Your Daily VMware quiz!

Lexluethar

Can i use a life line? I'll wait for the next round of questions to answer, i'm clueless on 6.0

Essendon

@Trev - Yep a fair bit's changed mate!

@Lex - Here ya go

Sticking to the same format

Question:

1. You are trying to vMotion a machine from one cluster to another but the migration wizard throws an error about networking not being available. What is the issue and how can you fix it?

2. In vSphere 6.0, the default isolation response for VMs is?

3. The default security settings for a standard switch are?

4. Can an ESXi 5.5 host run with a 6.0 vDS?

5. What's the name of the agent that manages HA on ESXi host?

Essendon

@Lex:


Let's go through your answers to the RP question.


1. VM2 wont boot up because although its RP is allowed to expand and ask the parent RP for more grunt, the parent RP doesnt have any more to go around. It only has 8GB available which has been evenly split between the child RPs, also note it cant ask the cluster for more even though the cluster may be be able to dish some RAM out.


2. Yep, you are right here. VM4 isnt overcommitting the RP, it wont be allowed to bootup so it cant overcommit. If a VM isnt powered on, it isnt consuming resources (barely any anyway).


3. VM5 will boot up or VM6 will boot up, but not both. Thing is when a VM's powered on, it also has the idea of "memory overhead". Since the memory it's been provsioned with + the overhead tips it over 2 GB, it doesnt leave enough for the other machine to power on.


4. Yeah, you are right here. Ballooning > compression > swapping can happen. Note TPS is now disabled after VMware came out with a security advisory about TPS being able to be exploited. In addition, TPS doesnt provide any savings because it doesnt work for large memory pages (which is what most modern guest OSs use - Windows 2008 etc). Why it's not used - because the chance of finding matches among large pages is very low incurring an avoidable tax. You'd think oh why cant ESXi break down the large pages into smaller ones and find matches? Yeah well it could, but it'd expend more than it would gain so it doesnt do it.


5. Yep, you could also increase the reservations, or better still not do reservations at all and go with shares only. Just know shares only kick in during times of contention only. With overcommitment, isnt this what virtualization is about?! It's how much and your goals that matter.


6. RPs are a pain in general, IMO. Nested RPs are worse. Dont do them. You are right about allowing departments to manage their RPs, well it sounds great and to some extent it is. But what would you let them do? Take a step back and think of the larger picture here. As for charge back - yep, that's a possible use for them. But given the management overhead, I'd employ a PowerShell script to **** out resource usage and charge off that instead. Big companies would almost never use RPs, contrary to what you'd think. They'd have disparate environments managed by different teams or maybe a vCD environment which splits resources up.


Hope this helps.

Lexluethar

Essendon wrote: »

@Trev - Yep a fair bit's changed mate!

@Lex - Here ya go

Sticking to the same format

Question:

1. You are trying to vMotion a machine from one cluster to another but the migration wizard throws an error about networking not being available. What is the issue and how can you fix it?

2. In vSphere 6.0, the default isolation response for VMs is?

3. The default security settings for a standard switch are?

4. Can an ESXi 5.5 host run with a 6.0 vDS?

5. What's the name of the agent that manages HA on ESXi host?

Awesome thank you for explaining those to me man - i appreciate it.

Here is my crack at your new questions:

1. Networking error can be a few different things. Off the top of my head it could be:
>You are using a standard switch and the networks you've configured aren't labeled exactly identical.
>You are using a DVS on the source Host and a standard switch on the destination - this will throw an error about the network not being available
>Your vMotion network is not configured correctly on one of the Hosts - if Host A cannot talk with host B you won't be able to vMotion. Easy test is to use esxi ping and verify connectivity from one Host to another (side note you shouldn't be able to test from your computer because the vmotion network shouldn't be routable beyond the host network for security purposes).
2. Not sure
3. You mean like is SSH enabled / disabled?
4. vDS is at the vCenter level right? So this is tough, normally vCenter has to be equal or greater than the Hosts, so assuming vCenter is 6.0 and your Hosts are still 5.5 your vDS would be 6.0 as well. I'm going to say yes it can because the vDS is at the vCenter level (IE if vcenter is down you can't add or remove from the vDS until vCenter is up)
5. Not sure - man i need to study more if i'm taking this exam in a few weeks

Lexluethar

Not to step on any toes but thought i'd put my own questions up here as well to keep the content rolling. I'm studying for the VCP 5.5 myself so I run into a lot of potential exam questions daily.

Here we go, all pertaining to 5.5 (answers in white):

1. What is the vCPU maximum per virtual machine? 64

2. What is the maximum RAM per virtual machine? 1 TB

3. What is the maximum virtual disk size? 62 TB

4. What is the maximum video memory per virtual machine? 512 MB

5. What is the maximum logical CPUs per host? 320

6. What is the maximum RAM per host? 4 TB

7. Explain what a NUMA Node is and why it's important to be aware of NUMA in a virtual environment?

Essendon

Thanks for stepping up Lex. I've been away since my wife gave birth to twin daughters on the 16th. I'll get back on the horse in a day or two.

Lexluethar

No worries essendon - i'm studying this crap daily so i have a dozen questions a day i'm reviewing in my head.

Essendon

Good questions, though just so you know VMware dont ask questions like these (config maximums) in the test nowadays. So while this information is good to know and I'd expect people to know some of it, it's unlikely you'd see it on the test. Point is, no need to memorize them.

Lexluethar

SOB for real!?!? Damn, thought they would throw dumb Microsoft exam questions at me, guess i won't study them then

Deathmage

Lexluethar wrote: »

SOB for real!?!? Damn, thought they would throw dumb Microsoft exam questions at me, guess i won't study them then

The VCP exam was no joke...

Lexluethar

I've heard it's not a cake walk, guess i'll find out in a few weeks when i take it. I've taken a good amount of MS exams but nothing from vmware so it's hard to gauge what to expect. MS was all over the place depending on what exam you were taking.

I have a solid understanding of vmware and how things work - what i'm trying to focus on are the specifics like services, how resource pools interact - things like that. Using the blueprint a lot as well.

Essendon

You must work on getting all those storage and networking concepts down pat too. Create scenarios for yourself based on things like:

- need to delete a LUN without downtime to VMs
- upgrade datastores
- migrate between arrays
- perform SDRS operations
- create a noisy neighbour using a load gen and configure SDRS as to how you can avoid the issue
- manual/auto SDRS and why?
- create datastore clusters and play with them and break them
- migrate between a vSS and vDS without downtime
- upgrade vDS
- play with NIOC to your heart's content, create user resource pools
- add/remove uplinks to switches
- swap between load balancing mechanisms and see what happens and why
- swap between PSPs and see what happens and why
- etc etc..

Get the drift? This things are important to know in the real world and on the exam too..

Lexluethar

Essendon wrote: »

You must work on getting all those storage and networking concepts down pat too. Create scenarios for yourself based on things like:

- need to delete a LUN without downtime to VMs - vmotion off servers, properly unmount in vmware prior to dismounting on storage array
- upgrade datastores - Right click within the storage view and upgrade (idk if i can do this with VMs on it though)
- migrate between arrays- storage vmotion from one to the other, unmount the old
- perform SDRS operation- definitely need to study this more, we don't use it so i don't have much experience
- create a noisy neighbour using a load gen and configure SDRS as to how you can avoid the issue
- manual/auto SDRS and why?- I would think auto SDRS if you don't have resource contention on the storage end and you ensure you've properly allocated your storage tiers to what your vms should have
- create datastore clusters and play with them and break them
- migrate between a vSS and vDS without downtime We've had to do this as well when we moved to a vDS - we had to 'walk the vms' from the vSS to the vDS by migrating one vmkernel port at a time, never migrating both on the same kernel at the same time otherwise the vm would drop
- upgrade vDS - haven't done this before - backed up before but not upgraded, need to do this as well
- play with NIOC to your heart's content, create user resource pools Definitely need to do more of this - investigated it but never implimented
- add/remove uplinks to switches
- swap between load balancing mechanisms and see what happens and why
- swap between PSPs and see what happens and why Unfortunately have had to do this a bunch lately b/c our storage array's upgraded their PSP's - so installing the vib (host has to be in maintenance mode) and restart. If the PSP is installed already you can change the storage multipathing on the fly but it won't stick on a restart. Also found out today installing a new VIB will most definitely reset your iSCSI login timeout period to 60 by default (might be the default of the VIB being installed).
- etc etc..

Get the drift? This things are important to know in the real world and on the exam too..

Sorry took your questions as a way to see what i already knew. Let me know if i was wrong on those above. I get your point though, my problem is being a our vmware engineer i only have experience in what we use - so things like NIOC, SDRS are things we don't do so i don't mess with. I have a lab and plenty of extra hardware at work to mess with though.

Thanks again for all of your help!

Deathmage

Q: In an EXSi 6.0 deployment with linked-mode vCenter and a standalone PSC VM, you get the full feature(s) of VUM from the VMware Web Client.

True or False?

Deathmage

Essendon wrote: »

Thanks for stepping up Lex. I've been away since my wife gave birth to twin daughters on the 16th. I'll get back on the horse in a day or two.

Ahhh I see, now I see why you've been slacking with the VCDX, you've been leveling up the wife the past few months. I guess you were getting wife aggro, and even with capped parry you still couldn't dodge it anymore. You need to stack more stamina, agility and intellect man, you need to feral tank better.

If you don't play World of Warcraft Manny you won't follow the above sentence at all...

Lexluethar

Updating my answer - true. Looks like now you can attach and scan items for remediation. You cannot though configure baselines from the web client - that is still done through the local vsphere client.

Nadzz

Excellent thread, thanks guys, keep it up