VCAP · VCF Storage · Exam 3V0-23.25 · VMware Cloud Foundation 9

This is the broadest of the advanced VCF exams. Its blueprint publishes thirty-four objectives and no sub-objectives at all, and it fills every one of the five blueprint sections: architecture, products, design, build and troubleshooting. Nothing else in the set asks for quite that range, and nothing else gives you so little nesting to hide behind. Hence a workbook of 214 tasks, one group per objective.

The examWhat you are signing up for

ShapeThirty-four objectives, and where they sit

Section 4 is more than half the blueprint and it is almost entirely vSAN: deploying clusters, stretched clusters, two-node clusters, encryption, File Service, the iSCSI Target Service, Data Protection, capacity sharing, policies, and the Day 2 work that follows. The nine objectives in sections 1 to 3 are written work you can do anywhere, and they are the cheapest marks in the exam.

ComparedAgainst the rest of the set

Lining the blueprints up by section, taken from the published guides rather than from memory:

Storage and VKS are the two that fill every section, which makes them the broadest preparation of the set: you are asked to choose an architecture, design against requirements, build the thing, and then fix it. Operations sits at the other extreme with everything in one section. If you are deciding which technology VCAP to sit first, that difference matters more than the subject does.

ESAThe thing to get right first

Objective 2.1 is a single line in the blueprint, differentiate between vSAN OSA and ESA, and it is the most load-bearing line in the document. ESA changes the data path, removes disk groups, changes how compression works, changes the RAID options and the host counts that go with them, and several services in section 4 exist only on ESA. Get this wrong in your head and you will get a question wrong in sections 2, 3, 4 and 5.

Section C of the workbook below is nine tasks on this one objective, and the first timed drill in section X is writing the comparison table from memory. That is deliberate.

PracticeWhere to get your hands on it

• A nested lab is viable here, more than for most of this set. vSAN runs nested well enough to practise cluster builds, policies, resynchronisation, fault domains and most of the troubleshooting. Performance figures will be meaningless, which does not matter for any objective on this blueprint.
• Your own environment, read-only, covers more than you would think: component layouts, policy compliance, health checks, capacity views and the whole of sections 1 to 3.
• VMware Hands-on Labs for the services you cannot easily nest, and for ESA if your lab hardware cannot meet its device requirements.

One warning that matters more on this exam than the others. Sections O, R, W and parts of U change or destroy data: enabling encryption rewrites a cluster, filling a datastore past its thresholds has consequences, and removing devices reduces redundancy while the rebuild runs. Do those in a lab. The tasks say so where it applies, but it is worth saying once here too.

Workbook214 tasks, in build order

Sections A and B are architecture, written work. C to G are the products and the vocabulary. H to J are design and sizing. K to U are the build, which is where most of the blueprint lives. V and W are monitoring and troubleshooting, and X is a set of timed drills. The numbers in grey at the right are the objectives, so a weak score report points you back at a group of tasks.

Objectives 1.1 and 1.2. Written work, and the only section you can do on a train.

1. 1Write what hyperconverged storage is, in one sentence, without using the word hyperconverged.1.1›
2. 2Write what traditional external storage is, in the same way.1.1›
3. 3Build a table comparing the two on scaling, failure domains, cost and operational ownership.1.1›
4. 4Record who owns storage in each model, and which team gets paged at night.1.1›
5. 5Given a greenfield site with no storage team, choose an architecture and justify it.1.2›
6. 6Given an estate with an existing array and three years left on support, choose and justify.1.2›
7. 7Write the one requirement that would overturn each of those two answers.1.2›

Objective 1.3. Five technologies, and the use case each one wins.

1. 8Write what vSAN is and the use case it wins.1.3›
2. 9Write what iSCSI is and when you would still choose it.1.3›
3. 10Write what NFS is and the workload that suits it.1.3›
4. 11Write what Fibre Channel is and what justifies its cost.1.3›
5. 12Write what NVMe over Fabrics brings, and what it demands of the network.1.3›
6. 13Record which of these VCF supports as principal storage and which as supplemental.1.3›
7. 14Build one table: protocol, typical latency, failure domain, and what it needs from the network.1.3›
8. 15Given a latency-sensitive database, pick a storage type and defend it against the other four.1.3›

Objective 2.1. Get this wrong and a third of the exam goes with it.

1. 16Draw the OSA data path, from guest write to capacity tier.2.1›
2. 17Draw the ESA data path, and mark where it differs.2.1›
3. 18Write what a disk group is, and record that ESA does not have them.2.1›
4. 19Record the device requirements for each architecture.2.1›
5. 20Record how compression works in each, and what is always on in ESA.2.1›
6. 21Record how deduplication differs between the two.2.1›
7. 22Record which RAID levels each supports, and the host count each needs.2.1›
8. 23Record which features are ESA only, and keep that list where you will see it.2.1›
9. 24Given an existing OSA cluster, write what moving to ESA would actually involve.2.1›

Objective 2.2. The vocabulary every other objective assumes.

1. 25Draw a vSAN cluster and label every component it is made of.2.2›
2. 26Write what an object is in vSAN, and list the object types.2.2›
3. 27Write what a component is, and how an object becomes several of them.2.2›
4. 28Record what a witness component does and why an object needs an odd number of votes.2.2›
5. 29Open an existing virtual machine and map its objects and components by hand.2.2›
6. 30Record where the vSAN traffic runs and what that VMkernel port needs.2.2›
7. 31Record what CLOM, DOM, LSOM and CMMDS each do, in one line each.2.2›
8. 32Record the storage pool in ESA and how it differs from a disk group.2.2›
9. 33Record the maximum component count per host for your version, and why it matters.2.2›

Objective 2.3. A VCF-specific distinction that catches people out.

1. 34Write what principal storage means in a VCF workload domain cluster.2.3›
2. 35Write what supplemental storage means.2.3›
3. 36Record which storage types VCF accepts as principal, and which only as supplemental.2.3›
4. 37Record when principal storage is chosen, and whether it can be changed afterwards.2.3›
5. 38Open your own workload domain and record what is principal and what is supplemental.2.3›
6. 39Record what the management domain requires, and whether that differs.2.3›
7. 40Given a cluster that must present both vSAN and an existing array, write how you would build it.2.3›

Objective 2.4. Know what each one is for before you are asked to configure it.

1. 41List the advanced vSAN services and write one line on each.2.4›
2. 42Record the use case for vSAN File Service.2.4›
3. 43Record the use case for the vSAN iSCSI Target Service.2.4›
4. 44Record the use case for vSAN Data Protection, and which architecture it needs.2.4›
5. 45Record the use case for a stretched cluster, and what it does not protect against.2.4›
6. 46Record the use case for a two-node cluster.2.4›
7. 47Record the use case for cross-cluster capacity sharing.2.4›
8. 48Record the use case for data-at-rest encryption.2.4›
9. 49Mark which of these your own estate actually uses, and which you have only read about.2.4›

Objectives 2.5 and 2.6. Where storage meets the rest of the platform.

1. 50Write what a datastore is and list the types VCF supports.2.5›
2. 51Write what a datastore cluster is and what it gives you.2.5›
3. 52Record what Storage DRS does, and what it must not do to a vSAN datastore.2.5›
4. 53Record the components of a supported storage solution, end to end.2.5›
5. 54Write how storage is presented to a vSphere Supervisor.2.6›
6. 55Record what a storage policy becomes on the Kubernetes side.2.6›
7. 56Record what a persistent volume maps to underneath.2.6›
8. 57Open a namespace and trace one persistent volume down to its vSAN objects.2.6›

Objective 3.1. Written design work, with no console open.

1. 58Read a requirements scenario and separate requirements from constraints and assumptions.3.1›
2. 59Record the availability requirement and turn it into a failures to tolerate number.3.1›
3. 60Choose OSA or ESA for the scenario and justify it in three sentences.3.1›
4. 61Choose a fault tolerance method and record what it costs in usable capacity.3.1›
5. 62Design the network for the cluster, and record the bandwidth it needs.3.1›
6. 63Decide the host count, and record what drives it.3.1›
7. 64Record the design decisions you made, each with its justification and its risk.3.1›
8. 65Write what you would change if the budget were cut by a quarter.3.1›
9. 66Write what you would change if the availability requirement doubled.3.1›

Objective 3.2. Arithmetic, and the place guesses get found out.

1. 67Record the raw capacity of one host, and of the cluster.3.2›
2. 68Apply the fault tolerance method and record the usable capacity.3.2›
3. 69Record the slack space the design needs and why.3.2›
4. 70Record what compression and deduplication do to your sizing, and whether you may count on them.3.2›
5. 71Size for the failure of one host, and record what must stay free.3.2›
6. 72Record the effect of a storage policy change on an already sized cluster.3.2›
7. 73Size the same workload under RAID-1 and under RAID-5, and compare.3.2›
8. 74Record what you would tell a customer who sized on raw capacity.3.2›
9. 75Produce one sizing you would put your name to, with every assumption listed.3.2›

Objective 3.3. The half of design that is not vSAN.

1. 76Read a scenario that rules out vSAN, and record why it is ruled out.3.3›
2. 77Choose a storage protocol for it and justify the choice.3.3›
3. 78Design the path from host to array, including multipathing.3.3›
4. 79Record the multipathing policy you chose and what the array vendor recommends.3.3›
5. 80Record how many paths each host needs and what a single point of failure would look like.3.3›
6. 81Design the datastore layout: how many, how big, and why.3.3›
7. 82Record what the design needs from the storage team, in their language.3.3›
8. 83Write the one thing most likely to be missed at implementation.3.3›

Objective 4.1. Section 4 begins, eighteen objectives of build work.

1. 84Record the prerequisites for a vSAN cluster in a VCF workload domain.4.1›
2. 85Check the hardware against the compatibility list and record anything that fails.4.1›
3. 86Configure the vSAN VMkernel networking and confirm it is working.4.1›
4. 87Deploy a vSAN cluster within a workload domain.4.1›
5. 88Record each decision the deployment asked for.4.1›
6. 89Confirm the datastore is created and its capacity matches what you expected.4.1›
7. 90Run the health checks and resolve or record every warning.4.1›
8. 91Record how long the deployment took, end to end.4.1›
9. 92Record what you would check first if the cluster formed but the datastore was undersized.4.1›

Objectives 4.2 and 4.14. The build and the day after.

1. 93Draw a stretched cluster: both sites, the witness, and the traffic between them.4.2›
2. 94Record the latency and bandwidth requirements for each link.4.2›
3. 95Record what the witness is, where it goes, and what it must not be.4.2›
4. 96Deploy a vSAN stretched cluster within a workload domain.4.2›
5. 97Configure the fault domains and confirm each host landed in the right one.4.2›
6. 98Create a storage policy with site mirroring and apply it.4.2›
7. 99Record what secondary failures to tolerate does within a site.4.2›
8. 100Fail one site deliberately and record what happened to the workloads.4.14›
9. 101Recover the site and record how long resynchronisation took.4.14›
10. 102Fail the witness and record what still works and what does not.4.14›
11. 103Record the Day 2 tasks specific to a stretched cluster, and who performs them.4.14›

Objective 4.3. Small, awkward, and on the exam.

1. 104Record what a two-node cluster is for and where it is typically deployed.4.3›
2. 105Record what the witness does here and where it can live.4.3›
3. 106Record the networking options, including direct connect between the nodes.4.3›
4. 107Deploy a two-node vSAN cluster.4.3›
5. 108Confirm the witness is working and the objects are placed as you expected.4.3›
6. 109Record what happens when one node fails.4.3›
7. 110Record what happens when the witness fails.4.3›
8. 111Record the storage policy limits of a two-node cluster.4.3›

Objective 4.15. The control surface for everything vSAN does.

1. 112List the rules a vSAN storage policy can contain.4.15›
2. 113Create a policy with a failures to tolerate setting and apply it.4.15›
3. 114Create a policy using RAID-5 or RAID-6 and record the host count it needs.4.15›
4. 115Configure disk striping in a policy and record what it changes.4.15›
5. 116Configure an IOPS limit and prove it is enforced.4.15›
6. 117Configure a space reservation and record its effect on the datastore.4.15›
7. 118Apply a new policy to a running virtual machine and watch the resynchronisation.4.15›
8. 119Record how much data moved and how long it took.4.15›
9. 120Record what happens when a policy cannot be satisfied.4.15›
10. 121Record which policy is the default and where that is set.4.15›

Objective 4.7. Key management is the part that goes wrong.

1. 122Record the difference between data-at-rest encryption and data-in-transit encryption.4.7›
2. 123Record the key provider options and what each one requires.4.7›
3. 124Configure a key provider and confirm it is trusted and available.4.7›
4. 125Enable vSAN data-at-rest encryption on a cluster.4.7›
5. 126Record what happened to the existing data, and how long it took.4.7›
6. 127Record the difference in how OSA and ESA perform encryption.4.7›
7. 128Perform a shallow rekey and record what it did.4.7›
8. 129Record what a deep rekey does instead, and what it costs.4.7›
9. 130Record what happens to the cluster if the key provider is unreachable at boot.4.7›

Objectives 4.8 and 4.17. A file server built out of the cluster you already have.

1. 131Record what vSAN File Service provides and which protocols it supports.4.8›
2. 132Record its prerequisites, including the network and the addresses it needs.4.8›
3. 133Configure the vSAN File Service on a cluster.4.8›
4. 134Record what it deployed, including the agent virtual machines.4.8›
5. 135Confirm the service is healthy and the file service domain is configured.4.8›
6. 136Create an NFS file share and mount it from a client.4.17›
7. 137Create an SMB file share and record what authentication it needed.4.17›
8. 138Apply a storage policy and a quota to a share.4.17›
9. 139Record how share permissions and access control are managed.4.17›

Objective 4.9. Block storage out of vSAN, for the things that need it.

1. 140Record what the vSAN iSCSI Target Service is for, and what it is not for.4.9›
2. 141Record its prerequisites and the network it needs.4.9›
3. 142Enable the iSCSI Target Service on a cluster.4.9›
4. 143Create a target and record what it asked for.4.9›
5. 144Create a LUN within the target and apply a storage policy to it.4.9›
6. 145Configure an initiator group and add an initiator to it.4.9›
7. 146Connect from a client and confirm the LUN is usable.4.9›
8. 147Record what authentication options exist and configure one.4.9›

Objectives 4.4, 4.10 and 4.16. Native snapshots, and the architecture they need.

1. 148Record what vSAN Data Protection provides, and which architecture it requires.4.4›
2. 149Record its prerequisites, including what must be deployed first.4.4›
3. 150Deploy the vSAN Data Protection components.4.4›
4. 151Confirm it is registered and healthy.4.4›
5. 152Configure vSAN Data Protection on a cluster.4.10›
6. 153Record how its snapshots differ from ordinary virtual machine snapshots.4.10›
7. 154Create a protection group and add workloads to it.4.16›
8. 155Create a recovery plan with a schedule and a retention.4.16›
9. 156Take a snapshot and record where it is stored and what it consumed.4.16›
10. 157Restore a workload from a snapshot and time the recovery.4.16›

Objective 4.6. Disaggregation, and the newest thing in this blueprint.

1. 158Write what vSAN cross-cluster capacity sharing does, in one sentence.4.6›
2. 159Write what a vSAN storage cluster is and how it differs from a normal cluster.4.6›
3. 160Record the requirements for both, including version and networking.4.6›
4. 161Configure cross-cluster capacity sharing between two clusters.4.6›
5. 162Mount the remote datastore on the client cluster and confirm it is usable.4.6›
6. 163Deploy a workload onto the shared capacity and record where its objects live.4.6›
7. 164Record what happens to the client cluster if the server cluster is unavailable.4.6›
8. 165Record when you would design for this, and when you would not.4.6›

Objectives 4.5, 4.11 and 4.12. The other half of section 4.

1. 166Record what VCF supports as principal storage for a non-vSAN workload domain cluster.4.5›
2. 167Record the prerequisites the storage must meet before VCF will accept it.4.5›
3. 168Deploy a workload domain cluster backed by non-vSAN storage.4.5›
4. 169Record each decision the deployment asked for, and how it differed from vSAN.4.5›
5. 170Present a new LUN or export to the hosts and rescan.4.11›
6. 171Create a datastore on it and confirm every host can see it.4.11›
7. 172Record the multipathing policy in use and change it deliberately.4.11›
8. 173Create a datastore cluster and add datastores to it.4.12›
9. 174Configure Storage DRS and record the automation level you chose.4.12›
10. 175Record what Storage DRS will and will not move, and what it needs to do so.4.12›

Objectives 4.13 and 4.18. The work that fills the years after the build.

1. 176List the Day 2 tasks a vSAN cluster needs, and how often each one is due.4.13›
2. 177Put a host into maintenance mode and record every data evacuation option.4.13›
3. 178Choose one option, run it, and record how much data moved.4.13›
4. 179Add a host to the cluster and confirm the capacity and components rebalanced.4.13›
5. 180Remove a disk or storage device and record what the cluster did.4.13›
6. 181Update vSAN as part of a cluster upgrade and record the order it enforced.4.13›
7. 182Record where resynchronisation is monitored and what a healthy one looks like.4.13›
8. 183Grow a non-vSAN datastore and confirm the hosts see the new size.4.18›
9. 184Unmount and detach a datastore cleanly, in the right order.4.18›

Objectives 5.1 and 5.2. Section 5 begins.

1. 185List the tools VCF gives you for monitoring vSAN and write what each is for.5.1›
2. 186Run the vSAN health checks and record every category they cover.5.1›
3. 187Open vSAN performance and record the metrics that matter at each layer.5.1›
4. 188Find the capacity view and record what it reports beyond used and free.5.1›
5. 189Set up a vSAN alert or alarm and trigger it.5.1›
6. 190Monitor a non-vSAN datastore and record which metrics are available.5.2›
7. 191Record what you can see about the array from VCF, and what you cannot.5.2›
8. 192Record which monitoring question needs the storage team, and what to ask them for.5.2›

Objectives 5.3 and 5.4. Create each fault, then diagnose it.

1. 193Write your diagnostic order for a vSAN performance complaint.5.3›
2. 194Break vSAN networking on one host and diagnose it from the symptoms.5.3›
3. 195Record what the health checks reported and how quickly.5.3›
4. 196Cause an object to become non-compliant and resolve it.5.3›
5. 197Cause a resynchronisation and diagnose why it is running.5.3›
6. 198Fill a vSAN datastore past a threshold and record what the cluster did.5.3›
7. 199Record what a failed capacity device looks like, and the recovery steps.5.3›
8. 200Break a path to a non-vSAN datastore and observe the failover.5.4›
9. 201Record what an all paths down event looks like, and how it differs from permanent device loss.5.4›
10. 202Diagnose a datastore that will not mount on one host only.5.4›
11. 203Write the five questions you would ask before escalating a storage problem.5.4›

Do this section twice in the final week, from a clean start, timing yourself.

1. 204Write the OSA against ESA comparison table from memory. 10 minutes.2.1›
2. 205Size a vSAN cluster for a stated requirement, showing your working. 25 minutes.3.2›
3. 206Deploy a vSAN cluster in a workload domain and pass the health checks. 30 minutes.4.1›
4. 207Create three storage policies and apply one to a running workload. 20 minutes.4.15›
5. 208Configure a stretched cluster fault domain layout and a site mirroring policy. 25 minutes.4.2›
6. 209Enable encryption with a key provider, then perform a shallow rekey. 25 minutes.4.7›
7. 210Configure File Service and publish an NFS share a client can mount. 30 minutes.4.8›
8. 211Configure the iSCSI Target Service and connect an initiator. 25 minutes.4.9›
9. 212Configure Data Protection and restore one workload from a snapshot. 30 minutes.4.16›
10. 213Put a host in maintenance mode with full data migration and bring it back. 20 minutes.4.13›
11. 214Diagnose one vSAN fault and one non-vSAN fault that somebody else created. 25 minutes.5.3›

Four weeksIf that is what you have

Two pieces of advice. Learn the OSA against ESA table until you can write it from memory, including the host counts for each RAID level, because that one table reaches into four of the five sections. And do the troubleshooting section by creating each fault yourself rather than waiting to meet it: a fault you built is one where you already know the answer, so you can judge whether your diagnostic order was any good.

Exam details are from the published Broadcom exam guide for 3V0-23.25 and are accurate at the time of writing. Broadcom revises blueprints regularly, so confirm the current version before you book. Practice tasks here are my own, written from the published objectives; they are not exam content.