Issues with workarounds
10UCP Release Notes
# Description
5771 Provisioning to a server may fail due to incorrect access rights in theDHCP-related directory
Applying service template may fail because the access permissions of the/etc/dhcp directory, and its files, are reset after updating DHCP services.
Workaround: Change the permissions of the directory /etc/hdcp and files underthe directory as follows:
chgrp -R admin_group /etc/dhcp
chmod -R 775 /etc/dhcp
chown -R icpscp /etc/dhcp
6215 "Change Cluster Image" and "Deploy Cluster Image" cluster actions aremissing from UI
"Change Cluster Image" and "Deploy Cluster Image" actions are missing fromthe cluster action list in vCenter.
Workaround: Use UCP CLI to perform the actions. The command to change thecluster image is Set-UCPClusterImage, and the command to deploy clusterimage is Deploy-UCPServer.
6711 In large UCP environments, the UCP inventory occasionally shows lesseritems than actual inventory in vSphere web client
Due to a vSphere web client issue, UCP inventory will occasionally show lesseritems than actual inventory.
Workaround: Sort the elements by a different parameter. If this action does notresolve the issue, close the browser session, clear the history and cookies, andretry the operation.
6951 Cannot apply HB1 and HB2 server profile to the other blade types
In this release, applying an HB1 server profile to an HB2 blade, or applying anHB2 server profile to an HB1 blade, is not allowed.
Workaround: Apply a server profile to a blade with the samemodel type as theserver profile.
Issues with workarounds
11UCP Release Notes
# Description
7472 Remove Server Profile job fails with"RemoveServerProfileFromServerFailedErrorEvent"
When running Remove Server Profile, either individually or as part of a batchoperation, occasionally you may see the job fail with error event"RemoveServerProfileFromServerFailedErrorEvent". This can intermittentlyoccur, especially when concurrent jobs are running at the time of the RemoveServer Profile operation.
Workaround: Verify that the association between the server profiles andservers no longer exists. If the association still exists, retry the Remove ServerProfile operation. If the association does not exist, the job completed thenecessary actions to remove the profile, and you can continue.
7675 During creation of a GAD replicated volume pair, the job fails but the groupis shown as being in the PAIR state in the UI
During the GAD replicated volume pair creation, UCP Director creates thereplicated volume pair. Then it tries to make it uniform by remotely attachingservers across the two UCPs, to the storage at the other site. If the uniformoperation fails for any reason, the job fails, but the replicated volume pairremains in PAIR status.
Workaround: Delete the replicated volume pair and retry the operation.
7776 Registration of UCP Director in VMware model fails if Compute Datacentername includes characters "%/\["
When you include characters%/\[ in the Compute Datacenter name, theregistration of UCP Director fails. This is a vCenter limitation.
Workaround: Do not use the characters%/\[ in the Compute Datacentername.
7882 vCenter occasionally fails to apply host profile to cluster node during clustercreation, leading to the host getting disconnected from vCenter
During the cluster creation operation in UCP Director, while applying the hostprofiles to the nodes in the cluster, vCenter occasionally fails to apply the hostprofile and leaves the host in a disconnected state. The host is still accessiblethrough SSH (if enabled) and the console. Any attempts to reconnect fromvCenter will fail.
Workaround:Reboot the ESXi host to reconnect to vCenter. In case the clustercreation job has failed at the apply host profile step, the host will have to beadded to the vDSmanually, and the vSwitch will have to be removedmanually.Alternatively, the cluster can be cleaned up by removing the server profile andthe cluster creation operation is then retried.
Issues with workarounds
12UCP Release Notes
# Description
7903 Unable to refresh Server Inventory when chassis and blades are down
Failure to retrieve information from chassis and blades results inUCP terminating processing formation for those resources. This then fails theRefresh Server Inventory job.
Workaround: Resolve the issue with the chassis and blades which are down, orremove the chassis and blades from the UCP resource group in HCSM until theissue has been resolved.
7929 Update Active Image action is missing from UCP Director UI
Update Active Images action is missing in the refreshed UCP Director UI
Workaround:Use the CLI command Update-UCPActiveImage to update theimage.
7943 VLAN not correctly configured on Brocade VDX switches during clusterdeployment
The default vCenter VDS VLAN range is 0-4094, however the allowed VLANs forthe Brocade VDX switches are 2-1001, and 1003-3961. Therefore, the clusterdeployments will succeed but fail to configure the VLAN range for the VDXswitches. This causes problems with cluster actions such as vMotion.
Workaround:When creating a VDS in vCenter for a Brocade UCP appliance,manually change the VLAN trunk range to be within the allowed VLAN ranges of2-1001, and 1003-3961.
8027 Windows deployment may fail when a blade server has a large amount ofmemory
DeployingWindows to a blade with a large amount of memory may fail becauseof the lack of C:\ volume capacity. Windows needs a paging file whose size isproportional to the blade memory size. It is located in C:\ volume as a hiddenfile.
Workaround: Change the boot volume size to 16GB, plus the size of bladememory, and retry.
8054 If a baremetal OS deployment fails due to a Fiber Channel zone creationerror, the zones are not cleaned up during rollback
If UCP Director fails to create fiber channel zones on one or more switches in thesystem during the baremetal OS deployment of Linux or Windows, the rollbackis completed without cleaning up the zones in the fiber channel switch.
Workaround: Manually clean up fiber channel zones created during the failure.
Issues with workarounds
13UCP Release Notes
# Description
8135 Provisioning of multiple UCP servers with additional volume provisioningfails due to volume name conflicts
During OS provisioning of multiple servers with additional volumes provisioned,the jobmay occasionally fail due to volume name conflicts. This is primarybecause UCP is performing OS provisioning and storage provisioning in parallelfor multiple blades. Due to timing issues, occasionally the volume crating withidentical names are issued at the same time. During this time, the volumes arecreated, but the second name is crated with an appended (1) in its name. Thissecond volume name fails validation , causing the job to fail.
Workaround:Perform the storage provisioning through UCP Director afterserver OS provisioning.
8162 Advanced Filter options cannot be selected in UCP Director
Rendering issues occasionally occur when UCP Director pages come underheavy load. The Advanced Filter options may not be able to be selected.
Workaround: Close the Advanced Filter andwait a few seconds beforereopening, or refresh the page.
8286 SCVMM - Detach Volume from Hyper-V Cluster job fails even after volume isunmounted from cluster nodes
Detach Volume job fails on the Hyper-V Cluster job even after the volume isunmounted from the cluster nodes: "Volume detach operation failed asvolume: <Id> is still mounted to the host(s): <hosts> in the cluster. Pleaseunmount from host then attempt volume detach operation again".
Workaround:Before running the Detach Volume operation on UCP and after thevolume is actually unmounted from cluster/servers, run the Server and StorageInventory Refresh operation. Run the Detach Volume operation again.
Previous release issues
The table below lists issues that were relevant to the 4.x releases forwhich workarounds exist. The issues are listed in order by referencenumber.
Issues with workarounds
14UCP Release Notes
# Description
472 UCP storage operations become inaccessible when all Hyper-V hostsare powered off
Under rare circumstances when ALL the Hyper-V hosts are powered off, UCPstorage operations become inaccessible and cause UCP Director to throw errors.This is due to the fact that UCP Director cache gets invalidated and needs to bere-built.
Workaround: Make sure one of the Hyper-V hosts is online for UCP Director tore-build the cache.
526 Create/Apply Server profile does not check if the IP being used for itis already in use
When a server profile is created and then subsequently being applied, it does notcheck if the IP address being applied is already in use. In case the IP address isalready in use, the job fails.
Workaround: Manually check that the IP address is not being used beforeapplying server profile
547 Windows HyperV service template deployment fails in anenvironment with a shared external AD if an IP is reused
In a UCP Director HyperV environment deployedwith an shared external activedirectory, the HyperV deployment fails if a previously used IP is used duringdeployment. This happens because while a Forward Lookup record issuccessfully created for the host during deployment, the corresponding ReverseLookup record is not created (if a pre-existing Reverse Lookup record exists). Sowhen UCP polls for the Host, the only item returned is for the old ReverseLookup record. This causes the apply service template job to fail.
Workaround: Delete existing Forward Lookup entries for IP address eitherwhen deprovisioning the host, or before deploying a service template.
564 Cannot log out of Datacenter Operations Center UI
It is not possible for a user to log out of the Datacenter Operations Center (DOC)UI. The user stays logged in even after the browser is closed.
Workaround: Clear browser history. If using a chrome browser - force closethe chrome process.
Issues with workarounds
15UCP Release Notes
# Description
713 HTTP action failed due to incorrect credentials (negotiate)
When performing an action through the Workstation VM with a user that doesnot have appropriate UCP credentials the UCP system will refuse an action fromthat user. The following error will show: "Unkown Error: The HTTP request isunauthorized with client authentication scheme "Negotiate". Theauthentication header received from the server was "Negotiate, NTLM".
Workaround:When using the Workstation VM for accessing UCP,make sureto log on to the VM with the same credentials that you'll use to authenticateUCP login. This may require first logging into the Workstation VM as domainadministrator and then granting your user the rights to remote-desktop intothe VM.
791 Running Attach-UCPVolume CLI command without passingparameters does not prompt for mandatory parameter 'serverId' or'clusterId'
While running the UCP CLI Command to attach volumes to a cluster or a hostwithout providing any parameters, the CLI command does not prompt formandatory parameter 'clusterId' or 'serverId', causing the command to fail.
Workaround: Use the CLI commandwith the required parameters as specifiedin the example for Attach-UCPVolume. It can be found by running the followingcommand: Get-Help Attach-UCPVolume -example
929 Username column is blank in UCP Switch Events page on VMwareWeb Client
The username column is blank within a Switch Summary monitoring pagewhen viewing events on said switch. This issue is seen only on the VMware webclient
Workaround: View the Switch summary page on the VMware Thick Client
940 Individual monitoring indicators set to Unknown still show up in UCPafter elements have been offboarded causing DOC to show incorrectnumber of Unknown indicators
If any specific indicators on an onboarded element in UCP are set to Unknownand the element is offboarded, the individual element indicators still linger onand show 'Unknown', causing DOC to show incorrect number of Unknownindicators.
Workaround: Set all indicators of onboarded elements to 'Monitoring' beforeoffboarding elements.
Issues with workarounds
16UCP Release Notes
# Description
1205
Unable to add online repositories for VMware in UCP 4.0
There is a regression with VMware code that causes PowerCLI to fail when anexternal VMware online repository is added. This leads to PowerCLI failureswhich can cause UCP API failures. More details about the issue can be foundhere: https://communities.vmware.com/thread/489560 Resolution: - Removeall VMware online repositories
Workaround/Recommendations going forward: Do not add any VMwareonline repositories into UCP
1317
Apply service template fails with error message "UCP failed to updateDHCP configuration."
This issue is seen when the UCP Utility VM which acts as the DHCP server in theUCP Management stack goes into recovery mode and the filesystem is in a readonly mode. Because of this, UCP cannot create DHCP entries for a blade whileapplying a service template and the job fails continuously.
Workaround: Reboot the UCP utility VM and verify that the file system ismounted in readwrite mode.
1347
Locales for DOC, UCP, and SQL VM's must be consistent in an HURDisaster Recovery configuration
Locales for DOC, UCP, and SQL VM's must be consistent in an HUR DisasterRecovery configuration. When performing journal operations all VM's must bethe same locale - English (US) to perform journal operations within an HURconfiguration.
Workaround: Configure DOC, UCP, and SQL VM to be the same locale - English(US)
Issues with workarounds
17UCP Release Notes
# Description
1599
vSphere Distributed Switch VLAN and MTU warnings when rebootinga compute blade
These warnings may be visible in the vSphere health check UI after a computeblade has been rebooted.
Workaround/Resolution: Ensure that all portgroups in the virtual distributedswitch with different VLANs have the same MTU in the physical switch becauseESXi will not detect the MTU mismatch of full paths and Jumbo Frame packetsmight forwarded to other physical switch ports which are out of the virtualdistributed switch. At those ports, there is a risk that the Jumbo Frame packetsmight be dropped if that port and VLAN do not enable Jumbo Frames.
Additionally, ensure the number of MAC addresses to be generated by thehealth check will be less than the size of the physical switch(es) MAC table.Otherwise there is a risk that the switches will run out of memory, withsubsequent network connectivity failures.
3102
ESXi host disconnected from vSphere client cannot be powered offfrom UCP UI
If a particular server is disconnected from vSphere client, when we try to powerit off from the UCP UI, it fails.
Workaround: Add the host back in vSphere inventory and then issue thepower off command. OR If the host cannot be added back to vSphere inventory,the UCP API or CLI can be used to force power off the server.
3169
DOC UI Allows user to select non-deployed blades as secondary hosts
In the DOC UI, while setting up replication between two hosts, the host at thesecondary site is allowed to be selected even if it is not deployed. This kicks offthe job, but fails since the secondary site does not have the host deployed.
Workaround: Only select hosts on the secondary sites which are deployed.
3746
UCP Director API for create and attach volumes does not immediatelyvalidate storage ports when specified manually
When storage ports are specifiedmanually while using the UCP Director API endpoint for create and attach volumes, the job is kicked off without verifying if theselected ports match UCP Director criteria for storage ports. As a result, the usercannot know until the job fails if the selected ports met the criteria or not. As aresult, the storage volume created as a part of the job remains in the pool aswell.
Workaround: Refer to the UCP Director API Reference Guide for create andattach volume to see UCP's storage port requirement and view the example.
Issues with workarounds
18UCP Release Notes
# Description
3992
While creating a server profile from the identity pool, UCP Directordoes not check if an identity is already in use.
When creating a server profile by using identities such as UUID, MAC address,WWNN,WWPN and IP address, UCP Director does not check if the randomlyselected resource is already in use. Though the probability of the same addressbeing selected is low, this has been seen occasionally. As a result, when theserver profile is applied, it fails with an error message that the identity is in useby other server.
Workaround: Delete the server profile with the duplicate identity
3997
Pressing Enter after entering the details while adding a site in DOCdoes nothing.
While adding a site in DOC, once all the necessary information is entered, hittingenter on the keyboard does not submit the information to DOC.
Workaround: Click on the 'Ok' button to submit the information to DOC.
3997
Pressing Enter after entering the details while adding a site in DOCdoes nothing
While adding a site in DOC, after all the necessary information is entered, hittingenter on the keyboard does not submit the information to DOC.
Workaround: Click the OK button to submit the information to DOC.
4088
Cannot reuse service templates with formatted volumes
When creating a new service template, UCP allows creation of raw or formattedvolumes which can be attached as a part of the apply service template workflow.Once this service template is deployed, in case the same service template isredeployed, though the OS deployment succeeds, the volume provisioning willfail due to conflicting names for formatted volumes, causing the apply servicetemplate job to fail.
Workaround: Re create formatted disks with unique name and attach them tothe deployed hosts.
4134
UCP log generation feature cannot handle invalid windows file namefor log bundle
The UCP log bundle generation feature does not verify invalid file names such asCON, PRN, AUX, NUL, COM1 - 9, LPT1 - 9
Workaround: Use valid windows file name for log bundle
Issues with workarounds
19UCP Release Notes
# Description
4182
Move server profile does not move associated trunk VLANs
When a server profile is moved from one blade to another, though the nativeVLAN is moved, the associated trunk VLANs are not moved.
Workaround: Identify all trunk VLANs before moving the server profile andapply them using configure switch VLAN after the server profile is moved.
4414
Concurrent storage operations using different intended use optionson the same storage ports may cause one of the storage operationsto fail
Under rare circumstances when two storage provisioning operations involvingthe same storage ports andmultiple intended use types (that are performedsimultaneously), one of the storage operations will fail. This is because of alocking condition where one operation which is creating the Host StorageDomain, locks it and the second operation cannot re-use the same Host StorageDomain.
Workaround: Re-run the storage provisioning operation.
4969
Cannot detach non-boot volumes from a Hyper-V host through UCPin SCVMM environment when it is powered off and not in SCVMMinventory
Under rare circumstances, e.g. when Hyper-V host is powered off in an SCVMMenvironment and removed from SCVMM inventory, UCP cannot remove non-boot volumes from the host.
Workaround: Power the host on and re-add the host to the SCVMM inventoryto be able to detach volumes from the SCVMM host.
4984
UCP storage operations become inaccessible when all Hyper-V hostsare powered off in an SCVMM environment
Under rare circumstances when ALL the Hyper-V hosts are powered off, UCPstorage operations become inaccessible and cause UCP Director to throw errors.This is due to the fact that UCP Director cache gets invalidated and needs to bere-built.
Workaround: Make sure one of the Hyper-V hosts is online for UCP Director tore-build the cache.
Issues with workarounds
20UCP Release Notes
# Description
5026
UCP Director contextual actions are disabled in SCVMM
When using SCVMMwith UCP Director, it is not possible to navigate to Fabrics>Select Elements > Select UCP Director contextual actions, e.g. Fabrics > AllHosts > Right click and perform actions.
Workaround: Directly open the UCP Director plugin to perform said actions.
5394
VLANS for VSAN are not set correctly.
If the user changes the VLAN ID for VSAN manually, UCP Directory mayconfigure incorrectly. This may lead to the loss of redundancy of storage traffic.It is not recommended tomanually change the VLAN ID for VSAN.
Workaround: Manually fix the VLAN ID for VSAN. The correct setting is that twoVLAN IDs for VSAN are set to two switches respectively.
11949
Expanding the datastore operation through UCP fails intermittentlyWhen expanding a datastore the underlying volume gets expanded but thedatastore expansion fails sometimes. This is a known bugwith the VMware API.
Login into the host using the vSphere client and follow VMware documentationon how to expand storage. For more information see this link:
http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1017662
12329
When an ESXi image is added to a UCP image repository, it cannot have thesame name as an existing ESXi image in UCP inventory.Displaying two images with the same name is not supported as this is alimitation in vCenter autodeploy rules.
If the image you plan to add has the same name as an existing image, rename itusing the vSphere PowerShell CLI. The image cannot simply be renamed inWindows since vSphere will not recognize the image afterwards.
12330
User can still make REST calls when permissions have been revokedIf UCP related permissions for example - network administrator privileges arerevoked from a user, the user will still be able to make network related calls tillabout 30 minutes from the time the privileges were removed. This issuehappens because UCP cache refresh is scheduled at a 30 minute interval.
Wait for 30 minutes and try again.
Issues with workarounds
21UCP Release Notes
# Description
12870
Granting UCP permissions in vCenterGranting UCP permissions to a user/group in vCenter does not work if the nodethe permissions are granted in are anything other than the vCenter root.
Grant the permissions in the vCenter root.
12982
Refresh Image inventory might fail if you have same image (differentcasing) in multiple repositoriesIf refresh image inventory failed because of an unexpected error, please verifythat the same image (zip file) is not present in multiple repositories (same filename different case).
Please remove one of the duplicate images from one of the repositories orrename the image profile and run Refresh Image Inventory again.
13192
UCP does not immediately report the correct total and used capacity for avolumeIf a volume is expandedmultiple times UCPmight not report the volumecapacity correctly as HDvM does not read the correct capacity from the array.
Use HDVM UI to resync with the array. It should be noted that forcing HDVM toresync with the array will degrade performance of the array for several minutesup tomore than an hour.
13571
UCP Director is not able to get array reservation for a long timeIf the SVP on the VSP is put in a modify mode by a customer service engineerand the mode is not released, UCP Director will be stuck in a endless loop tryingto get the reservation from the array.
Customer engineer should NEVER put SVP in modify mode when storageoperations are running on UCP.UCP will only be able to perform storageoperation when the SVP is in the view mode.
14678
Changing the boot type of a server may succeed immediately after the Bladeis powered ONIf the boot type for a server is changed too soon after the Blade is powered on,HCSM and UCP Director may be out of sync on the power status of the blade andUCP Director may allow changing boot type between custom boot and ESXistateless.
Wait for the blade to be powered ON completely and ensure that UCP shows thecorrect power status before attempting to change the boot type for the server.
Issues with workarounds
22UCP Release Notes
# Description
14697
After recovery from unexpected network outage, the power status of someblades may be 'unknown'When UCP regains connectivity to HCSM, it may not pick up the correct powerstates of some blades. Even after a server inventory refresh, the blades' powerstates may still be reported as 'unknown.' An additional side affect is that mostof the Hitachi server actions through the web client will not be available.
1.) Open a web browser to HCSM and log in as ucpadmin
2.) Go to the Resources tab
3.) Under "General Tasks," click Refresh\Refresh All Chassis and Servers
4.)When the task completes, return to UCP Director web interfaceUCP Directorweb interface and click Refresh Server Inventory in the UCP Server Inventorytable
5.) The UCP display of current power states for all blades should be corrected
15051
Restoring the configuration on a Brocade switch can cause the switch to notsend any trapsIf users try to take a backup of the switch and restore back the configuration onthe Brocade switch, the SNMP agent on the switch will get removed.
The next topology scan will fix the issue. It will detect that there is no SNMPagent on switch and create the SNMP agent.
15232
Running Ethernet switch inventory refresh enables CDP feature on Ethernetswitches even though the feature is disabled on switchesThe UCP orchestrator service settings are stored in its web.config onUCPManagementVM. One of settings in this file is 'enableCDPonVirtualSwitch'.This has default value of 'true' and Ethernet switch inventory refresh operationenables CDP setting on switches based on 'enableCDPonVirtualSwitch' setting.
Confirm the user installation does not require enabling CDP on the Ethernetswitches. Change the 'enableCDPonVirtualSwitch' setting to 'false' in UCPorchestratorservice's web.config. Rerun Ethernet switch inventory refresh.
Issues with workarounds
23UCP Release Notes
# Description
15242
If VPC Domain can disable on one of the Cisco 5K switches, it will causeboth switches to become the Cisco switches to get disconnected fromexternal networkIf VPC domain gets disabled on one of UCP's Cisco switches due to some out ofband activity, the other switch will remain connected to the network outside ofthe appliance till the next topology scan. However, on UCP's next topology scan,UCP will recreate the VPC peer link between UCP's two Cisco switches. This willcause the VPCs to become inconsistent between UCP's Cisco switches, and alluplinks to the outside network will be disabled.
Manually recreate VPCs to the outside network via a SSH client.
15305
If Fibre Channel switch refresh happens during the same time as a serverreboot then the FC switch cache might get invalidated for that serverDuring server reboot, if a FC switch refresh happens concurrently then the FCswitch cache might remove zones for that particular server from the cache. Thismight lead to storage attach operation failing immediately after the host comesback up.
Refresh FC switch inventory after the server reboot, this will update the cachewith the zone information for the server.
15948
vSphere web client runs older version of UCP after re-installIntermittently after an uninstall and reinstall of UCP web client might still runthe older version of UCP.
Delete "com.hds.ucpdirector.service" and "com.hds.ucpdirector.ui" from folder"C:\ProgramFiles\VMware\Infrastructure\vSphereWebClient\server\work\org.eclipse.virgo.kernel.deployer_3.0.3.RELEASE\staging\global\bundle" and re-registerUCP.The new version of UCP web client will get picked up.
16139
Trusted domain users cannot be restricted to vCenter resources at any levellower than the root.When a trust relationship is established between an existing domain and UCP,users from the existing domain can be added to administrative OU groups withinthe UCP domain. However, their privileges to vCenter objects cannot berestricted to any datacenter, folder, or cluster.
Users from trusted domains must always be granted access at the root level ofvCenter. Individual privileges can be restricted, but this will apply from the root-level on down the inventory tree.
Issues with workarounds
24UCP Release Notes
# Description
16197
Configure host network when switches are in initializing state, fails but jobevent does not clearly specify whyConfigure network on server host requires the Ethernet switch inventory in UCPneed to have complete information about host to Ethernet switch connectivity.
Refresh Ethernet switch inventory, wait for 30 seconds and then rerun'configure host network'
16319
Setting weak SNMP passwords for Cisco Ethernet switches succeeds task,but fails to configure SNMP settings on the switchOn a UCP instance, a user can input a weak authentication/privacy password forCisco Ethernet switches. The task will report as success, but the switch will notbe configuredwith the supplied password. This will cause all SNMP collectiontasks to fail (SNMP traps, performance, health) for the Ethernet switch.
Use an Authentication/Privacy Password that meets the following criteria:
1.) It cannot be based on a dictionary word - example: "Password1"
2.) It MUST include multiple characters of at least three of the following classes:lower case letters, upper case letters, digits and special characters
3.) It cannot have repeated characters of three or more times consecutively
16661
Detach/Delete volume fails after tearing down replication group.When a replication group is deleted or a replicated volume is removed from areplication group, users may not be able to detach and/or delete the underlyingvolumes as the volumes still believe that they are being replicated.
Wait up to 24 hours for for HDvM to resync with the storage array or use HDVMUI to resync with the array. It should be noted that forcing HDVM to resync withthe array will degrade performance of the array for several minutes up tomorethan an hour.
17037
REST call to clusters/{clusterId}/networks incorrectly reports resource notfound when network cache is not readyAPI call to get cluster networks using REST call to clusters/{clusterId}/networksincorrectly reports resource not foundwhen network cache is not ready.
Run Ethernet switch inventory refresh andwait 5 minutes
Issues with workarounds
25UCP Release Notes
# Description
17490
Removing HUR replication group sometimes fails to remove journalRemoving replication group or replicated volume pair on HUR sometimes failswith error that "Cannot delete journal {journalId} since the journal is beingused for replication".
Follow these steps to manually remove the journal volumes:
1.) Log into HDvM andmanually refresh the Storage system.Wait for the taskto complete
2.) In HDvM, obtain the serial number of the storage system
3.) Open a PowerShell session using an account with UCP Admin role or vCenterAdministrator role
4.) Load the Hitachi.UCP.CLI
5.) Obtain the UCP instance with Get-UCPDirector CLI cmdlet
6.) Remove the Journal volume with the Remove-UCPJournal cmdlet, using thestorage system id obtained earlier, and the Journal ID as shown in the previousfailure event
18125
Setting DPM (Distributed Power Management) to Automatic can cause thehosts to stay in Suspended stateConfiguring an HA/DRS cluster with DPM (Distributed Power Management) setto Automatic can cause the hosts to go in a bad state. DPM causes unutilizedhosts to go into "Suspended" mode. However, because UCP uses ESXi stateless,there is no disk to resume from and hosts cannot come out of Suspended state.
Do not use the 'Automatic' setting when configuring Distributed PowerManagement for a cluster of stateless ESXi hosts.
18405
Removing replication for a volume that is added to SRM protection groupmay fail after removing the associated test volume pairsAn attempt to remove replication for a volume that is added to SRM protectiongroupmay fail after removing the test volume pair but before removing thereplicated volume pair. This leaves the volume pair in erroneous state.
Delete the SRM protection group or remove the volume from SRM protectiongroup and retry removing volume replication.
Issues with workarounds
26UCP Release Notes
# Description
18410
PowerShell Session fail because session object reference is NULLOn a UCP instance, various jobs will start to fail includingmonitoring, disasterrecovery and attempts to gather information from VMware. These failures willall contains unexpected errors and refer to an error number. In theOrchestration log files, messages surrounding the unexpected error number willrefer to a NULL session value.
Wait for existing jobs to complete. Connect to the UCPManamgement VM. Fromthe command line run the iisreset command.
18517
New UCP privileges are not added to custom roles in vCenter after anupgradeUCP v3.0.1 has added new privileges to vCenter - UCP Network Console privilegeand UCP Reset VLAN privilege, UCP Storage Console privilege and UCP ServerConsole privilege. These privileges were included in the following privileges forUCP v2.2. UCP Network Administrator privilege, UCP Storage Administratorprivilege and UCP Server Administrator privilege respectively. If users hadcreated a custom role that uses any of the UCP v2.2 privileges they will not getthe new UCP v3.0.1 privileges added by default.
Users can go into the edit role option in vCenter and add the new privileges.
18553
The Hitachi UCP Director Servers page may display an HCSM connectivityerror while there is an ongoing server firmware update procedure.When using the vSphere thick client, the Hitachi UCP Director Servers page maydisplay an HCSM connectivity error while there is an ongoing server firmwareupdate procedure. This error message says "Sorry, an error occurred whileprocessing your request. The server inventory cannot be displayed due to a lossof connectivity with HSCM. Inventory details will be refreshedwhen HCSMconnectivity is restored". This is a temporary state that will be corrected oncethe server firmware update procedure is complete.
Use the vSphere Web Client to access the Hitachi UCP Director server inventorywhile there is an ongoing server firmware update procedure.
18721
Setting FC Privacy protocol to AES 128 not supported for 5460When setting the Priv Protocol of FC switches to AES128, the task succeedswithout errors, and the 6510 SNMP configuration is properly set, but the 5460SNMP configuration is set to Priv Protocol 3DES. This causes all refresh healthand refresh FC Performance tasks to fail.
Use UCP Director to change the Fibre Channel privacy protocol to "NoPriv" or"DES"
Issues with workarounds
27UCP Release Notes
# Description
18976
PathSelectionPolicy and SATP configuration gets set on ESXi server hostafter creating ESXi cluster using service templateESXiCluster service template requires server base host profile. When base hostprofile has both 'PSP' and 'PSP, SATP' configurations not selected, the createdESXi cluster enabled these two settings.
Manually disable these two configuration settings from the cluster host profileafter cluster creation is complete. Then re-apply the host profile to hosts in thecluster.
19210
Apply service template fails when Ethernet feature is disabledApplying service template with Ethernet-Feature disabled is NOT supported.UCP has setting to turn OFF configuring any settings on Ethernet switches ininventory. This is available in Ethernet-Switch inventory page as'EthernetFeature' button. Service template feature is required to configureEthernet switch port/s connected to host. Turning OFF or disabling'EthernetFeature' will make service template feature not function correctly.
Do not disable 'EthernetFeature' if using service templates
19213
Cannot update site pair inventoryUCP Datacenter Director site pair inventory cache cannot be updated if any UCPsite in site pair is down.
Ensure UCP Director is up and running at both sites.
19252
Test volume not created in a SRM Reprotect operationWhen a Reprotect operation is performed in SRM on a Recovery Plan, a Testvolume is not created on the Recovery site. This prevents a Test operation to berun, however it does not affect the successful outcome of a Disaster Recovery orPlannedMigration in SRM.
If a Test operation must be supported, create a new Replication Group in UCPDatacenter Director, and in SRM create a new Recovery Plan to hold thisReplication Group.
20505
vSphere web access error message when not using signed certificateWhen using the vCenter web client to access UCP from a client workstation thatdoes not already have a signed certificate installed, the operation will fail withthe following error: “Content was blocked because it was not signed by a validsecurity certificate. For more information, see “About Certificate Errors” inInternet Explorer help.”
On the client workstation, install a signed certificate. Either obtain a digitalcertificate from your trusted third party certificate authority (CA), or export aself signed certificate from UCPManagement and import it onto the clientworkstation.
Issues with workarounds
28UCP Release Notes
# Description
20721
Servers may loose I/O access to storage during firmware upgradeESXi hosts or bare metal servers may loose I/O access to storage for 14 secondswhile updating firmware on an Access Fibre Channel switch.
For every host that accesses the target Fibre Channel switch, disable the twoHBA paths that access that switch. Then update the firmware on the switch.Then re-enable the paths on the hosts. Repeat these steps for the other switch.
21685
Opening the blade console with the UCP Plug-in for SCVMM requires arefresh
When you open a console window for a server using the UCP Plug-in for SCVMM,the resulting web page shows that the page is unreachable.
Workaround: Refresh the browser window.
21756
Multiple DNS records for a Hyper-V host may get mapped to the same IP
If Hyper-V templates are redeployed to the same host, multiple DNS recordsmay get mapped to the same IP. This can happen when Active Directory (AD) orDNS zones have not been properly cleaned up.
Workaround: Clean up a Hyper-V host, including AD and DNS records byfollowing the steps below:
1.) Log onto the Hyper-V host and unformat all data volumes
2.) Using UCP Director, detach all data volumes that are attached to the server
3.) In SCVMM, remove the host from SCVMM inventory
4.) Login to the Hyper-V host as a domain admin and unjoin the computer fromthe domain
5.) Power off the server
6.) Detach and delete the boot volume
7.) Remove the server profile
8.) Check DNS and AD to ensure that all records for that object have beenremoved
Issues with no workaroundsThe table below lists existing issues that are relevant to this release forwhich no workarounds exist. The issues are listed in order by referencenumber
Issues with no workarounds
29UCP Release Notes
# Description
2571 UCP Monitoring service reports errors during Firmware update
UCP Director's Orchestrator service prevents external jobs from running if anupgrade job is currently running. This causes all Monitoring jobs to fail andfalsely report that UCP Director or the SQL database is down.
Workaround: None
8093 In the Batch detach volume UI, if a filter is applied, previouslyselected volumes are hidden, but remain selected
In the UCP Storage interface, when a batch detach volume operation isperformed, if a filter is applied to search for volumes, the previously selectedvolumes are hidden, but remain selected. As a result, when the job issubmitted, previously selected volumes are submitted, but the user is unable toconfirm if the volumes were selected or not.
Workaround: None
Previous release issues
The table below lists issues that were relevant to the 4.x releases forwhich no workarounds exist. The issues are listed in order by referencenumber.
# Description
252 Task/Event names in the VMWare web Client are not displayed as afriendly name
Task/Event names in the VMWare web Client are not displayed as a friendlyname. This has been reported to VMware support team and is not fixed in thecurrent release of VMware 6.0
992 Cannot remove IP pool range if identities are in use
An IP pool range can be addedwhen the IPs in the range are currently in use,but an IP pool range cannot be removed if any of the IPs are in use.
Issues with no workarounds
30UCP Release Notes
# Description
2813 If UCP Replication service is installed during installation butDisaster Recovery is not set up, and error is seen in the UCP logs
During UCP installation, if the option to install replication is selected, theinstaller will install the Replication Manager Service and also add a registry keyindicating DR is set up. During UCP storage operations, if the registry key isset, UCP tries to contact the DR service to check which volumes are being usedin replication and this call fails causing an error to be seen in the logs.
This should not affect UCP operations.
3018 DOC UI becomes slow and unresponsive when one of the sites isunreachable
When one of the sites registered in DOC is unreachable, the DOC UI waits for apredetermined timeout to restore communication with the offline site, causingit to become slow and unresponsive.
3938 UCP does not show user friendly job descriptions in vSphere WebClient occasionally
In certain environments, UCP jobs show names that are not user friendlywithin the vCenter jobs while being viewed from the web client. They appear inthe com.hitachi.ucp.[job].label format.
4018 Cannot move server profile across racks with different VLANs in anL3 configuration
In a UCP appliance with a Cisco 9k switches, if the 9332 ToR switches areconfigured in L3 mode with different VLANs across each rack, server profilescan be moved, but the server will fail to boot since it has the incorrect VLANset.
4038 Incorrect error message when trying to remove a server profilefrom a host already deleted from vSphere client.
When trying to remove a server profile from a host that has already beenremoved from vSphere, an incorrect error message appears saying the serverprofile remove failed instead of a warning which says that the host is not invSphere inventory.
4209 Performance aggregation jobs fail in UCP in rare conditions
Performance aggregation jobs fail in UCP in rare conditions due to an internalSQL exception
Issues with no workarounds
31UCP Release Notes
# Description
4329 Selecting multiple versions of same image package fails withadministrator error
The UCP Director's Images page does not correctly validate identical packageswith different versions when they are enabled via the API/UI.
The user should not enable two identical packages of different versions.
13886 The locale of the vCenter client used to access UCP Director should be thesame as the locale of the OS running on the UCP Management VMUCP Director can only be accessed in the language of the OS that is running onthe UCP Management VM. Currently UCP Director supports Japanese andEnglish languages. If a vCenter client tries to access UCP in a language that isdifferent than the language of the UCP VM the format of events, alarms andtasks will not be accurate.
15095 Volume detaches from a cluster can intermittently lead to false failuremessagesIf a user tries to detach multiple volumes that share the same storage pathconcurrently or immediately one after the other, the detach succeeds but itcan intermittently throw a false failure message.
15162 Same server ID with different host names causesInvalidOperationExceptionEach server host in UCP server inventory is identified with server ID. If adminchanges server ID directly on the host after it boots into ESXi it's possible toassign server ID. Doing this will create conflict in vCenter server hostinventory which in turn result in server host being not correctly available toUCP. Each server host should have unique server ID.
15540 Manual editing of HORCM files used for replication may cause replicationto break and be reported successfulUCP Disaster Recovery owns the HORCM file configurations used to setupreplication. If the HORCM files are editedmanually, UCP does not have theknowledge of known states andmay result in unpredictable behavior, such asbreaking replication and causing it to be reported as successful. Please do notmodify the HORCM files used for replication.
16013 Current image reads: "Updated" HitachiESXi image" if the image is clonedCurrent image might have the following text prepended "(Updated)" once theimage has been cloned. This is an intermittent issue at VMware’s end.
Issues with no workarounds
32UCP Release Notes
# Description
16130 Ignore system partition failure errors when deploying windows using UCPservice templateDuring windows installation a warningmessage might popup on the consolestating - “Setupwas unable to create a new system partition or locate anexisting system partition. See the Setup log files for more information.” or
The user should ignore these messages. UCP handles this error by rebootingthe blade, loading the appropriate HBA driver, and dealing with MPIO issuesand disk initialization. The user should not interfere by attempting othertroubleshooting steps.
17433 Uninstalling UCP Disaster Recovery from UCP Director only results in UCPDatacenter Director failureUCP Disaster Recovery will not work in UCP Datacenter Director if UCP DisasterRecovery is not also installed in UCP Director. Uninstalling UCP DisasterRecovery from UCP Director at a site that has been added to UCP DatacenterDirector, without fully uninstalling it from UCP Datacenter Director will resultin an error when using the disaster recovery feature of UCP DatacenterDirector. Uninstalling UCP Disaster Recovery from UCP Datacenter Director willdisable disaster recovery functionality in UCP Datacenter Director.
18035 Password for SCP server cannot contain the "@" characterFirmware update and switch configuration restore operations will not work onthe Brocade switches if SCP password has an "@" character in it. Please useother characters when setting a new password.
18331 Replication will fail if the volume size is 1.2 GBIf user tries to replicate a volume that is 1.2 GB in size replication will fail asUCP does not detect the volume capacity accurately.
18612 The Hitachi UCP Director Status Monitor page occasionally displays anAMQP communication error messageThe Hitachi UCP Director Status Monitor page may occasionally display anAMQP communication error message in the All Events table. This error eventsays "Failed to communicate with the AMQP service at IP Address: 127.0.0.1".If the UCP System is registered into a UCP Datacenter Director service, there isa small chance that the Datacenter Director service will not receive a monitorstate message from the UCP system. This will be corrected during the nextregularly scheduled UCPmonitor state refresh task.
18648 Attaching storage volume between different OS version succeedsFor example, applying volume created earlier (not currently used by any Linuxhost) for Linux OS can be attached to non-Linux hosts. This is because UCPrestricts only the currently in use storage volumes to be not shared acrossdifferent OSs.
Issues with no workarounds
33UCP Release Notes
# Description
18848 Cannot view Performance Graphs on MacOpening up the performance graphs on a Mac results in an error popup, andthe graphs do not properly display.
19486 No monitoring support for depletion thresholds in HUS storage arrayfamilyWhen depletion thresholds are exceeded on a HUS storage array, monitoringstatus shows "Normal" state. This is expected behavior since HDvM does notsupport monitoring for depletion thresholds.
20519 Identity conflict after removing server profile from serverRemove Server Profile removes the virtual identity from the server leaving theserver only with physical identity. The physical identity of the server couldhave been virtualized to another server causing a an identity conflict.
20991 Topology UI has been deprecated in 3.5 releaseGoing forward from 3.5 we will be using vC OPs for topology information andnot the topology maps that show up in the vSphere web client UI. The UIoptions will be removed in the 3.5.1 release.
22412 Re-applying server profile fails when changing slots for blades currentlyexisting in UCP server inventoryWhen a blade that currently exists in UCP server inventory is moved to adifferent slot, re-applying the server profile that is already associated to thatslot fails. This is due to duplicate MAC addresses in the database.
Getting helpHitachi Data Systems Support Portal is the destination for technicalsupport of products and solutions sold by Hitachi Data Systems. To contacttechnical support, log on to Hitachi Data Systems Support Connect forcontact information: https://support.hds.com/en_us/contact-us.html.
Hitachi Data Systems Community is a global online community for HDScustomers, partners, independent software vendors, employees, andprospects. It is the destination to get answers, discover insights, and makeconnections. Join the conversation today! Go to community.hds.com,register, and complete your profile.
Getting help
34UCP Release Notes
UCP Release Notes
Hitachi Data Systems
Corporate Headquarters2845 Lafayette StreetSanta Clara, California 95050-2627U.S.A.www.hds.com
Regional Contact Information
Americas+1 408 970 [email protected]
Europe, Middle East, and Africa+44 (0) 1753 [email protected]
Asia Pacific+852 3189 [email protected]