vSphere 4.0

I attended this vSphere 4.0 course some time ago and took down some notes in the form of questions and answers. Shall share it here. Note, don't take the short information here as final answers, i might be wrong, the instructor might be wrong.

Always verify with your own setup or official documentation.

Q8	How long is the validity of VMware authorized courses before the relevant certification exam is no longer valid?
	Till the next major version
Q9	What kind of hardware does each virtual machine emulate?
	BX440 motherboard from Intel
Q11	Can we anyhow use hardware for kernel? Can we use any available server
	It has to be certified hardware only, else not supported.
Q12	The configuration file, how different is it from VMware Server and other desktop vmware products?
	Use the free VMware convertor if you want to convert to and from different versions.
Q13	Is it better to suspend a VM or shut it down? Does it release the resources?
	This suspend is different from OS suspend. The entire state is written to disk.
Q14	What BIOS is used for the vms?
	Phoenix BIOS , stored in NVRAM file for each VM. Note that each vram is created upon VM starting.
Q15	Are all the configuration files dependant on each other?
	No. Decoupled. But if the vmdk file is lost, its all gone. Vmx file just stores the hardware configuration of the VM (Not OS).
Q16	What is DPM?
	Dynamic Power Management. Automatically power on/off physical servers accordingly to load requirements. Requires DRS clusters to be up and running.
Q18	What factors does V Motion take into account other than CPU allocation?
	Memory and CPU allocation…. Nothing else.
Q19	What is VMM?
	Virtual Machine Monitor. Supposedly do everything required of the VM (CPU/RAM and IO operations). Although certain items are just pass-thru. In other words, it presents the hardware.
Q20	Difference b/w the Hypervisor OS and Service console OS (Redhat)
	Hypervisor is the real underlying OS, service console is just like another VM that takes care of monitoring and administration of ESX host.
Q21	What is ESXI direct console?
	The screen of the physical ESXI server. To setup basic network and hardware settings. Other settings can be done via the vSphere client
Q22	Can the users of ESX management be linked to a LDAP server?
	Not for ESX, only for VC. But it is strictly limited to Microsoft Active Directory services only
Q23	Seen issues of hidden network adaptors, even from GSX times, still there? How to avoid?
	This is a Microsoft Windows issue. Cos windows see different network adaptors and thus will reassign new network adaptors. Having said that, you will need to remove all the hidden adaptors and recheck every time you clone.
Q24	How do you join an ESX to an Active Directory domain??
	You can’t. ESX only let u enter DNS information, the domain entry in the ESX is only for DNS purposes.
Q25	What is Firewall in service console?
	It only protects the connection to service console, doesn’t protect the VMs.
Q26	What is VMA
	vSphere Management Assistant. Run command lines against the ESXI servers. ESX got service console to do this.
Q28	Permissions?
	Although u can tap user info from active directory, the actual access control is set in VC only.
Q29	Why do I need to join to an AD domain for vCenter?
	Because to manage multiple ESX servers, you need a domain that would make it easier to manage due to user permissions on different machines. In other words, our infra design will need to incorporate an AD just for virtualization. Note: VMware only support Microsoft Active Directory, nothing else.
Q30	For vCenter, is it better to install the database on the machine itself? Port issues if any?
	Yes. This is especially so if we are using vCenter Server Heartbeat
Q31	What is linked mode?
	To link up different VCs, which then in turn manage their ESXs under those VCs.
Q32	Why can’t I log on as root in ESXi?
	By default, ESXi hosts does not allow root access
Q33	What is hostd?
	A very very important process that handles all execution in the ESX host. A watchdog will attempt to restart the process if it fails.
Q34	What is a VM Appliance?
	A specifically built VM that performs optimally a specific range of operations.
Q35	Where can we learn the command line commands of the vCLI?
	There are documentation on the website. But you don’t really need it most of the times.
Q36	What is VMA?
	VMware Management Application, a Linux based virtual appliance. Contains vCLI, vSphere SDK and vSphere API
Q37	What is vCLI?
	A shell program that can be installed to windows or Linux for command line control to the ESX. Note, whatever you can do in the command line, you can do it in VI Client. The only advantage of vCLI is when you want to write specific scripts to automate some processes.
Q38	What controls the MAC address of the Virtual NICs?
	They use a combination of GUID and a whole bunch of other IDs to generate ‘unique’ MAC addresses. This also means there is still a chance of MAC address conflict, though the chance is extremely low. ESX inherently cannot detect such conflicts, its up to the experience of the network guys should there be an un-resolvable network issue.
Q39	How many ports does each virtual switch support?
	Up to 4088 ports (It should be 4096 but 8 ports are reserved by VMKernel), but only 56 ports is given by default. Note, changing the maximum number of ports will require a reboot of the ESX server.
Q40	Does it make any sense or benefits if we group several VMs to a switch or port group?
	Policies like teaming and security can be set.
Q41	Can a vswitch talk to another vswitch?
	Not possible, unless they are connected via the physical NIC.
Q42	How should I connect up my management network?
	Management network should only connect to the service control port (ESX) or the VMotion Port (ESXi)
Q43	How are switches connected to physical adaptors?
	Each switch can only be assigned to unique adaptors. This is useful for physical separation. Note: 2 different virtual switches cannot communicate with each other without going through their physical adaptors. Good. If more than 1 nics are assigned to a single vswitch, they will be automatically teamed. Note again that NIC teaming and VLAN trunking needs to be supported by the physical switches.
Q44	VLANS
	They can only be configured on predefined port groups on the switch.
Q45	What’s the diff between VMKernel/VMotion and Service Console port group?
	For ESX: The management network should be connected to the Service Console port group. For ESXi: The management network should be connected to the VMKernel port group. Yes, there is a slight difference between the VMKernel port groups in both ESX and ESXi.
Q47	What is EVC as seen in ESX summary
	Cluster feature that prevents VMotion migrations failing due to incompatible CPUs. E.g. If ESX A,B,C has the same vendor type of CPU but of generation 1, 2 and 3. I can choose to set EVC (at DRS level) on generation 2 cpu. So ESX B and C can be placed in same drs cluster, ESX A will be rejected.
Q49	By default, the configured no of ports is 64.. Also, theory says default is 56….???
	This is because 8 ports are used up by the kernel, thus 56 is the usable port number.
Q50	By default, the configured no of ports for a pre created switch is 32. But theory says 24.
	Cos only 24 is usable? Since 8 are taken up by kernel?
Q51	What are the virtual mac addressed made up?
	00:50:56. It always start with this number.
Q52	Benefits of Distributed switch over Standard switches?
	The real benefits comes if u use it with Nexus 1000, a CISCO distributed switch. Note: The configuration is done on the VC. So when VC goes down, the operations on the ESX still carry on as per normal, but no more configuration is possible unless using the vCLI on each ESX. Note: You can only set no of ports on a port group of a distributed switch, but you cannot set the no of ports for the switch itself.
Q53	Uplink and Ports on Distributed switch
	Generally, there is no hard and fast rule on how many uplinks is required of how many ports. Note: I can assign uplink A to port group B. But there won’t be FT if there’s only 1 uplink assigned to a port group.
Q54	Bad thing on setting ‘reject’ for promiscuous mode.
	Up to 10% of CPU cycles would be used for this inspection of packets. But if an application tries to forge MAC Address, all egress traffic will be dropped. Same goes for ingress traffic.
Q55	How to properly use VLAN and PVLAN (Isolated, Promiscuous, Community)
	Can consider using VLAN and PVLAN for N-Tier applications
Q57	Storage?
	Transport Data Control FC FC SCSI Block/Raw iSCSI IP SCSI Block/Raw NAS IP Filesystem
Q58	Fabric?
	Collection of one or more FC channels.
Q59	What are the file systems ESX can read?
	Only VMFS and NFS. ONLY. Note: VMFS supports up to 2TB only. Note: VMFS can only be created on SCSI disks, not on any other types (e.g. IDE) Note: Only NFS version 3 is supported on ESX.
Q60	Does this means that the local hdd on esx servers are formatted in VMFS?
	YES.
Q61	Does RDM only apply to SAN? How about local SCSI raw devices or even partitions of them?
	Yes, just to SAN only. Not even NFS.
Q62	SANs?
	Same as what you know, except that there is a human LUN id and a system(SAN) LUN id Note: ESX can only support 2TB storage devices due to VMFS and the SCSI tool used, doesn’t matter if RDM is used or not (E.g. LUNs). But SAN can join different LUNs and the eventual extends can be seen as a single VMFS unit.
Q63	We experienced disconnection from NAS quite frequently. NAS is still accessible via its share but ESX simply couldn’t rescan it till we reboot it. Why? How does rescan work?
	Probably because the NAS doesn’t support NFS 3. Buffalo NAS not supported by VMWare!
Q64	Storage Runtime name/Canonical name/SCSI id?
	Runtime name: vmhbaN:C:T:L N is HAD card id. C is the channel. Each HBA only uses 1 channel. Usually zero. T is the Storage processor on the SAN L is the LUN Number ** Be careful, the runtime name will always be represented even though its not a Fibre channel storage device. ** Canonical Name: Usually naa, but if no naa, then represented as mpx SCSI id: This is provided by the ESX.
Q66	Does third party MPP , SATP and PSP always be installed with patches of VMKernel?
	No, these plug-ins usually comes at a separate charge. Have to approach the vendors separately.
Q67	Difference b/w Fibre and iSCSI SANs nomenclature
	Fiber iSCSI NFS Storage Processor Target NIL HBA Initiator (Just a modified NIC) Normal NIC WWN MAC address MAC Address Does not require vm port Requires a VMkernel Port Requires a VMkernel port
Q68	Does ESX support software/hardware initiator for iSCSI?
	Yes. But software initiator disabled by default. Also note that Software initiators take up CPU cycles considerably, as compared to hardware initiators. A native HBA iSCSI driver is provided in kernel
Q69	Did you say cannot use DNS names for a iSCSI server? Book says can.
	Can. As long as VC can resolve
Q70	Cons of iSCSI?
	The data transfer is in clear only (via the IP network of this iSCSI)
Q71	Is there a one stop info site on all the ports which vSphere is using?
	Search thru the document page of VMware site. http://pubs.vmware.com/vsp40/wwhelp/wwhimpl/js/html/wwhelp.htm#href=server_config/c_firewalls_for_configurations_with_a_virtualcenter.html
Q72	Authentication on iSCSI SAN?
	Using CHAP. By setting authentication for iSCSI on esx, you also need to set the authentication on the iSCSI SAN itself
Q73	Multipathing for iSCSI networks?
	Only Active-Passive multipathing. No load balancing.
Q74	VMFS?
	Meant for large size storage, DO NOT save documents on this vmfs!!
Q75	Can I use only a portion of a LUN to create my datastore?
	Datastore can be created to a partial sized LUN. The remaining space of the LUN can be used for creating new datastores or ‘growing’ the datastore on this LUN. But note that the datastore on each LUN is 2TB maximum. After that, you should extent other LUNs to this LUN to form bigger datastore. Note: The datastore (VMFS) can also grow dynamically (automatically). Note: So long as the ESX can see the LUNs, they can be ‘extended’ together. But remember, each LUN size should be 2TB max. Some advice: Create a LUN just large enough so that the total usability is about 60% of the total space.
Q77	I notice that sometimes creation of new VMs may result in weird vmdk naming? How does vSphere handle naming of vmdk file?
	If the VM has > 1 disk file, the file pair for the second disk file and on is named vm_#.vmdk and vm_#-flat.vmdk, where # is the next number in sequence , starting with 1.
Q78	Why is there a need to specify Guest OS when installing?
	Because the VMkernel talks to this VM (The VMM) based on this parameters. Thus it is highly recommended that the user sets the correct OS for each VM. If not, there might be problems later, though it does not show immediately.
Q79	Why can’t we put spaces in VM names?
	Because scripts won’t work with this.
Q80	Difference between vm.vmdk and vm-flat.vmdk?
	The flat file is actually the main file that stores the data, the vmdk (no flat) is just a meta file that describes the disk’s parameters (E.g. cylinders, rpm…etc)
Q81	What is .vswp?
	VM swap file. Meant for the kernel to present a bit of the hard disk to the guest OS as physical ram , as a last resort. Note that this is different from the swap file used in windows OS.
Q82	How many vmdk files can a virtual machine take?
	Up to 60. Because each VM supports 4 SCSI adaptors, which in turn supports 15 hdds. 15x4=60
Q83	How do we do clustering of VMs?
	Use RDM. All other forms are not supported by VMware officially.
Q84	What is .vmxf?
	Used by the kernel only. An extended configuration file. DO NOT touch this file manually.
Q85	What is thin and thick disks?
	Thin is grow on demand disk, thick is pre-allocated disk
Q86	What are the considerations for hardware/software as compared to physical? E.g. Do you need a 1 to 1 comparison?
	Software for a fact, needs to take advantage of this. For MS os, generally more then 4 CPU won’t do much significant improvement. (Don’t quote Bert…lol) Hardware, of course needs to support.
Q87	Virtual disk mode setting?
	Non independent – For snapshots Independent (Persistant) – Changes written to disks is immediate and permanent. Independent (non persistant) – After reboot, the changes written to disks will be lost. Good for clean slate operations.
Q88	Virtual NIC?
	E1000 and vmxnet3 are more for newer OS like windows server 2008 Use flexible for most cases.
Q89	DUMB Question. Can I install ESX or ESXi on a VM???
	Yes, but only specific versions of ESX and VMware are possible. Google it.
Q90	What are appliances?
	They are ‘Just Enough OS’ (JEOS) which are highly optimized for the application they are meant for. E.g. Weblogic
Q100	How to make sure every Windows VM deployed from template is ‘resetted’?
	Can install Microsoft Sysprep on the VCenter, then vSphere will handle all those for us. Note: the sysprep cab file needs to be extracted to c:\doc and settings\all users\application data\vmware\vmware virtual center\sysprep\svr2003 Note: Sysprep in vSphere also requires the installer sometimes. So its always a good idea to make an ISO out of the official Windows Setup and keep it in a dedicated software spot.
Q101	Converting from physical to virtual, windows reactivation issue?
	Have to ask Microsoft.
Q102	Does conversion work for old OS (Windows NT?)
	Yes. In fact, the older the better.
Q102	Does conversion changes MAC addresses?
	Yes, the physical MAC addresses cannot be reused in virtual machines. Yes, software licenses tied to mac addresses will need to be redone.
Q103	Will all other network and computer information be saved, in the virtual machine?
	Yes, but there’s the option of changing it.
Q104	Support for clustered machine conversion?
	No. Have to de-cluster, convert and reform cluster.
Q105	Support for USB cold conversion?
	No.
Q106	Do you actually get to map each volume to another volume in the ESX?
	Yes.
Q107	Whats a VMCI device?
	For VM comms. Now not being used.
Q108	Hot extend of Virtual Disks reversible?
	No, unless you use convertor to convert into another VM.
Q109	Changing of virtual machine names after they are created?
	Not recommended, usually not all the dependencies are changed as well.
Q110	What is para-virtualisation or VMI?
	Allows you to create a Guest OS to send commands that are virtualized ready. This increases the VM’s performance. Note: Currently MS Windows is not ready for paravirtualisation.
Q111	Snapshots?
	Snapshots contain new updates from the last image it took. So its actually not a clone from the previous. Note that a snapshot is always dependant on the previous image/snapshot from where it’s taken. Note that when a snapshot is deleted (Using VMWARE’s command!!), the changes are fused into the ‘child’ snapshot. (Note that vmware calls it the parent instead). Be careful! If you try to delete snapshots on your own, be prepared to lose the child snapshots.
Q112	Is there a clean up function to find all orphaned vmx/vmdks?
	Scripting from google
Q113	How stringent are CPU compatibility requirements for VMotion and stuff?
	See pg 560 – 562 of course book 2.
Q114	Can role privileges be assigned down to the Datacenter/ESX/Folder/VM level?
	Yes. But if I vmotioned my VMs from ESX A to B, then the permissions will be renewed and inherited according to the new ESX.
Q115	Can users/groups be distinguished between domains? E.g. alpha\user1, beta\user1.
	Not really, but VMware’s direction is that we should bring in users from other domains into a localized domain.
Q116	Permissions.
	Specific permission defined on a specific object overrides all inherited permissions, union is out.
Q117	I suppose the permissions are stored on VC. So permission setting is not possible on independent ESXs?
	Not really. But the users will be that of the esx SAM. And you cannot administer the roles created on the vc and vice-versa
Q118	Is there a generation of roles and permissions report?
	Sadly no. And don’t know when there will be one.
Q119	CPU presentation
	For a Dual Core physical CPU, vmkernel presents it as 2 logical CPUs. For a Quad Core physical CPU, vmkernel presents it as 4 logical CPUs. The VMs can use any of the logical CPUs presented (Maximum is 8). If a physical CPU is Hyper Threading enabled, then the no of logical CPUs will be doubled. Note: The HyperThreading function can only be turned off at the server BIOS level. (Also known as Enable Logical Processors) Note that if a VM has multiple CPUs, the vmkernel must be able to assign that number of logical CPUs to that VM before that VM can continue executing.
Q121	Planning of VMs?
	In general, there is a best practice that each VM should be allocated to just 1 logical CPU. This will allow the vmkernel a greater chance of allocating CPU cycles to the VMs, as compared to VMs with multiple cores (cos need to make sure that number of logical cores is available). So saying so, each VM should be configured such that it only performs one type of operation. (E.g. App server, DB Server…etc)
Q122	If I assign 1GB to my VM, and as time goes, the physical ram is not enough to support this requirement. Will the VM crash?
	No. Cos with the starting of each VM, a VM swap file will be created. The size of this swap file will be the difference between the max ram assigned to the VM and available physical ram. If this swap file cannot be created on starting, the VM will not start at all. Also, vSphere has other ways of getting memory. Transparent Memory paging -> Balloon driver (Only works for inactive pages) -> VMkernel Swap (last resort)
Q123	System monitoring tools
	IOMeter. Perfmon – note that after you install vmwaretools, then you can get additional stats such as the actually physical ram usage as compared to your vm usage. Iostat – to check ios. AppSpeed – A vmware appliannce
Q124	Ability to monitor applications and services on the vms?
	In the AppSpeed appliance. But this is still vague as a lot of features are currently only on the roadmap.
Q125	How to change the scale and refresh rates of the performance charts?
	Nope u cannot. These options are locked in.
Q126	Can i print a performance report?
	Yes.
Q127	If a reservation is a guaranteed value, does this means the total reserved value of all the VMs on a single ESX must be less than the total physical memory on the ESX?
	Yes. And this means that if you try to migrate to an ESX that cannot satisfy the reservation requirements, the system will not allow you to do so. Only way is to reduce the reservation requirements.
Q128	Reservation on a multi-cpu VM?
	The reservation is divided among the number of CPUs on that VM.
Q129	Resource pools?
	It’s a logical abstraction of resources, that indicate the kind of requirements (cpu, i/o, ram) Can set reservation and limit as per normal, loads can span across esxes. You can create hierarchies of resources also. Just be careful that the summation of sources of a child resource pool must be enough to cover this.
Q130	Thin disks: Will the disks fail when the actual storage runs out.
	Yes. So u will have to monitor and increase it as time goes.
Q131	Thin disks cannot defrag?
	Yes. But VMware is coming up with something.
Q132	Are there performance issues for using thin disks?
	Yes. Cos thin disks would also mean more meta updates.
Q134	What happens to roles after VCcenter is isolated?
	The original roles are maintained, newly merged roles discarded.
Q135	How to see a server (esx) cpu specifications as required by vSphere?
	Use the VMware bootable CPUID utility at http://vmware.com/download/shared_utilities.html
Q136	Does the system give u specific error when vmotion fail?
	Sadly, probably not. Can try looking into the logs instead.
Q137	Can I enable/disable vmotion at the VM level? Fix an ESX for a VM?
	Yes. But u have to fix this at the DRS level where you can set it for each VM in the DRS(ESX that’s inside that DRS)
Q138	Maintenance Mode?
	To put esx in a state for management. Virtual machines continue to run till u power off or migrate them out. No VMs can be powered on henceforth.
Q139	With VMware HA, does it mean I don’t need Active-Passive Clustering on my guests?
	Not true. Cos MSCS is FT, and check into service level. Currently, VMware HA by itself can’t do it. We really need to think if we want to do MSCS. Without MSCS, we must accept non FT situations and lack of application level monitoring. Warning: Cold migration might not even be supported for MSCS VMs.
Q140	Is there a way to prevent other VC from ‘stealing’ my ESX?
	No software methods to do that.
Q141	VMware cCenter Server Heartbeat?
	Clustering solution to protect vCenter only
Q142	Does vCenter Server Heartbeat comes with vSphere package? If not, can use MSCS?
	Actually yes, but no answers to the pros and cons
Q143	Vmware HA slot size availability?
	Slot size – Will be activated if under ‘Configuring VMware HA setting’ -> Admission control -> Prevent VMs from being powered on if they violate availability constraints. Slot size are calculate in following manner - If I had 5 VMs with ram reserved value of 1,2,3,4,5 GB respectively. - vSphere will take the value of the largest. 5GB. - So if the esx can support up to 10GB ram, then 10/5=4 slots are available. - This means the esx can only handle up to 4 VMs running on it, this will be the factor to decide if a VM can be powered on. - Generally, slot computation gives a feel of how much host failures it can handle. (Need to factor in the computation of the VMs in the different ESX) - Note that the slot size can be manually defined though in advanced settings. (See page 645 of course book 2)
Q144	HA with DRS?
	If host fails, HA kicks in first (Even if it means breaking DRS rules), then DRS will come in and reschedule the vms on different ESXs.
Q145	Does HA detect OS failures like Blue screen? Can VMware HA be enhanced to support application level monitoring? Are there partner products doing this?
	VM HA can monitor VMs only if a vmware tools is installed. VM HA will monitor the heartbeat from the VM. Although this can have a side effect that the heartbeat service does not get cpu cycles as another app is hogging the resources, VM HA will think the VM failed and restarts the whole VM in the midst of a busy application running. [This info is confidential] - For application level monitoring, can wait for appSpeed (1/2 a year), or in a quarter or 2, for the next version.
Q146	HA. What happens if a clustered ESX is disconnected from heartbeat network?
	The ESX goes into ISOLATED mode and will be considered as a failure. The VMs on that ESX will be powered off and then powered on in other ESX servers in the same cluster.
Q147	For HA, when the situation in Q146 happens, does the booting of the shutdown’ed VMs happen immediately when the VM is shutdown or until all the VMs in that failed ESX is shutdown?
	No idea. You have to test it out to find out.
Q148	For HA, if my heartbeat network is unstable, but all the ESX are running ok. Will HA kick in accidentally?
	Possible, but u can set a value to determine the time outs for HA to happen. By default is 11secs, you can increase it.
Q149	Can I control which VM boot up first?
	Yes, but indirectly, using priority bootup. This is configured at the cluster level. Not the esx level.
Q150	In a Vmware HA cluster, how do the ESX communicate?
	First 4 ESX added is primary, the rest all are secondary. Secondary talk to Primary, Primary talk to other primary. Warning: If primary ESX fail, the cluster will fail for a short while. In the failure moments, re-election of primary esx will occur and then the cluster should recover again. This primary/secondary thing is not configurable. The information is not shown also.
Q151	What are the uses of VMkernel network?
	VMotion, IP Storage and Mgmt in the case of ESXi
Q152	Pinging of isolation address?
	To determine if the ESX itself is somehow network isolated. This pinging is 1 sec interval, in addition to the heartbeat testing.
Q153	Is FT supported on a wide range?
	No, only supported on newest ESX hardware
Q154	Why or why not Vmware FT ?
	Cos it’s a possible replacement for MSCS? Actually no. But there’s a whole lot of prerequisites (limitations) for FT. (See pg 678 on course book 2)
Q155	vCenter Heart Beat server - Does the creation of the active server automatically create the passive server via network?
	Yes, this is possible. You can also choose to clone it. Note: For heartbeat to protect, the SQL server needs to be installed on the VC itself. The interesting thing is there is no need for shared storage.
Q156	What are the implications of merging different port groups (vmkernel, vmchannel…etc) on the same network?
	Its ok as long as u separate the production and management and protect them accordingly.
Q157	Why is there a hardware requirement for domain controller for Server heartbeat?
	Not a hardware requirement, but it can be any accessible GC domain controller.
Q158	Can I install Server Heartbeat on VM?
	Yes. Combinations can include VM-VM, VM-Physical, Physical-VM
Q159	Can Heart beat protect the Oracle instance installed with VC?
	This may not work on Oracle, at least not now
Q161	Backup ESX?
	It is far easier to reinstall an ESX than to restore it from a backup. But you can backup ESX configuration and proc info instead.
Q162	VCB and vStorage APIs?
	VCB is old technology. All vendors has promised to move on to vStorage APIs. Need to check with vendors if they will offer upgrade options when this kicked into force is a few quarters. If using integration scripts, then its not using the latest technology.
Q163	Host profiles?
	Some kind of configuration that is created and can be applied to all other ESX hosts if you want them to comply to a standard configuration. This can also be used to check for incompliance when adding a new ESX server.
Q164	Update Manager
	To be able to patch the ESX and the virtual machines. You can put a Patch DB on the internet network and let it download EVERYTHING. Then copy them out on DVD or USB. Then transfer all these patches to the actual patch DB in the network. Note that you can install the update manager on the VC as well. For VM level, if the patch breaks anything, you can actually rollback. This is done by setting a snapshot taking before patching and store that snapshot for a configured amount of time. For ESX level, having DRS enabled will allow a zero downtime patching. Note: Doesn’t work for MSCS clusters since it doesn’t support VMotion.
Q165	For ESXi embedded, patching supported by VMware?
	No. For such hosts, the patching has to be provided by the vendors only.