A Storage Account is the top-level container for all Azure storage services. One account = one namespace, one set of access keys, one firewall, one redundancy configuration. Everything storage-related sits inside it.
A Storage Account is not storage itself. It is the container, the envelope, the building that holds your storage. Inside one storage account you can have blob containers, file shares, queues, and tables, all completely different storage types, managed through one account with one set of access keys, one firewall, and one redundancy configuration. Think of it as a warehouse building. The building has one address, your storage account name, which becomes a globally unique DNS prefix like kubestgprod.blob.core.windows.net. The building has one security system, your account keys and firewall rules. Inside are different rooms, the blob room for unstructured data, the files room for shared drives, the queue room for messages, the table room for key-value data. The name is tested on almost every exam. Between 3 and 24 characters. Lowercase letters and numbers only. No hyphens, no underscores, no uppercase. Globally unique because it becomes part of a public DNS hostname. For account type, you almost always want GPv2, General Purpose version 2. It supports all four storage services, all access tiers, and all redundancy options. It is the default and right choice 95 percent of the time. One account equals one redundancy setting. You cannot have some blobs in LRS and others in GRS inside the same account. The redundancy choice is account-wide. That is why teams often use multiple storage accounts, hot operational data in a cheaper LRS account, compliance archives in a GRS account. The two account keys are like master keys. Anyone with either key has complete access to everything in the account. No expiry, no scope limitation. Treat them like the root password. Never put them in application code. Use SAS tokens for scoped access or Managed Identity for zero-credential access.. Remember: Storage Account is the container, not the data. Name must be 3 to 24 characters, lowercase and numbers, globally unique. GPv2 is your default. One account equals one redundancy, one firewall, one set of keys.
📖 Storage Account — Complete Explanation
Every piece of data in Azure Blob, Azure Files, Azure Queues, or Azure Tables lives inside a Storage Account. It's the top-level container — not the data itself, but the envelope that holds everything.
One account = one namespace. Your storage account name becomes the globally unique DNS prefix: kubestgprod.blob.core.windows.net. This is why names must be globally unique — two storage accounts cannot share a public hostname.
One account = one set of keys. There are exactly two account keys (Key1 and Key2). Anyone with either key has full access to everything inside — all blobs, all files, all queues. This is why you almost never use account keys directly. Instead you use SAS tokens (scoped, time-limited) or Managed Identity (no keys involved).
One account = one firewall. Storage firewall rules apply at the account level, not per container. If you restrict to a VNet subnet, that restriction applies to all blobs, files, and queues in that account.
One account = one redundancy configuration. You can't have some blobs in LRS and others in GRS inside the same account. The redundancy choice is account-wide. This is why large organisations often have multiple storage accounts — one for hot operational data (LRS, cheaper) and one for compliance archives (GRS, more durable).
🏭
The Metaphor
A Storage Account is like a warehouse building. The building has one address (endpoint), one security system (access keys + firewall), and one backup policy (redundancy). Inside the building are different rooms: the Blob room (open shelves — any box, any size), the Files room (filing cabinets — familiar Windows share), the Queue room (conveyor belt — messages in order), and the Table room (index cards — simple key-value data).
Storage Account — Complete Architecture
Storage Account Types
Type
Services
Access Tiers
Redundancy
Use
GPv2 (General Purpose v2)
Blob, Files, Queue, Table
Hot/Cool/Cold/Archive
All options
DEFAULT — always use this
Premium Block Blob
Blob ONLY
Always Hot
LRS / ZRS only
IoT telemetry, analytics, low latency <1ms
Premium File Shares
Files ONLY
N/A
LRS / ZRS only
High-performance SMB file shares
GPv1 (legacy)
All
None
All options
Avoid — upgrade to GPv2
💡
Kube context
For Kube IoT sensor data (high write throughput, sub-millisecond latency required): Premium Block Blob. For general config files, reports, exports: GPv2. Premium cannot have Archive tier — all data stays Hot.
Naming Rules — Exam Tested
Storage Account Naming
3–24 Characters · Lowercase + Numbers Only · Globally Unique
Why globally unique: the storage account name becomes part of the public DNS endpoint (name.blob.core.windows.net). Two accounts cannot share a DNS name.
⚠️
Most Failed Naming Exam Questions
kube-storage — INVALID (has hyphen) KubeStorage — INVALID (uppercase letters) ps — INVALID (only 2 chars, minimum is 3) Common trap: they give you 4 names and ask which is valid — look for hyphens and capitals first.
Azure replicates your data automatically. You choose HOW MANY copies and WHERE. More geographic spread = higher durability = higher cost.
Azure automatically makes copies of your data. The question is how many copies and where. LRS, Locally Redundant Storage, makes three copies all in the same datacenter. Spread across different racks, so a single server or rack failure is fine. But if the datacenter itself has a catastrophic failure, all three copies go down together. LRS is the cheapest option, fine for data you can recreate or data that does not need to survive regional disasters. ZRS, Zone-Redundant Storage, makes three copies spread across three different availability zones in the same region. These are physically separate datacenters connected by high-speed private fibre. If one datacenter goes down, the other two have your data. ZRS keeps your data in one region, which matters for data sovereignty. GRS, Geo-Redundant Storage, makes six copies total. Three in your primary region using LRS, then Azure asynchronously replicates to a paired region hundreds of kilometres away. Three more copies there. If East US has a complete regional disaster, your data survives in West US. But here is. the exam trap with plain GRS: you cannot read from the secondary region during normal operations. Only after a failover. RA-GRS, Read-Access Geo-Redundant Storage, adds a readable endpoint for the secondary at all times. Your application can read from the secondary endpoint right now without waiting for failover. GZRS combines zone redundancy in the primary region plus geo-replication to a secondary. RA-GZRS adds read access to that secondary. Decision tree: data must stay in one region, use LRS or ZRS. Survive regional outage, use GRS or GZRS. Read secondary without waiting for failover, use RA-GRS or RA-GZRS. Premium storage only supports LRS and ZRS.. Remember: LRS is one datacenter, ZRS is three zones, GRS is two regions, RA-GRS adds readable secondary. Premium storage only gets LRS or ZRS.
💾
The Metaphor
LRS: 3 copies in the same room. Room burns down → all gone. ZRS: 3 copies in 3 different buildings (zones) in the same city. City floods → all gone. GRS: 3 copies in your city + 3 copies in another city 500km away. Your city destroyed → second city has it. GZRS: Best of both — 3 copies across 3 buildings in your city AND 3 in another city.
Redundancy — Visual Comparison
SKU
Copies
Zones
Regions
Read Secondary
Data stays in 1 region?
LRS
3
1
1
✗
✓ Yes
ZRS
3
3
1
✗
✓ Yes
GRS
6
1
2
Failover only
✗ No
GZRS
6
3
2
Failover only
✗ No
RA-GRS
6
1
2
✓ Always
✗ No
RA-GZRS
6
3
2
✓ Always
✗ No
⚠️
Premium Storage + Redundancy Limitation
Premium Block Blob and Premium File Shares only support LRS or ZRS. No GRS, no RA-GRS. If compliance requires geo-redundancy AND you need premium performance — you cannot have both in a single account. Use GPv2 GRS for compliance or accept LRS for premium.
📋
Scenario → Answer Map
"Data MUST stay in one region (compliance/sovereignty)" → LRS or ZRS only
"Survive complete regional outage" → GRS or GZRS
"Read from secondary during normal operations" → RA-GRS or RA-GZRS
"Cheapest possible" → LRS
"IoT data, highest write throughput" → Premium Block Blob (LRS)
Blob = Binary Large OBject. Unstructured data — any format, any size. The workhorse of Azure storage. IoT sensor data, images, videos, backups, logs — all go here.
Blob stands for Binary Large Object. Basically any file, any format, any size. Images, videos, sensor data, backups, log files, database exports, all of it is blob storage. Three blob types. Block blobs are the default, made of blocks of data, up to about 190 terabytes. Use these for virtually everything, IoT sensor data, images, videos, backups. Append blobs are write-once, append-only. You can add data to the end but never modify existing content. Perfect for log files and audit trails where tamper-evidence matters. Page blobs store data in 512-byte pages and support random reads and writes. Designed specifically for virtual machine disks. Access tiers determine cost and availability. Hot tier is for frequently accessed data. Highest storage cost, lowest access cost. Active dashboard data, recent sensor readings. Cool tier is for data you access occasionally. Lower storage cost, higher access cost. Minimum storage duration of 30 days. Delete before day 30, you still pay for 30 days. Cold tier is for rarely accessed data. Minimum 90 days. Even lower storage cost, still immediately readable. Archive tier is the cheapest available. Minimum 180 days. But you cannot read an archived blob directly. You must first rehydrate it, change its tier to Hot or Cool. Standard priority rehydration takes up to 15 hours. High priority takes under an hour but costs more. This is the most tested Archive gotcha on AZ-104. Lifecycle management policies automate tier transitions. Write a rule once: move blobs to Cool after 30 days, Archive after 90 days, delete after 7 years. Azure evaluates these rules daily.. For Kube IoT data: fresh sensor readings stay Hot, move to Cool after 30 days, Archive after 90, delete after 7 years. One lifecycle policy, no manual work ever.. Remember: Block blobs for almost everything. Append blobs for logs. Page blobs for VM disks. Archive requires rehydration before reading, up to 15 hours standard priority. Early deletion penalties apply.
Hot = items on your desk. Instant access, most expensive real estate. Cool = items in a nearby filing cabinet. Still accessible, cheaper space. Cold = items in the office storeroom. Rarely needed, cheaper rent. Archive = items in an off-site warehouse. Cheapest storage, but you need to order them and wait for delivery (rehydration).
Access Tiers — Storage Cost vs Access Cost Tradeoff
⚠️
Archive Rehydration — Exam Always Tests This
Cannot read Archive blobs directly — get error 409 BlobArchived.
Must rehydrate first (change tier to Hot or Cool): Standard priority: up to 15 hours — cheaper High priority: under 1 hour — costs more, for urgent recovery
Early deletion penalty: Move blob to Archive at day 1, then delete at day 50 — you still pay for 180 days. Same for Cool (30 days) and Cold (90 days). Azure charges the MINIMUM duration regardless of actual time stored.
Lifecycle Management Policy
Automated Tier Movement
Set Rules Once — Azure Moves Blobs Automatically
Lifecycle policies define: IF blob is X days old AND matches filters → THEN move to tier or delete.
Example For Kube IoT sensor data:
Day 0–30: Hot (active queries)
Day 30–90: Cool (occasional access)
Day 90–365: Archive (compliance)
Day 365: Delete
Important: Policies run ONCE PER DAY. Changes take up to 24 hours to apply. Not real-time. New policy can take up to 24 hours before its first evaluation.
Azure Files provides fully managed file shares accessible via SMB or NFS — the same protocol as Windows shared drives. Lift-and-shift file servers to cloud without changing how apps access them.
Azure Files is Microsoft's managed file share service. The key insight: it works exactly like the Windows shared drives you already know, but instead of a file server in a cupboard, it is a fully managed cloud service. When you mount an Azure File Share, it appears as a drive letter on Windows or a mount point on Linux. Applications do not know they are talking to the cloud. They think they are using a local network drive. Zero application changes when migrating from on-premises file servers. Two protocols. SMB, Server Message Block, is the Windows file sharing protocol running on port 445. Azure Files supports SMB 3.0 with encryption in transit. You can mount from Windows, Linux, and macOS. NFS, Network File System, is the Linux world's protocol running on port 2049. Azure Files supports NFS 4.1, but only on Premium tier accounts and only for Linux clients. The most common real-world problem: port 445 is blocked by many internet service providers and corporate firewalls. If someone tries to mount from home, they often fail silently. The fix is either VPN to the corporate network first, or use Azure File Sync. Azure File Sync installs an agent on your Windows Server. The agent creates a bidirectional sync between a local folder and an Azure File Share. All files exist in Azure at all times as the source of truth. On the local server, cold files are replaced with lightweight placeholders that stream from Azure transparently when you open them. This is called cloud tiering. Your server has a small fast SSD with only the hot files, while the full dataset lives in Azure. Fast local performance for active files, cloud economics for everything else, Azure durability for disaster recovery. For authentication: on-premises AD DS, Entra ID Kerberos, or Entra Domain Services.. Remember: Azure Files is managed SMB or NFS. Port 445 is often blocked, use VPN or File Sync. File Sync installs an agent on Windows Server and enables cloud tiering. Authentication supports on-premises AD, Entra ID Kerberos, and Entra Domain Services.
🗂️
The Metaphor
Azure Files is like your office's shared network drive (\\server\shared) — but instead of a physical server in the cupboard, Microsoft runs it. Your Windows laptop maps it as Z:, your Linux VM mounts it at /mnt/share. Both see the same files. No server to patch, no disk to replace, no backup to configure.
Feature
SMB (Standard)
NFS (Premium only)
Protocol
SMB 3.0 / 2.1
NFS 4.1
OS
Windows, Linux, macOS
Linux only
Storage tier
Standard GPv2
Premium File Shares only
Port
445 (often blocked by ISPs)
2049
Auth
AD DS / Entra ID / storage key
Storage key only
Encryption
In-transit (TLS)
In-transit NOT supported
⚠️
Port 445 Blocked at Home — Most Common Azure Files Issue
Many ISPs and home routers block TCP port 445 (SMB) as an anti-malware measure. Symptom: Azure file share mounts fine in office, fails at home.
Fix options: (1) Connect via VPN to corporate network, then mount. (2) Use Azure File Sync with local cache instead of direct mount. (3) Use Azure Bastion/Jump box to access from within the VNet.
Azure File Sync
🔄
The Metaphor
Azure File Sync turns your Windows Server into a local cache of the Azure file share. Hot files stay on local SSD (fast access). Cold files exist only in Azure (cloud tiering). When someone opens a cold file — it streams down transparently. Like Google Drive File Stream, but for enterprise file servers.
Azure File Sync — Bidirectional Sync Architecture
Blobfuse2
Mount Blob Container as Linux Filesystem
Blobfuse2 = a FUSE (Filesystem in Userspace) driver that mounts a Blob container as a Linux directory.
Your app reads/writes files normally (/mnt/blobdata/sensor.json) — under the hood it's actually reading/writing Azure Blob over HTTPS.
Compare: Azure File Sync = bidirectional sync with local cache. Blobfuse2 = direct mount, no local copy, every read/write goes to Azure. Different use cases.
Use: ML training jobs accessing large datasets, Linux apps that expect a filesystem path but data lives in Blob.
Four ways to access Azure Storage. Choose based on how much access you need to grant, to whom, and for how long.
How you secure access to Azure Storage is where most security incidents begin. Not from sophisticated attacks, but from mismanaged keys and overly permissive access. Four ways to control access: account keys, Shared Access Signatures, Entra ID, and network controls. Account keys are master keys. There are exactly two. Whoever has either key has complete unrestricted access to everything in that storage account. No expiry. No scope limitation. These must never appear in application code, environment variables, or source control. Shared Access Signatures, SAS tokens, are the right tool for most scenarios. A SAS is a URL granting specific scoped time-limited access. Read access only, to this specific container, for the next two hours, only from this IP address. Far safer than an account key. Three types of SAS. Account SAS covers all services, signed by account key. Service SAS covers one service, also signed by account key. User Delegation SAS is the most secure, signed by an Entra ID credential, not the account key. When. the exam asks for the most secure SAS type, the answer is always User Delegation SAS. Critical exam scenario: how do you revoke a SAS before it expires? If the SAS is backed by a Stored Access Policy, delete the policy and every SAS referencing it is immediately invalidated. If the SAS has no stored access policy, your only option is rotating the account key that signed it, which invalidates every SAS signed by that key, including ones used by other applications. Always back SAS tokens with Stored Access Policies. Entra ID authentication uses RBAC roles like Storage Blob Data Reader directly assigned to identities. For VM applications, use Managed Identity. No credentials in code at all. Network controls: IP rules, Service Endpoints requiring two steps, or Private Endpoints giving the storage account a private IP inside your VNet.. Remember: Account keys are master keys, treat accordingly. User Delegation SAS is most secure. Stored Access Policy is the only way to revoke SAS without rotating the key.
SAS Tokens — Shared Access Signatures
🎟️
The Metaphor
Your account key is a master key that opens everything forever. You'd never hand that to a contractor.
A SAS token is a scoped, time-limited visitor pass: "You can enter the Blob room only, read files only, for the next 2 hours, only from this IP address." When it expires — gone. No need to change the master key.
SAS Token Types — Security Ranking
Animated — SAS Revocation: With Policy vs Without Policy
⚠️
How to REVOKE a SAS Token — Critical Exam Question
If Stored Access Policy exists: Delete the policy → SAS immediately revoked ✅ (fastest, no impact on other apps)
If raw SAS (no policy): Only option is rotate the account key → ALL SAS tokens signed by that key are invalidated → breaks other apps using same key ⚠️
Best practice: Always back SAS tokens with a Stored Access Policy for easy, targeted revocation. One policy can control many SAS tokens.
Azure CLI — Generate SAS + Stored Access Policy
# Create Stored Access Policy on containeraz storage container policy create \
--account-namekubestgprod \
--container-namesensor-data \
--name"read-policy-2026" \
--permissionsr \
--expiry"2026-12-31"# Generate SAS backed by that policyaz storage container generate-sas \
--account-namekubestgprod \
--namesensor-data \
--policy-name"read-policy-2026" \
--outputtsv# Revoke: just delete the policyaz storage container policy delete \
--account-namekubestgprod \
--container-namesensor-data \
--name"read-policy-2026"# SAS tokens backed by this policy now INVALID ✅
Storage Firewall & Virtual Network Access
Method
What It Does
Cost
Traffic Path
On-Prem Access
No restriction
Storage accessible from anywhere (default)
Free
Public internet
✓
IP firewall rules
Allow specific public IPs only
Free
Public internet from allowed IPs
✓ if IP listed
Service Endpoint
VNet traffic via MS backbone, restrict to subnet
Free
VNet → MS backbone → Storage public IP
✗ Cannot use
Private Endpoint
Storage gets private IP inside your VNet
Hourly + data
Stays entirely in VNet
✓ Via VPN/ER
⚠️
Service Endpoint Requires TWO Steps — Most Missed
Step 1: Enable Service Endpoint (Microsoft.Storage) on the VM's subnet Step 2: Add VNet/subnet rule in Storage Account firewall → Selected Networks → Add VNet
Step 1 alone = VMs route via MS backbone but Storage still accepts traffic from anywhere.
Step 2 alone (without Step 1) = Rule exists but traffic still routes over internet. Both required for private-backbone + restricted access.
Identity-Based Access for Azure Files
Azure Files — Auth Options
Share-Level + File/Dir Level Permissions
Azure Files supports identity-based access — no storage key needed:
On-premises AD DS: join storage account to on-prem AD, use Windows NTFS permissions + SMB ACLs Entra ID Kerberos: for cloud-only (no on-prem DC needed), hybrid identities Entra Domain Services: Azure-managed AD DS
Two layers of permissions: Share-level RBAC: Storage File Data SMB Share Contributor/Reader Directory/file NTFS ACLs: granular file-level permissions via Windows File Explorer or icacls
📊 Access & SAS Tokens — was this clear?
✓ Saved
2.6 — Migration
Data Transfer & Migration Tools
Azure provides several tools for moving data into and out of Azure, and choosing the right one depends on how much data you have and how fast your internet connection is. AzCopy is the command-line tool for moving data to and from Azure Storage. It supports blobs, files, and tables. It uses parallel transfers to maximise throughput. Key commands: azcopy copy to upload or download, azcopy sync to synchronise a local folder with a blob container keeping only changed files, azcopy make to create a container. AzCopy authenticates using SAS tokens or Managed Identity.. For Kube, AzCopy is how CI/CD pipelines upload IoT firmware images to blob storage. Storage Explorer is the graphical alternative to AzCopy. A desktop application that connects to your storage accounts and lets you browse, upload, download, copy, and manage blobs, files, queues, and tables visually. No command line needed. Useful for one-off tasks and for people who prefer a GUI. Azure Import/Export service is for very large data movements where the internet is too slow. You physically ship hard drives to an Azure datacentre. For import, you copy data to drives, ship to Microsoft, they upload to your storage account. For export, Microsoft copies your data to drives and ships them to you. This is for petabyte-scale migrations or locations with very limited bandwidth. Azure Data Box takes this further with purpose-built hardware. Microsoft ships you a Data Box appliance, you connect it to your network, copy data locally at network speeds, then ship it back. Data Box holds 80 terabytes usable. Data Box Heavy holds 770 terabytes. Decision guide: network migration under a few terabytes, use AzCopy. One-off large upload, use AzCopy or Storage Explorer. Petabyte scale or poor internet, use Import/Export or Data Box.. Remember: AzCopy is the CLI tool, Storage Explorer is the GUI, Import/Export ships drives, Data Box ships appliances.
AzCopy: one-time or scheduled one-way copy. Not continuous. CLI-based. Azure File Sync: continuous bidirectional sync with local caching. Server-based agent.
Exam scenario: "Copy 500GB from storage account A to storage account B" → AzCopy
Exam scenario: "Windows file server needs to sync with Azure Files ongoing" → Azure File Sync
AzCopy — Key Commands
# Copy local folder to Blob containerazcopy copy'C:\sensor-data\*' \
'https://kubestgprod.blob.core.windows.net/sensor-data?<SAS>' \
--recursive# Copy between two storage accounts (no local download)azcopy copy \
'https://source.blob.core.windows.net/container?<SAS>' \
'https://dest.blob.core.windows.net/container?<SAS>' \
--recursive# Sync (like rsync — only copies changed/new files)azcopy sync \
'C:\reports' \
'https://kubestgprod.blob.core.windows.net/reports?<SAS>' \
--delete-destinationtrue# Authenticate with Managed Identity (recommended)azcopy login--identity# Then copy without SAS tokenazcopy copy'source''destination'
Click to reveal. Based on real AZ-104 exam patterns from last 2–3 years.
QA company needs to store 7 years of IoT sensor logs for compliance. The data is never accessed after 90 days. Minimize cost. What storage tier and settings?
ANSWER
GPv2 storage account + Archive tier after 90 days + Lifecycle Management Policy.
Configure lifecycle rule: move to Cool at 30 days, Archive at 90 days, delete at 7 years (2555 days).
Also enable WORM immutability policy (locked) if regulatory compliance requires data cannot be modified/deleted during retention period (e.g., SEC 17a-4).
⚠️ Trap: if using Archive, you cannot read data directly. Any compliance read access requires rehydration (up to 15 hours). Factor this into SLA planning.
QA storage account is shared with an external partner via SAS token. The partner's credentials were compromised. The SAS has no Stored Access Policy. How do you revoke access immediately?
ANSWER
Rotate the storage account key that was used to sign the SAS token.
This immediately invalidates ALL SAS tokens signed by that key. Side effect: any other app using that key or SAS tokens from it will also break — must update those apps with new key/SAS.
Lesson: Always create SAS tokens backed by Stored Access Policy. Then you can delete the policy to revoke specific SAS without affecting others.
QA VM cannot access Azure Storage even though a Service Endpoint is configured on the subnet. What is the most likely missing step?
ANSWER
Storage Account firewall not configured to allow the VNet/subnet.
Service Endpoint has 2 required steps:
✅ Step 1 done: Service Endpoint enabled on subnet
❌ Step 2 missing: Storage Account → Networking → Selected Networks → Add the VNet and subnet
Without Step 2, storage still accepts traffic from everywhere (default allow) but the Service Endpoint provides no restriction OR the firewall may be set to deny all, blocking the VM.
QYou need to grant an external vendor read access to specific blobs for 48 hours. The most secure method that doesn't use account keys?
ANSWER
User Delegation SAS token (signed by Entra ID credentials, Blob only, expiry 48 hours).
User Delegation SAS = most secure because:
1. Not signed by account key (key not exposed)
2. Backed by Entra ID identity (audited)
3. Max validity = 7 days (capped by Azure)
4. Revocable by revoking the delegating user's permissions
⚠️ User Delegation SAS works for Blob service only. Not supported for Files, Queue, or Table.
QAn admin deleted critical production blobs by mistake. Soft delete was NOT enabled. Is recovery possible?
ANSWER
If versioning was enabled: yes — restore from previous version. If soft delete AND versioning both disabled: no recovery possible.
Recovery options in order of preference:
1. Blob versioning: previous versions available if enabled
2. Azure Backup: operational backup for Blob (requires Backup Vault)
3. Object Replication: if configured to another account, copies exist there
4. No options if none of the above were configured
Best practice: enable both soft delete AND versioning on production storage.
QWhich is correct for a storage account name? (a) Kube-Storage (b) kubestorage2026 (c) ps (d) kube_storage
ANSWER
(b) kubestorage2026
(a) INVALID — uppercase P + hyphen
(b) VALID — lowercase letters and numbers, 22 chars ✅
(c) INVALID — only 2 chars (minimum 3)
(d) INVALID — underscore not allowed
Rules: 3–24 chars · lowercase letters and numbers ONLY · no hyphens · no underscores · globally unique
QA company needs storage that survives complete regional outage AND allows reading from secondary region without waiting for failover. What redundancy option?
ANSWER
RA-GRS (Read-Access Geo Redundant Storage) or RA-GZRS.
RA-GRS = 6 copies across 2 regions, secondary region readable always via -secondary suffix URL
RA-GZRS = same but primary region uses 3 availability zones (more resilient)
⚠️ GRS alone does NOT allow reading secondary. Must be RA-GRS or RA-GZRS for live read access.
Phase 2 — Cheat Sheet
Storage account name has hyphen — valid?INVALID — lowercase + numbers only, no hyphens