Your Daily VMware quiz!

kj0

Alright, Answer time.

Whilst the VM is sitting on the host, if you have deselect "Sync Guest OS with host" VMTools option, the time will sync with where you have pointed the guest OS for NTP server.
When you vMotion or perform a series of events, the guest will sync with the host by default - this can cause issues if the host time is out of Sync.

To prevent the guest OS ever syncing with the Host time, you need to add additional lines to the VMX of the guest VM. VMware KB: Disabling Time Synchronization

Deathmage

kj0 wrote: »

Alright, Answer time.

Whilst the VM is sitting on the host, if you have deselect "Sync Guest OS with host" VMTools option, the time will sync with where you have pointed the guest OS for NTP server.
When you vMotion or perform a series of events, the guest will sync with the host by default - this can cause issues if the host time is out of Sync.

To prevent the guest OS ever syncing with the Host time, you need to add additional lines to the VMX of the guest VM. VMware KB: Disabling Time Synchronization

adding this too the toolbox. I would have never thought of that!

DevilWAH

Question. (hopping the answer might help others in the progress of upgrading)

you update esxi5.5 to version 6. upon reboot you notice that the hosts not longer have there physical nics attached to the distributed switch and attempting to re-add them end in an error message.

What setting might be causing this, and haw can you resolve it?

Essendon

Unaware of a setting, but are they on the HCL? Gotta look at the error message first!

DevilWAH

Its not specifically to do with HCL, its a network setting that is global per host. The error you will see when trying to add a physical nic from an affected host will contain the line. " Unsupported address family".

this kind of gives it away if you want to google it so the KB is here describing the issue and fix..

But any guess before you look? I really posted this as I know people will be looking to upgrade and this is a good one to check before you do to prevent issues with the process.

phack

Data Virtual Volume Size MAX = 62TB

https://www.vmware.com/pdf/vsphere6/r60/vsphere-60-configuration-maximums.pdf

piedthepiper

Hello guys I am new here and this is an epic thread!

Can anyone join in with questions?

You have a host that no one knows the root password for, what method can you use to recover that are/are not covered by vmware?

Essendon

Let's revive this thread, let's inject some life again!

@piedpiper - anyone can chime in with questions and answers. Just make the question's font red so it stands out.

Questions:



The Visio's mostly obvious. The RAM numbers are reservations at all levels of the hierarchy. The cluster has 24GB RAM to dish out.

Consider all VMs are currently powered off. The power-on events will go like this, VM1 > VM2 > VM3.... > VM8

1. You power on VM1, it boots up no worries. Will VM2 boot up too? Why?

2. Say VM2 was able to power on, you then power on VM3, will VM4 then be able to power up too?

3. Will you be able to power up VM5 and VM6? Why?

4. Will you be able to power up both VM7 and VM8?

5. What can you do to ensure all VMs power up?

6. Does this make you think resource pools arent a good idea? If so, why?

iBrokeIT

Wow! That's a great resource pool question Essendon, I hope you guys are tossing him some +Rep

I really dislike resource pools because most people don't really understand them and misuse them as folders to organize their VMs then have performance issues with their VMs during contention or during a host reboot.

Essendon

Thanks mate! Gotta spread the knowledge. If I can assist one person in getting their VCP and hopefully their VCAP/VCIX, this thread and my time has been well worth it.

dave330i

Guess I take a crack at the RP questions. Haven't worked with RP in a while so I'm rusty. I'll keep my answers in white.

1. VM2 will not boot since the RP doesn't have enough resources, the parent pool doesn't have enough resources and the parent isn't expandable.

2. VM4 won't boot since the RP doesn't have enough resources and it is not expandable.

3. This one is tricky. VM5 will power on, while VM6 won't due to overhead resource requirement and the RP isn't expandable.

4. VM7 & 8 should power on since the RP and parent RP are expandable and there are enough resources within the cluster.

5. You could increase the RP reservations, enable expandable reservations on all RPs, **** reservations and go with shares.

6. Nested RPs are a pain. If you use share vs. Reservations and size them correctly, they can be useful.

Essendon

Done well Dave.

Everyone else - review the question and answer 10 times if you need to. You must understand RPs for your VCPs and VCAPs..

Essendon

Question:

You are the administrator of an environment that is located across 2 physical datacenters. Datacenter A is your production site and Datacenter B is the DR site. The vSphere version at both sites is at 5.5 and SSO operates in basic mode (vCenter and SSO on same VM). You need to upgrade to 6.0 and there's a new VDI project that is shaping up that's going to need a dedicated vCenter. How would you do this?

1. Upgrade vCenter and SSO to 6.0 embedded PSC and point the 2nd vCenter to the PSC. Upgrade VUM, use it to upgrade hosts to 6.0 and move VMs to hardware version 11
2. Deploy standalone instance of SSO 5.5, repoint vCenter 5.5 to new SSO 5.5, decommission old SSO, upgrade SSO 5.5 to 6.0, upgrade VUM, upgrade vCenter, upgrade hosts and move VMs to hardware version 11
3. Upgrade vCenter and SSO to 6.0 embedded PSC and point the 2nd vCenter to the PSC. Upgrade VUM, use it to upgrade hosts to 6.0 and move VMs to hardware version 12
4. Deploy standalone instance of SSO 5.5, repoint vCenter 5.5 to new SSO 5.5, decommission old SSO, upgrade SSO 5.5 to 6.0, upgrade vCenter, upgrade VUM, upgrade hosts and move VMs to hardware version 11

Choose one option.

Deathmage

Wish I could use Resource Pools here, I love them at home. I have to run in EVC mode, so no pools for me.

Essendon wrote: »

Thanks mate! Gotta spread the knowledge. If I can assist one person in getting their VCP and hopefully their VCAP/VCIX, this thread and my time has been well worth it.

Cough, you already have with me.

Deathmage

Essendon wrote: »

Question:

You are the administrator of an environment that is located across 2 physical datacenters. Datacenter A is your production site and Datacenter B is the DR site. The vSphere version at both sites is at 5.5 and SSO operates in basic mode (vCenter and SSO on same VM). You need to upgrade to 6.0 and there's a new VDI project that is shaping up that's going to need a dedicated vCenter. How would you do this?

1. Upgrade vCenter and SSO to 6.0 embedded PSC and point the 2nd vCenter to the PSC. Upgrade VUM, use it to upgrade hosts to 6.0 and move VMs to hardware version 11
2. Deploy standalone instance of SSO 5.5, repoint vCenter 5.5 to new SSO 5.5, decommission old SSO, upgrade SSO 5.5 to 6.0, upgrade VUM, upgrade vCenter, upgrade hosts and move VMs to hardware version 11
3. Upgrade vCenter and SSO to 6.0 embedded PSC and point the 2nd vCenter to the PSC. Upgrade VUM, use it to upgrade hosts to 6.0 and move VMs to hardware version 12
4. Deploy standalone instance of SSO 5.5, repoint vCenter 5.5 to new SSO 5.5, decommission old SSO, upgrade SSO 5.5 to 6.0, upgrade vCenter, upgrade VUM, upgrade hosts and move VMs to hardware version 11

Choose one option.

Too me I'm not as smart as some of you, yet, in regards to VMware but the question stipulated a new dedicated 2nd vCenter so to me that's answer 2 and 4 right off the list right away. answers 1 and 3 are tricky if you don't know hardware versions, version 11 as far as I recall is for ESXi 6.0, version 10 was the highest for ESXI 5.5 so that's only logically. I'm pretty darn sure hardware version 12 is for things like vFusion, vWorkstation player, workstations.

I know the 1st step in upgrading you always upgrade vCenter 1st so this was a dead giveaway on two question wrong right away, so with the above mentioned and just by looking at how similar the answers are, I'm basing my answer on hardware versions, so I'm going with answer 1.

iBrokeIT

Deathmage wrote: »

I have to run in EVC mode, so no pools for me.

What?....

dave330i

Not VCP related.

Question:

If you install every VMware product, how many ports will it use?

Answer in white:

735 unique ports.

jibbajabba

iBrokeIT wrote: »

What?....

Sorry Trevor, don't get it either. Because you got EVC enabled you can't use RPs ?!?

That would be pretty bad if that's the case

Deathmage

jibbajabba wrote: »

Sorry Trevor, don't get it either. Because you got EVC enabled you can't use RPs ?!?

That would be pretty bad if that's the case

Yup, can't do a Resource Pool.

here are some screenies, they wanted to reuse last generation Dell servers so EVC mode it is, and EVC mode has limitation.





I do remember vividly the EVC mode drove me nutso because when I 1st deployed the cluster I figured the were the same generation and I had to destroy the cluster, move it from VMware free on one host, recreate the cluster, enable EVC mode, add the new host to the cluster and then migrate all of the VM's in a balanced stance..... it was utterly annoying....

But defiantly something I learned how to do... I only did it once in the lab and never had it on the VCP test so it was allot of bricks sweating out of my head.

dave330i

It's not EVC. It's DRS: http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1004098

Deathmage

dave330i wrote: »

It's not EVC. It's DRS: VMware KB: Cannot create a Resource Pool in a vCenter Server cluster

It says something like incompatible CPU or something (I'll go recreate it and upload the picture), drove me nutso when I 1st deployed it, wouldn't let me enable DRS with EVC for one of the hosts. Even submitted a ticket to VMware, but there cluster is so small here I just left it be.

It was a few months ago maybe they did a bugfix or hotfix, might do a VUM update this weekend when I update the Equalogic, just been doing project after project, like last week I did a MPLS VPN to a remote location 1600 miles away, so been dealing with WAN-based Terminal Services configurations and tweaking!!!!!!!!

Maybe next year when we need vCloud Director for our SaaS I'll get Enterprise Plus and it will go away when I upgrade to 6.0. Right now we have Enterprise 5.5.

VCP6-DCV is next after the CCNA: R&S, already upgrade the home-cluster to 6.0 but I just want to test the piss out of it at home before I deploy it at work....then I'll do the MCSA after that....

I know I'm loopy I'm a system administrator and I'm doing Microsoft last, go figure.

iBrokeIT

Deathmage wrote: »

...wouldn't let me enable DRS with EVC for one of the hosts.

Sounds like your EVC cluster isn't properly configured.

Personally, when I'm building a cluster that will use EVC I start with the lowest common host that will be in the cluster, create the cluster with EVC mode set to the level of that host and then add the rest at the same level.

If you have VMs already running at a higher processor family level then they will need to be restarted to downgrade to the lower level (which might have been your issue).

Deathmage

iBrokeIT wrote: »

Sounds like your EVC cluster isn't properly configured.

Personally, when I'm building a cluster that will use EVC I start with the lowest common host that will be in the cluster, create the cluster with EVC mode set to the level of that host and then add the rest at the same level.

If you have VMs already running at a higher processor family level then they will need to be restarted to downgrade to the lower level (which might have been your issue).

ya it's a bug sadly I wish I did misconfigure something, I had the VMware guy remote into my PC back in May and they said they send me a hotfix, they never did. I also haven't dont VUM in a few months either.





..

Deathmage

dave330i wrote: »

Not VCP related.

Question:

If you install every VMware product, how many ports will it use?

rofl, insert wise ass remark, however many you give it!!!! I block the piss out of mine....

But not sure how I remember this, but I had notes for VCP5 on this, looked in my notes but its 148, as of ESXi 5.5.


my screen isn't big enough for this picture I have in my VCP5 notes folder back from Jan.





I figured knowing the ports and having this picture would be nice to blow up some day once I get to VCAP level considering it's probably highly unlikely VMware will chane the bulk of them anytime soon......

jibbajabba

Deathmage wrote: »

ya it's a bug sadly I wish I did misconfigure something, I had the VMware guy remote into my PC back in May and they said they send me a hotfix, they never did. I also haven't dont VUM in a few months either.

iBrokeIT wrote: »

Sounds like your EVC cluster isn't properly configured.

Personally, when I'm building a cluster that will use EVC I start with the lowest common host that will be in the cluster, create the cluster with EVC mode set to the level of that host and then add the rest at the same level.

If you have VMs already running at a higher processor family level then they will need to be restarted to downgrade to the lower level (which might have been your issue).

Then I can only imagine EVC has been enabled after VMs were powered on so VMs don't get the new instruction sets for the CPUs and as a result some VMs wouldn't be able to vmotion from one host to another.

Having said that - I had that before and all I got was an error that the VM couldn't be migrated live and I had to power them off ...

So you got different CPUs in your cluster or why EVC ?

Deathmage

jibbajabba wrote: »

So you got different CPUs in your cluster or why EVC ?

They initially had no idea the concept of VMware or what the 'cloud' was....

the VMware cluster was proof on concept and I did it for them on the basis they'd add more hardware later, so I did it with an existing server, a 2009 era R710, and one brand-new Dell R720xd. I knew going into it they would have a generation mismatch just didn't think it would cause this kind of problem. the next N+2 server is another R720xd.

Basically they wanted the 'cloud' for less, so it was a proof on concept really and now they love the stability and performance, I also sacrificed on the hosts to get a next-level higher Equalogic SAN. So yes I know EVC isn't the best option but I didn't want to gimp the core storage fabric and the 10G iSCSI has a big cost factor too.

So that end I do think I made the right choice all things considering, just moving forward if they want more control over resources they will need to invest in newer hosts, but at-least the core storage fabric was built correctly from the get-go and can addon a few more Equalogic's if need be without a performance issue..

Lexluethar

Essendon wrote: »

Let's revive this thread, let's inject some life again!

@piedpiper - anyone can chime in with questions and answers. Just make the question's font red so it stands out.

Questions:



The Visio's mostly obvious. The RAM numbers are reservations at all levels of the hierarchy. The cluster has 24GB RAM to dish out.

Consider all VMs are currently powered off. The power-on events will go like this, VM1 > VM2 > VM3.... > VM8

1. You power on VM1, it boots up no worries. Will VM2 boot up too? Why?

2. Say VM2 was able to power on, you then power on VM3, will VM4 then be able to power up too?

3. Will you be able to power up VM5 and VM6? Why?

4. Will you be able to power up both VM7 and VM8?

5. What can you do to ensure all VMs power up?

6. Does this make you think resource pools arent a good idea? If so, why?

1. VM2 will boot up too. The pool is allowed to expand beyond the 4 GB allocated, so having a total of 5 is okay.

2. VM4 won't be able to power on because there won't be enough resources in the pool, the RP only has 4GB non expandable and VM4 is requiring 1 GB too much. Also VM4 is over committing the parent RP.

3. VM5 and VM6 will power on, both have enough resources under that RP to power on even though it's not expandable.

4. Yes both VM7 and VM8 will power on, while this exhausts the current RP it's expandable and as long as there is enough RAM to lend out (24 GB) you are okay (22 GB used and the cluster has 24 GB - which it would still allow you to turn on the VM's right? Just once you start allowing more memory to be consumed things like ballooning, transparent page sharing, compression and hot swapping start to play a bigger factor - right?)

5. I would allow every RP to be expandable as long as you aren't over committing the Hosts in the cluster

6. I think in specific instances RP are a great tool. If you allow departments to manage their own virtual infrastructure RP is a great policing policy. You can easily create RP for departments and configure memory allocations per pool. This way departments can manage the resource pool given to them. In smaller environments or environments. This can be a great 'charge back' policy too where each resource pool is charged based off consumption. We don't use it because we police these things ourselves, but i can see in very dispersed environments this type of setup could be beneficial (thinking big ass companies like Dell, HP, ect).

I'm not 100% sure on some of these but those are my best answers.

Essendon

Deathmage wrote: »

Too me I'm not as smart as some of you, yet, in regards to VMware but the question stipulated a new dedicated 2nd vCenter so to me that's answer 2 and 4 right off the list right away. answers 1 and 3 are tricky if you don't know hardware versions, version 11 as far as I recall is for ESXi 6.0, version 10 was the highest for ESXI 5.5 so that's only logically. I'm pretty darn sure hardware version 12 is for things like vFusion, vWorkstation player, workstations.

I know the 1st step in upgrading you always upgrade vCenter 1st so this was a dead giveaway on two question wrong right away, so with the above mentioned and just by looking at how similar the answers are, I'm basing my answer on hardware versions, so I'm going with answer 1.

Yeah this question is one of those in which you need to have played with vSphere 6 and/or read the documentation. In one line, VMware don't recommend and soon won't support an embedded PSC talking to another vCenter Server.

1. This clearly wont cut it as pointed out just above
2. You need to upgrade vCenter first, then VUM. (notice the subtle change in order in the answer choice)
3. Again won't work for the reason mentioned above
4. This works

Answer is 4.

Essendon

Sorry for the delay in updating this thread, I've been flat out with work.

Question:

Let's change the format for this question to a rapid fire one (answers in white)

1. What's the name of the vSphere component that can serve as the public key infrastructure for your vSphere 6.0 deployment? VMCA

2. In vSphere 6.0, VMware HA requires a robust DNS infrastructure? No, it hasn't depended on DNS since 5.0

3. In vSphere 6.0, the names of the lockdown mode options are? Normal, strict and exception users

4. In vSphere 6.0, Single Sign On is now called? Platform Services Controller

5. In vSphere 6.0, host certificates are stored where? On the hosts

Deathmage

Essendon wrote: »

Sorry for the delay in updating this thread, I've been flat out with work.

Question:

Let's change the format for this question to a rapid fire one (answers in white)

1. What's the name of the vSphere component that can serve as the public key infrastructure for your vSphere 6.0 deployment? VMCA

2. In vSphere 6.0, VMware HA requires a robust DNS infrastructure? No, it hasn't depended on DNS since 5.0

3. In vSphere 6.0, the names of the lockdown mode options are? Normal, strict and exception users

4. In vSphere 6.0, Single Sign On is now called? Platform Services Controller

5. In vSphere 6.0, host certificates are stored where? On the hosts

Looks like I do indeed need to read the ESXi 6.0 book, a ton has changed now.