What an IBM DS8000 is and where you’ll meet it
IBM DS8000 storage is the big disk system behind the most important computers in a bank, an airline or a government office. When you pay by card, book a train ticket or check your salary slip, there’s a fair chance the data came off a DS8000, and it’s built to keep running even when a part inside it breaks.
This first lesson gives you the big picture: what sits inside the frame, how a server reads and writes data on it, and why companies spend so much on one. You’ll also set up the SaatVedha practice shell on your laptop and type your first DS CLI commands, the same ones storage teams use at work.
No storage background is needed here. If you know what a hard disk is and what a server is, that’s enough to start.
| Time | You need | By the end you can |
|---|---|---|
| 60 to 75 minutes | A Windows, macOS or Linux laptop with Python 3.8 or newer | Name the main parts of a DS8000, follow one write from the host to flash, and run lssi, lsserver and lsarraysite in the practice shell |
What it is, in one line
An IBM DS8000 is an enterprise storage system: two Power servers, a protected cache and flash drives in one frame that give many servers fast, always-on disks over Fibre Channel and FICON.
In Simple Words
A school library with two librarians
Picture your school library at lunch break, with hundreds of students wanting books at once. One librarian can’t keep up with that queue, so the school hires a second one, and the two of them sit at the same desk and split the work.
When a student returns a book, the first librarian writes it in her notebook and the second copies the same line into his, and only then do they tell the student “done”. So if one of them goes home sick, the other still has every record.
The books live on strong shelves at the back, but the librarians keep the popular ones on the desk, so most students get theirs in seconds. Nobody walks to the shelves; you ask for a book number at the desk.
In a DS8000 the two librarians are the two servers, their notebooks are its memory, the shelves are the flash drives and the book number is the disk number (LUN) your server asks for.
Engineer's View
What is really inside a DS8000 G10
The current family is called DS8000 G10 (tenth generation), and its models are the DS8A00 family. All models are all-flash and use machine type 5341:
| Model | Class | Processor cores (both CPCs) | System memory |
|---|---|---|---|
| DS8A10 (A01 racked, A00 rack-mounted) | Flexibility | 20 (10 per CPC) | 256 GB or 512 GB |
| DS8A50 (A05, optional E05 expansion frame) | Agility | 40 (20 per CPC) | 1 TB, 2 TB or 3.5 TB |
| DS8A80 (A08, optional E05 expansion frame) | Analytic | 48 (24 per CPC) | 4.8 TB |
Every DS8A00 has the same building blocks:
- Two CPCs (central processor complexes): IBM Power9+ servers called CPC 0 and CPC 1. Each one runs a partition with AIX and IBM’s storage code, known as the storage node (server0 and server1), and the two work at the same time, ready to take over from each other.
- Cache and NVS: most of the memory is read and write cache. A slice of it is NVS (non-volatile storage), where each write is kept as a second copy in the other CPC, and NVDIMMs with backup power modules save it if the power fails.
- I/O enclosures: these come in pairs, hold the host adapters and the flash RAID adapters, and connect to both CPCs over PCIe Gen4 cables.
- Host adapters: 4-port 32 Gbps cards. You set each port to FCP for open systems (Linux, Windows, VMware, AIX) or to FICON for IBM Z mainframes, and it runs at 32, 16 or 8 Gbps.
- HPFE Gen3 (High-Performance Flash Enclosures): pairs of 2U drawers holding 16, 32 or 48 NVMe flash drives, from 1.6 TB up to 19.2 TB each. The bigger drives are IBM FlashCore Modules (FCM), which compress data on their own.
- Two HMCs (Hardware Management Consoles): small servers at the bottom of the rack. You manage the system through them with the DS GUI and the DS CLI, and they call IBM when a part fails, but no data ever flows through an HMC.
A few limits worth remembering: up to 255 LSSs, 65,280 logical devices, LUNs up to 16 TiB and CKD volumes up to 1,182,006 cylinders (about 1 TB). RAID 6 is the default. IBM’s service team (the SSR) installs the hardware, and you look after the planning, the licence keys and the logical setup.
On DS8900F (the previous generation, which you’ll still see a lot at work) the design is the same with Power9 servers and PCIe Gen3, so nearly everything here applies to it as well.
How it works, step by step
Follow one piece of data through the system. A Linux database server wants to save a 4 KB block on its disk /dev/mapper/mpatha, which to the server looks like a normal local disk, but inside the DS8000 quite a lot happens.
1The host sends the write over the SAN
The server’s HBA (host bus adapter) sends the write as a Fibre Channel frame through a SAN switch (a mainframe would use FICON instead), and the frame arrives at one port of a host adapter in an I/O enclosure.
A good setup gives every host at least two paths, through two host adapters in two different I/O enclosures, so if one cable or card dies the other path carries on.
2The host adapter hands it to the right server
Every volume belongs to one of the two CPCs, and the LSS number decides which: volumes in even-numbered LSSs (logical subsystems, like 20 or 2A) belong to server0 in CPC 0, and those in odd LSSs (21, 2B) belong to server1 in CPC 1.
The host adapter knows who the owner is and passes the write over PCIe Gen4 to that server. LSSs get a proper treatment in Lesson 14; for now, remember “even goes left, odd goes right”.
3Two copies, in two different servers
The owning server puts the data in its cache, and at the same moment the other server stores a copy in its NVS, so the data now sits in two separate boxes with two separate power supplies.
4The host hears “done” right away
Once both copies exist, the DS8000 tells the host the write is complete. The host never waits for a flash drive, which is why writes on a DS8000 finish at memory speed.
5Destage to flash, then drop the spare copy
A little later the owning server writes the data out (destages it) through a flash RAID adapter to an array of NVMe drives in an HPFE Gen3 enclosure. That array uses RAID 6, so even two failed drives lose nothing, and once the destage is finished the NVS copy is thrown away to make room for new writes.
Reads are easier. If the data is already in cache (a cache hit), the host gets it straight from memory; if not, the server reads it from flash, keeps a copy in cache for next time and sends it on to the host.
IBM’s cache rules (SARC, AMP and IWC) stop a big backup job from pushing everyone else’s hot data out of cache.
All this time the two HMCs are watching the hardware, but they sit completely outside the data path. You could unplug both and the servers would keep reading and writing; you’d only lose the GUI, the DS CLI and Call Home until an HMC came back.
If a whole CPC fails, the surviving server already holds a copy of the other server’s writes in its NVS, so it saves them to flash, takes over all the LSSs and keeps serving the hosts. This is failover, and Lesson 18 walks through it step by step.
Hands-on lab: meet your practice DS8000
Before you run any command, double-check it
Read every command twice before you press Enter. Check the target: the system or account, the IDs, the names and the sizes. These commands are for learning and were tested in our lab, not on your system. At work, run the read-only list commands first, follow your change process, make sure a backup exists, and ask a second person to check anything that changes or deletes data. You run them at your own risk.
You’ll use the SaatVedha practice shell in nearly every lab of this course. It’s a small Python program we wrote that acts like the DS CLI on a DS8A50 called ds8k-g10-lab: you type the same commands a storage admin types at work, and the shell remembers what you build.
Practice shell note: the practice shell is our teaching tool, not IBM software. Real DS CLI output looks very close to this but isn’t identical, and some message numbers are ours (we list them on the practice shell page). What you see in the boxes below is what the shell prints.
1Install Python 3
First check whether you already have Python: open PowerShell on Windows, or Terminal on macOS or Linux.
Run this:
python --versionYou should see something like this:
Python 3.13.2If you get an error, install Python 3 from python.org. On Windows, tick Add python.exe to PATH on the first installer screen and then open a new PowerShell window. On macOS and Linux the command is often python3, so use that wherever this course says python.
✅ Checkpoint: python --version shows 3.8 or newer.
2Get the practice shell
Download ds8sim.zip from the DS8000 practice shell page and unzip it into a new folder, C:\ds8k-lab on Windows or ~/ds8k-lab on macOS and Linux. It only uses Python’s standard library, so there’s nothing else to install.
Run this (Windows PowerShell):
cd C:\ds8k-lab
dirYou should see the file:
ds8sim.py✅ Checkpoint: ds8sim.py is in your lab folder.
3Log in for the first time
Start the shell. It asks the same four questions as the real DS CLI; press Enter to skip the secondary console.
Use admin as the user name and admin as the password, as on a brand-new DS8000, and then choose your own password with at least 8 characters, a letter and a number. The password won’t show on screen while you type.
Run this:
python ds8sim.pyYou should see this:
SaatVedha DS8000 practice shell — not IBM software. Real DS CLI output looks like this but is not identical.
Enter the primary management console IP address: 10.0.0.1
Enter the secondary management console IP address:
Enter your username: admin
Enter your password:
CMUN00247W Your password has expired. You must change it now.
Enter a new password:
Re-enter the new password:
CMUC00249I Your password was changed successfully.
dscli>The dscli> prompt means you’re in interactive mode, so you no longer type dscli in front of each command.
✅ Checkpoint: You see the dscli> prompt and your new password works.
4Find out which DS8000 you are talking to
At work the first thing you check is that you’re on the right system, and two commands tell you that.
Type these at the dscli> prompt:
ver -l
lssiYou should see this:
dscli> ver -l
Date/Time: 24 October 2026 10:34:14 UTC IBM DSCLI Version: 7.10.10.348 DS: IBM.2107-78SV001
DSCLI 7.10.10.348
StorageManager 10.2.2507181727
HMC DSCLI 7.10.10.348
==================Version==================
Storage Image LMC Bundle Version
===========================================
IBM.2107-78SV001 7.10.10.348 10.10.106.0
dscli> lssi
Date/Time: 24 October 2026 10:34:14 UTC IBM DSCLI Version: 7.10.10.348 DS: IBM.2107-78SV001
Name ID Storage Unit Model WWNN State ESSNet
====================================================================================
ds8k-g10-lab IBM.2107-78SV001 IBM.2107-78SV000 A05 5005076303FFF0DD Online EnabledIBM.2107-78SV001 is the storage image ID, which you’ll use in profiles and scripts, and model A05 means DS8A50. The WWNN is the system’s worldwide node name that SAN engineers see in their switches, and the bundle version is the code level, here Release 10.1. The Date/Time line on top is only the banner, so your date and time will be different.
✅ Checkpoint: You can name the model (A05 = DS8A50) and the storage image ID.
5Meet the two servers
Run this:
lsserver -lYou should see this:
dscli> lsserver -l
Date/Time: 24 October 2026 10:34:19 UTC IBM DSCLI Version: 7.10.10.348 DS: IBM.2107-78SV001
ID Image ID Image Name Power Control SFI State LIC Version OS Version Bundle Version
===================================================================================================
server0 IBM.2107-78SV001 ds8k-g10-lab Online Online 7.10.10.348 AIX 7.3 10.10.106.0
server1 IBM.2107-78SV001 ds8k-g10-lab Online Online 7.10.10.348 AIX 7.3 10.10.106.0There they are: server0 and server1, the storage partitions inside CPC 0 and CPC 1. Both are Online and on the same code level, and if one ever shows anything else, the other is carrying all the work.
✅ Checkpoint: Both servers show Online.
6Look at the two HMCs
Run this:
lshmcYou should see this:
dscli> lshmc
Date/Time: 24 October 2026 10:34:14 UTC IBM DSCLI Version: 7.10.10.348 DS: IBM.2107-78SV001
Name State Role Release Management-IP Location-Code Machine-Type Machine-Model Memory Disk
================================================================================================================================
DS8A50-1A.LAB Online Primary(1) R10.1 bundle 10.10.106.0 10.0.0.1 U7063CR2.78SV1DD 7063 CR2 33GB 1TB
DS8A50-1B.LAB Online Secondary(2) R10.1 bundle 10.10.106.0 10.0.0.2 U7063CR2.78SV2DD 7063 CR2 33GB 1TBYou logged in through HMC 1 at 10.0.0.1, and HMC 2 at 10.0.0.2 is the backup. At work you put both addresses in your DS CLI profile so you can still manage the system when one HMC is down.
✅ Checkpoint: You can tell which HMC is primary.
7Count the drives
Run this:
lsarraysite -lYou should see this:
dscli> lsarraysite -l
Date/Time: 24 October 2026 10:34:14 UTC IBM DSCLI Version: 7.10.10.348 DS: IBM.2107-78SV001
arsite DA Pair dkcap (10^9B) diskrpm State Array diskclass encrypt
===============================================================================
S1 0 4800.0 65000 Unassigned - FlashCoreModule supported
S2 0 4800.0 65000 Unassigned - FlashCoreModule supported
S3 0 4800.0 65000 Unassigned - FlashCoreModule supported
S4 0 4800.0 65000 Unassigned - FlashCoreModule supported
S5 2 4800.0 65000 Unassigned - FlashCoreModule supported
S6 2 4800.0 65000 Unassigned - FlashCoreModule supported
S7 2 3840.0 65000 Unassigned - FlashTier2 supported
S8 2 3840.0 65000 Unassigned - FlashTier2 supportedEach line is an array site, a group of 8 identical flash drives that IBM sets up at install time. This lab system has six sites of 4.8 TB FlashCore Modules and two sites of 3.84 TB industry-standard flash, so work out the raw capacity yourself: 6 × 8 × 4.8 TB = 230.4 TB, plus 2 × 8 × 3.84 TB = 61.44 TB, which comes to about 291.8 TB raw.
Every site still says Unassigned, meaning no RAID arrays exist yet. You’ll build them in Lesson 10.
✅ Checkpoint: You get 291.84 TB raw from the 8 array sites.
8Check the host ports, then leave
Run this:
lsioportYou should see this:
dscli> lsioport
Date/Time: 24 October 2026 10:34:14 UTC IBM DSCLI Version: 7.10.10.348 DS: IBM.2107-78SV001
ID WWPN State Type topo portgrp Security
===========================================================================================
I0010 50050763030130DD Offline Fibre Channel-LW - 0 Disabled
I0011 50050763030170DD Offline Fibre Channel-LW - 0 Disabled
I0012 500507630301B0DD Offline Fibre Channel-LW - 0 Disabled
I0013 500507630301F0DD Offline Fibre Channel-LW - 0 Disabled
I0030 50050763030330DD Communication established Fibre Channel-SW SCSI-FCP 0 Disabled
I0031 50050763030370DD Offline Fibre Channel-SW - 0 Disabled
I0032 500507630303B0DD Communication established Fibre Channel-SW SCSI-FCP 0 Disabled
I0033 500507630303F0DD No light detected Fibre Channel-SW SCSI-FCP 0 Disabled
I0040 50050763030430DD Offline Fibre Channel-SW - 0 Disabled
I0041 50050763030470DD Communication established Fibre Channel-SW FICON 0 Disabled
I0042 500507630304B0DD Communication established Fibre Channel-SW SCSI-FCP 0 Disabled
I0043 500507630304F0DD No light detected Fibre Channel-SW SCSI-FCP 0 Disabled
I0050 50050763030530DD Offline Fibre Channel-SW - 0 Disabled
I0051 50050763030570DD Offline Fibre Channel-SW - 0 Disabled
I0052 500507630305B0DD Communication established Fibre Channel-SW SCSI-FCP 0 Disabled
I0053 500507630305F0DD Offline Fibre Channel-SW - 0 DisabledEach I00xx line is one port on a host adapter. Communication established means a cable is plugged in and the switch answered, while No light detected means the port is set up but sees no signal, which points to a missing or broken cable. topo shows how the port is used: SCSI-FCP for open systems, FICON for mainframes and - if it isn’t set yet.
Then type:
quit✅ Checkpoint: You can point to one FCP port, one FICON port and one port with a cabling problem.
Break It, Then Fix It
Now break two things on purpose. These are the first two errors nearly every new DS8000 admin hits, usually when they’re in a hurry.
1Break it: log in with the wrong password
Start the shell again and type the old password admin on purpose.
Run this:
python ds8sim.pyYou’ll get this error:
SaatVedha DS8000 practice shell — not IBM software. Real DS CLI output looks like this but is not identical.
Enter the primary management console IP address: 10.0.0.1
Enter the secondary management console IP address:
Enter your username: admin
Enter your password:
CMUN00244E Authentication failure: the user name or password is not correct.Why: you changed the password in step 3, so admin no longer works. On a real DS8000 each failed login also counts against you, and after a set number of failures (shown by showpass, 5 on this lab system) the account locks until an administrator releases it with chuser -unlock.
People lock out the shared admin account this way during an outage, which is one reason teams give each person their own user (Lesson 5 of the Implementation course).
Fix: run python ds8sim.py again and type the password you chose. If you’ve forgotten it, delete ds8sim-state.json in your lab folder and start again with admin / admin.
✅ Checkpoint: You are back at the dscli> prompt.
2Break it: type a command that does not exist
Plural or singular? Everyone gets this wrong at least once.
Type this at the dscli> prompt:
lsarraysitesYou’ll get this error, then the fix:
dscli> lsarraysites
CMUN00014E lsarraysites: Unknown command. Enter "help" for a list of commands.
dscli> help -s lssi
SaatVedha DS8000 practice shell — not IBM software. Real DS CLI output looks like this but is not identical.
lssi Displays a list of storage images.
dscli> lsarraysite -l -diskclass FlashTier2
Date/Time: 24 October 2026 10:34:19 UTC IBM DSCLI Version: 7.10.10.348 DS: IBM.2107-78SV001
arsite DA Pair dkcap (10^9B) diskrpm State Array diskclass encrypt
==========================================================================
S7 2 3840.0 65000 Unassigned - FlashTier2 supported
S8 2 3840.0 65000 Unassigned - FlashTier2 supported
Why: DS CLI command names have to match exactly. The list commands are always ls plus a singular object (lsarraysite, lsarray, lsrank, lsextpool, lsfbvol), the same pattern holds for mk to make, ch to change and rm to remove, and show gives the details of one object.
Fix: when you’re not sure, ask the shell: help lists every command, help -s name gives a one-line description and name -h shows the full usage. The last command above also shows a handy filter, -diskclass, which lists only the array sites of one drive type.
✅ Checkpoint: You can list only the FlashTier2 array sites.
This happened at work
The night someone rebooted the wrong box
This was at a mid-sized bank that ran its card authorisation system on AIX servers with DS8000 storage. One Friday night the card system slowed right down: payments that normally took 200 ms were taking 4 seconds, and shops started phoning in.
A new engineer on the night shift remembered someone saying “the storage has two consoles at the bottom”, decided one of them was stuck and rebooted HMC 1, then HMC 2. Nothing improved, and the team lost the DS GUI for twenty minutes while the HMCs came back up. By then forty minutes had gone.
When the senior storage engineer joined the call, she looked at the performance data from the host side first and saw one AIX server pushing every I/O down a single path. A SAN switch port had gone bad earlier that evening, and the server’s second path had been missing since a cable change two weeks before, so the DS8000 had been healthy the whole time.
Moving the cable to a good switch port and bringing the second path back took ten minutes. After that night the team wrote two rules on the wall: the HMC isn’t in the data path, and every host must show two healthy paths after every change. If the night shift had known what each part of the DS8000 does, most of that hour would have been saved.
Common mistakes and troubleshooting
| Symptom | Likely cause | What to check / fix |
|---|---|---|
| “The storage is slow, reboot the HMC” | Thinking the HMC carries data | The HMC only manages and monitors, so look at host paths, port errors and performance data instead. |
| A host loses all its disks when one cable is pulled | Only one path, or both paths on the same host adapter | Zone every host to two host adapters in two different I/O enclosures. |
| All busy volumes sit on one server | Volumes created only in even LSSs (or only one pool) | Spread volumes over even and odd LSSs, so both CPCs share the work. |
| Confusion about “server”, “CPC”, “node” | Several names for the same thing | CPC 0/1 is the Power box; server0/server1 (node 0/1) is the storage partition running inside it. |
| Someone plans a RAID 5 array | Habit from older arrays | DS8A00 supports RAID 6 (default) and RAID 10 only. |
| DS CLI says it can’t connect | Wrong HMC address or HMC down | Use the second HMC (hmc2 in your profile). Hosts are not affected by an HMC outage. |
One habit helps every time: before you change anything, ask yourself which server owns it and what else depends on it. That question explains most of how a DS8000 behaves.
Interview questions
Click a question to open it. Each is worded the way an interviewer would ask it, and underneath you’ll find why they ask it and one way you could answer in your own words.
Fresher level
1. Can you explain in simple words what an IBM DS8000 is?
Why they ask this: They want to hear that you see it as a complete storage system and not a box of disks, and explaining it in plain words shows you understand it.
How you might answer: I’d call it an enterprise storage system that gives disks to lots of servers over Fibre Channel or FICON. Inside there are two Power servers sharing the work, a large cache with a protected copy of every write and NVMe flash drives in RAID 6, and if one part fails another takes over, so the servers using it keep running.
2. What does the HMC do on a DS8000?
Why they ask this: They’re checking whether you can tell the management path from the data path, which a lot of beginners mix up.
How you might answer: The HMC is the management console. I connect to it with the DS GUI or the DS CLI to configure and monitor the box, and it sends Call Home messages to IBM when a part fails. It’s not in the data path, so host I/O carries on even if both HMCs are down.
2–5 years experience
3. A host writes a block to a DS8000. Walk me through what happens until the host gets the acknowledgement.
Why they ask this: It shows whether you understand cache, NVS and why writes are fast, and it comes up in nearly every storage interview.
How you might answer: The write arrives on a host adapter port and goes over PCIe to the server that owns the volume’s LSS, even LSSs to server0 and odd to server1. That server puts it in cache while the other one keeps a copy in its NVS, and as soon as both copies exist the host gets “complete”. The destage to the RAID 6 array on flash happens later, and then the NVS copy is dropped.
4. Your manager asks why the bank should pay for a DS8000 when cheaper arrays exist. What would you say?
Why they ask this: They want to see whether you can link technical features to business value, because sooner or later you’ll have to explain things to managers.
How you might answer: I’d say the DS8000 is made for systems that can’t stop, like card payments and core banking. Its two active servers fail over without the hosts noticing, the write cache is protected, the flash is RAID 6 and the replication with Metro Mirror and Global Mirror is very strong. It’s also one of the few arrays that supports IBM Z with FICON and zHyperLink, so mainframe shops often don’t have a real alternative.
Senior level
5. How would you size the first DS8000 for a company that runs both IBM Z and VMware?
Why they ask this: Senior people have to think about mixed workloads, growth and keeping both servers busy, so they want a structured approach rather than a model number.
How you might answer: I’d begin with workload data from both sides: capacity, IOPS, read/write mix and growth over three to five years. From that I’d pick the model by cores and cache, say a DS8A50 with enough memory, and plan the FICON and FCP host adapter ports separately. I’d also plan separate extent pool pairs for CKD and FB so each workload is spread over both servers, and I’d run the numbers through IBM’s modelling tool before we ordered anything.
6. What would you check first if someone says “the DS8000 is down”?
Why they ask this: They want calm, step-by-step thinking, and they want to know you understand what “down” can mean on a system with no single point of failure.
How you might answer: First I’d find out what’s failing: one host, one application or everything. I’d log in through the DS CLI or GUI on either HMC and look at lsserver, the open problems and the events. Usually the array is fine and the problem is a path, a zone or a host. If a CPC has failed, the other one has already taken over, so I’d open a case with IBM and keep an eye on the surviving server.
Quick quiz
Q1. How many CPCs (Power servers) does every DS8A00 have?
- One
- Two
- Four
- It depends on the model
Show answer
B — Every DS8A00 has two CPCs, CPC 0 and CPC 1, which share the work and can take over from each other.
Q2. Where is the second copy of a new write kept before it reaches flash?
- On the HMC
- In the NVS of the other CPC
- On a spare flash drive
- In the host’s memory
Show answer
B — The owning server keeps the write in cache and the other server keeps a copy in its NVS.
Q3. Which models belong to the DS8000 G10 family?
- DS8880 and DS8900F
- DS8A10, DS8A50 and DS8A80
- DS8910F only
- FlashSystem 9500
Show answer
B — The G10 generation is the DS8A00 family: DS8A10 (Flexibility), DS8A50 (Agility) and DS8A80 (Analytic).
Q4. What happens to host I/O if both HMCs go offline?
- All I/O stops
- Only reads stop
- I/O continues; you lose management until an HMC is back
- The CPCs fail over
Show answer
C — The HMC is only for management and monitoring. It is not in the data path.
Q5. A volume in LSS 21 is owned by which server?
- server0
- server1
- Both at the same time
- The HMC
Show answer
B — Odd LSSs belong to server1 in CPC 1; even LSSs belong to server0 in CPC 0.
Quick recap
- A DS8000 is an enterprise storage system with two Power9+ CPCs, a large cache with NVS, host adapters and HPFE Gen3 flash.
- The G10 models are DS8A10, DS8A50 and DS8A80, all-flash, machine type 5341.
- A write lands in the owning server’s cache and the other server’s NVS, and then the host hears “done”; the destage to flash comes later.
- Even LSSs belong to server0, odd LSSs to server1.
- The two HMCs manage and monitor the system; they are not in the data path.
- With the practice shell you can run real DS CLI commands like lssi, lsserver and lsarraysite on your own laptop.
Next lesson
Lesson 2: DS8000 models: DS8A10, DS8A50, DS8A80 and the family tree → You compare the three models and see how the DS8000 grew from DS8880 to DS8A00.
Further reading: IBM DS8000 storage
To go deeper into IBM DS8000 storage, IBM’s own guide covers every part of this family: IBM Storage DS8A00 Architecture and Implementation Guide (IBM Redbooks). And if SANs are new to you, our SAN Storage course explains hosts, HBAs and LUNs from the start.
You can also see the full plan for this course on the IBM DS8000 Architecture course page.