iSCSI on Linux: open-iscsi and dm-multipath
Storage people often stop at “the LUN is mapped”. The ticket comes back an hour later because the Linux team can’t see the disk, or worse, the disk vanishes after the next reboot.
This lesson is the host side of iSCSI. The practice shell plays both sides: a two-node ONTAP cluster in one terminal, and the Ubuntu server linux1 in a second terminal. Every command and output below comes from that simulator, and they’re the same commands you’d type on a real Ubuntu or Rocky Linux server.
Where this lesson fits
Lesson 25 of 56 in NetApp ONTAP Storage Administration · Phase 2: Protocols and day-to-day administration · Before this: Lesson 24, ALUA and multipath I/O · Next: Lesson 26, iSCSI on Windows Server
| Time | You need | By the end you can |
|---|---|---|
| 75 minutes | The practice shell (version 1.2 or later) in two terminals | Discover and log in to an iSCSI target, make a file system that mounts after reboot, and read sessions on both sides |
What it is, in one line
On Linux, the open-iscsi package logs the server in to an iSCSI target over TCP port 3260, the kernel turns each LUN path into an sd device, and dm-multipath joins those paths into one disk you can format and mount.
In Simple Words
Dialling into a conference call
Think of a weekly conference call. First you look up the number (discovery). Then you dial in and say your name, and the host lets you in only if your name is on the invite list (login and the igroup).
If you dial in from your desk phone and your mobile, you’re on the call twice; your headset just picks the clearer line. That’s multipath.
And if you don’t save the number in your phone’s auto-dial, you’ll miss next week’s call. On Linux that’s the node.startup setting.
Engineer's View
How an engineer describes it
The initiator name lives in /etc/iscsi/initiatorname.iscsi. iscsiadm -m discovery -t sendtargets -p IP asks one portal for every target and portal it offers and saves them as node records under /etc/iscsi/nodes. iscsiadm -m node --login opens a session per portal.
Each record has node.startup: automatic logs in at boot, manual doesn’t. RHEL-family systems default to automatic in /etc/iscsi/iscsid.conf; Ubuntu ships manual, which is why Ubuntu disks often “disappear” after a reboot.
dm-multipath builds /dev/mapper/... devices from paths with the same WWID. NetApp’s settings are built into multipath-tools, so you only need a basic /etc/multipath.conf.
In /etc/fstab, mark iSCSI file systems _netdev so they mount after the network is up, and nofail so a missing disk doesn’t stop the boot. NetApp’s Linux Host Utilities add the optional sanlun tool.
How it works, step by step
1Install the tools and read the IQN
open-iscsi and multipath-tools (Ubuntu) or iscsi-initiator-utils and device-mapper-multipath (RHEL, Rocky). Give the IQN to the storage team for the igroup.
2Discover and log in
Discovery from one LIF returns every iSCSI LIF of the SVM. Log in to all of them so you get a path through each node.
3Rescan and check multipath
New LUNs show up as sd devices, one per path, and multipath groups them by WWID. multipath -ll should show two priority groups for an ONTAP LUN.
4Make it permanent
Set node.startup to automatic, enable the iSCSI and multipath services, and mount with _netdev,nofail. Then reboot once on purpose, during the build, so you know it comes back.
| Task | Ubuntu / Debian | RHEL / Rocky / Alma |
|---|---|---|
| Packages | open-iscsi, multipath-tools | iscsi-initiator-utils, device-mapper-multipath |
| Default node.startup | manual | automatic |
| Boot login service | open-iscsi | iscsi |
| Create multipath.conf | sudo touch /etc/multipath.conf (then edit) | sudo mpathconf --enable |
Hands-on lab: connect linux1 over iSCSI with multipath
Before you run any command, double-check it
Read every command twice before you press Enter. Check the target: the server name, the user, the disk or device name, the path and the sizes. These commands are for learning and were tested in our lab, not on your system. At work, run the read-only 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.
This lab uses the SaatVedha ONTAP practice shell, our own learning tool. It is not NetApp software: real ONTAP output looks like this but is not identical.
You’ll use two terminal windows.
The first runs the cluster shell (python ontapsim.py). The second runs a client host: python ontapsim.py --host linux1 opens a root prompt on linux1, an Ubuntu 24.04 server at 192.168.56.21. Both read the same lab file, so what you map on the cluster shows up on the host.
1Build the storage side
In C:\ontap-lab, save these ten lines as setup-25.txt. They make two aggregates, an SVM with one iSCSI LIF per node, and a 10 GB LUN mapped to linux1’s IQN, which is everything Lessons 23 and 24 taught.
storage aggregate create -aggregate aggr1_01 -node saat-cluster-01 -diskcount 8 -disktype SAS
storage aggregate create -aggregate aggr1_02 -node saat-cluster-02 -diskcount 8 -disktype SAS
vserver create -vserver svm1 -aggregate aggr1_01 -rootvolume-security-style unix
network interface create -vserver svm1 -lif svm1_iscsi1 -service-policy default-data-iscsi -home-node saat-cluster-01 -home-port e0d -address 192.168.56.131 -netmask 255.255.255.0
network interface create -vserver svm1 -lif svm1_iscsi2 -service-policy default-data-iscsi -home-node saat-cluster-02 -home-port e0d -address 192.168.56.132 -netmask 255.255.255.0
vserver iscsi create -vserver svm1
volume create -vserver svm1 -volume vol_web1 -aggregate aggr1_01 -size 12g -space-guarantee none -snapshot-policy none
lun create -vserver svm1 -path /vol/vol_web1/lun_web1 -size 10g -ostype linux -space-reserve disabled
lun igroup create -vserver svm1 -igroup ig_linux1 -protocol iscsi -ostype linux -initiator iqn.2004-10.com.ubuntu:01:5a7e1a2b3c4d
lun mapping create -vserver svm1 -path /vol/vol_web1/lun_web1 -igroup ig_linux1Reset the lab and run the file from your normal terminal:
python ontapsim.py -c "simctl reset"
python ontapsim.py --run setup-25.txtEach line prints its result. You should see no line that starts with Error.
✅ Checkpoint: The run ends with the lun mapping create line and no errors.
2Open linux1 and read its IQN
Open a second terminal in the same folder and start the host:
python ontapsim.py --host linux1You’re now root on linux1. On a real server you’d use sudo; the practice host logs you in as root so you can skip it. Check the initiator name and that the iSCSI daemon is running:
cat /etc/iscsi/initiatorname.iscsi
systemctl status iscsidYou should see this (times will differ):
root@linux1:~# cat /etc/iscsi/initiatorname.iscsi
## DO NOT EDIT OR REMOVE THIS FILE!
## If you remove this file, the iSCSI daemon will not start.
## If you change the InitiatorName, existing access control lists
## may reject this initiator. The InitiatorName must be unique
## for each iSCSI initiator. Do NOT duplicate iSCSI InitiatorNames.
InitiatorName=iqn.2004-10.com.ubuntu:01:5a7e1a2b3c4d
root@linux1:~# systemctl status iscsid
● iscsid.service - iSCSI initiator daemon (iscsid)
Loaded: loaded (/usr/lib/systemd/system/iscsid.service; enabled; preset: enabled)
Active: active (running) since Sat 2026-11-07 11:24:19 IST; 37min ago
TriggeredBy: ● iscsid.socket
Docs: man:iscsid(8)
Main PID: 1639 (iscsid)
Tasks: 2 (limit: 4613)
Memory: 3.7M (peak: 4.5M)
CPU: 58ms
CGroup: /system.slice/iscsid.service
└─1639 /usr/sbin/iscsidThe IQN on the last line of the file must match the igroup exactly, character for character. On a fresh Ubuntu install you’d first run apt install -y open-iscsi multipath-tools; the practice linux1 already has both.
✅ Checkpoint: The InitiatorName ends in 5a7e1a2b3c4d and iscsid is active (running).
3Discover the target and log in to every portal
iscsiadm -m discovery -t sendtargets -p 192.168.56.131
iscsiadm -m node --login
iscsiadm -m sessionYou should see this:
root@linux1:~# iscsiadm -m discovery -t sendtargets -p 192.168.56.131
192.168.56.131:3260,1033 iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3
192.168.56.132:3260,1034 iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3
root@linux1:~# iscsiadm -m node --login
Logging in to [iface: default, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.131,3260]
Logging in to [iface: default, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.132,3260]
Login to [iface: default, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.131,3260] successful.
Login to [iface: default, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.132,3260] successful.
root@linux1:~# iscsiadm -m session
tcp: [1] 192.168.56.131:3260,1033 iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3 (non-flash)
tcp: [2] 192.168.56.132:3260,1034 iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3 (non-flash)You asked one LIF, and it answered with both LIFs of the SVM, one per node. --login without a portal logs in to every discovered portal, so you get two sessions.
✅ Checkpoint: iscsiadm -m session lists two sessions, one to .131 and one to .132.
4Check the disks and multipath
lsblk
multipath -llYou should see this:
root@linux1:~# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS
sda 8:0 0 25G 0 disk
├─sda1 8:1 0 1M 0 part
├─sda2 8:2 0 2G 0 part /boot
└─sda3 8:3 0 23G 0 part
└─ubuntu--vg-ubuntu--lv 252:0 0 11.5G 0 lvm /
sdb 8:16 0 10G 0 disk
└─3600a09806e7553666d67586b74577143 252:1 0 10G 0 mpath
sdc 8:32 0 10G 0 disk
└─3600a09806e7553666d67586b74577143 252:1 0 10G 0 mpath
sr0 11:0 1 1024M 0 rom
root@linux1:~# multipath -ll
3600a09806e7553666d67586b74577143 dm-1 NETAPP,LUN C-Mode
size=10G features='3 queue_if_no_path pg_init_retries 50' hwhandler='1 alua' wp=rw
|-+- policy='service-time 0' prio=50 status=active
| `- 30:0:0:0 sdb 8:16 active ready running
`-+- policy='service-time 0' prio=10 status=enabled
`- 31:0:0:0 sdc 8:32 active ready runningsdb and sdc are the same 10 GB LUN, seen once through each session. dm-multipath joined them into one device named after the LUN’s WWID. The path with prio=50 goes through the LUN’s own node (active/optimized); prio=10 goes through the partner, exactly the ALUA picture from Lesson 24.
✅ Checkpoint: multipath -ll shows one NETAPP device with two paths: prio=50 active and prio=10 enabled.
5See the same login from ONTAP
Back in the first terminal, in the cluster shell:
vserver iscsi session show -vserver svm1
lun igroup show -vserver svm1 -igroup ig_linux1 -instanceYou should see this:
saat-cluster::> vserver iscsi session show -vserver svm1
Tpgroup Initiator Initiator
Vserver Name TSIH Name Alias
---------- ------------- ---- ---------------------------------------------- ----------
svm1 svm1_iscsi1 1 iqn.2004-10.com.ubuntu:01:5a7e1a2b3c4d linux1
svm1 svm1_iscsi2 2 iqn.2004-10.com.ubuntu:01:5a7e1a2b3c4d linux1
2 entries were displayed.
saat-cluster::> lun igroup show -vserver svm1 -igroup ig_linux1 -instance
Vserver Name: svm1
Igroup Name: ig_linux1
Protocol: iscsi
OS Type: linux
Portset Binding Igroup: -
Igroup UUID: 9cb5f1f0-3cfe-5eb6-6aab-dd1205720888
Initiators: iqn.2004-10.com.ubuntu:01:5a7e1a2b3c4d (logged in)One session per LIF, and the igroup now says (logged in). When a server team says “we can’t see the disk”, these two commands tell you in seconds whether the host ever logged in.
✅ Checkpoint: Two sessions and the initiator shows (logged in).
6Make a file system that survives a reboot
On linux1. Always use the /dev/mapper name, never sdb or sdc: those names can change and they bypass multipath. Use the WWID from your own multipath -ll output.
mkfs.xfs -L webdata /dev/mapper/3600a09806e7553666d67586b74577143
mkdir -p /mnt/web
echo "/dev/mapper/3600a09806e7553666d67586b74577143 /mnt/web xfs _netdev,nofail 0 0" >> /etc/fstab
mount -a
df -h /mnt/web
findmnt /mnt/web
iscsiadm -m node --op update -n node.startup -v automatic
systemctl enable iscsid open-iscsiYou should see this:
root@linux1:~# mkfs.xfs -L webdata /dev/mapper/3600a09806e7553666d67586b74577143
Discarding blocks...Done.
meta-data=/dev/mapper/3600a09806e7553666d67586b74577143 isize=512 agcount=4, agsize=655360 blks
= sectsz=4096 attr=2, projid32bit=1
= crc=1 finobt=1, sparse=1, rmapbt=1
= reflink=1 bigtime=1 inobtcount=1 nrext64=1
data = bsize=4096 blocks=2621440, imaxpct=25
= sunit=0 swidth=0 blks
naming =version 2 bsize=4096 ascii-ci=0, ftype=1
log =internal log bsize=4096 blocks=16384, version=2
= sectsz=4096 sunit=1 blks, lazy-count=1
realtime =none extsz=4096 blocks=0, rtextents=0
root@linux1:~# mkdir -p /mnt/web
root@linux1:~# echo "/dev/mapper/3600a09806e7553666d67586b74577143 /mnt/web xfs _netdev,nofail 0 0" >> /etc/fstab
root@linux1:~# mount -a
root@linux1:~# df -h /mnt/web
Filesystem Size Used Avail Use% Mounted on
/dev/mapper/3600a09806e7553666d67586b74577143 10G 118M 9.9G 2% /mnt/web
root@linux1:~# findmnt /mnt/web
TARGET SOURCE FSTYPE OPTIONS
/mnt/web /dev/mapper/3600a09806e7553666d67586b74577143 xfs rw,relatime,attr2,inode64,logbufs=8,logbsize=32k,noquota
root@linux1:~# iscsiadm -m node --op update -n node.startup -v automatic
root@linux1:~# systemctl enable iscsid open-iscsi_netdev tells systemd to wait for the network before mounting, and nofail lets the server boot even if the storage is missing. node.startup automatic makes the sessions log in again at boot; on Ubuntu the default is manual, which is the classic “it vanished after the reboot” ticket.
✅ Checkpoint: df -h /mnt/web shows a 10G XFS file system on the /dev/mapper device.
7Optional: the same LUN on Rocky Linux with NetApp Host Utilities
Red Hat style servers use dnf and mpathconf, and NetApp’s Linux Host Utilities add sanlun, which maps every Linux device back to its ONTAP LUN and LIF. First give rhel1 its own LUN in the cluster shell (never share one LUN between two servers without a cluster file system):
saat-cluster::> lun create -vserver svm1 -path /vol/vol_web1/lun_db -size 5g -ostype linux -space-reserve disabled
Created a LUN of size 5GB (5368709120)
saat-cluster::> lun igroup create -vserver svm1 -igroup ig_rhel1 -protocol iscsi -ostype linux -initiator iqn.1994-05.com.redhat:8f3d2c1b9a7e
saat-cluster::> lun mapping create -vserver svm1 -path /vol/vol_web1/lun_db -igroup ig_rhel1Then open a third terminal with python ontapsim.py --host rhel1 and run:
dnf install -y iscsi-initiator-utils device-mapper-multipath
mpathconf --enable --with_multipathd y
systemctl enable --now iscsid
iscsiadm -m discovery -t st -p 192.168.56.131
iscsiadm -m node -L all
rpm -ivh netapp_linux_unified_host_utilities-8-0.x86_64.rpm
sanlun lun show -pYou should see this:
[root@rhel1 ~]# dnf install -y iscsi-initiator-utils device-mapper-multipath
(dnf output trimmed: 5 packages installed)
Complete!
[root@rhel1 ~]# mpathconf --enable --with_multipathd y
[root@rhel1 ~]# systemctl enable --now iscsid
Created symlink /etc/systemd/system/multi-user.target.wants/iscsid.service → /usr/lib/systemd/system/iscsid.service.
[root@rhel1 ~]# iscsiadm -m discovery -t st -p 192.168.56.131
192.168.56.131:3260,1033 iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3
192.168.56.132:3260,1034 iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3
[root@rhel1 ~]# iscsiadm -m node -L all
Logging in to [iface: default, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.131,3260]
Logging in to [iface: default, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.132,3260]
Login to [iface: default, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.131,3260] successful.
Login to [iface: default, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.132,3260] successful.
[root@rhel1 ~]# rpm -ivh netapp_linux_unified_host_utilities-8-0.x86_64.rpm
Verifying... ################################# [100%]
Preparing... ################################# [100%]
Updating / installing...
1:netapp_linux_unified_host_utiliti################################# [100%]
[root@rhel1 ~]# sanlun lun show -p
ONTAP Path: svm1:/vol/vol_web1/lun_db
LUN: 0
LUN Size: 5g
Product: cDOT
Host Device: mpatha
Multipath Policy: service-time 0
Multipath Provider: Native
--------- ---------- ------- ------------ ----------------------------------------------
host vserver
path path /dev/ host vserver
state type node adapter LIF
--------- ---------- ------- ------------ ----------------------------------------------
up primary sdb host30 svm1_iscsi1
up secondary sdc host31 svm1_iscsi2sanlun lun show -p is the command support engineers ask for first: it shows the ONTAP path, the multipath device and which LIF each path uses, with primary meaning the optimized path. On a real server you download the Host Utilities rpm from the NetApp support site.
✅ Checkpoint: sanlun shows two paths that are up: primary through svm1_iscsi1 and secondary through svm1_iscsi2.
Try it yourself: Build the ONTAP side of an iSCSI LUN in the Command Builder.
Break It, Then Fix It
The most common iSCSI ticket of all: “the host logs in, but there’s no disk”. Nine times out of ten the IQN on the host doesn’t match the igroup, usually after someone edited the file, cloned a VM or rebuilt the server.
1Break the initiator name
On linux1, unmount the file system, let simctl put a one-character typo in the IQN, then log out and in again the way a reboot would:
umount /mnt/web
simctl break host-iqn-wrong linux1
iscsiadm -m node -U all
iscsiadm -m node -L all
lsblk
cat /etc/iscsi/initiatorname.iscsiYou’ll get this:
root@linux1:~# umount /mnt/web
root@linux1:~# simctl break host-iqn-wrong linux1
simctl: fault "host-iqn-wrong" is active on linux1. Now find it from the host and the cluster.
root@linux1:~# iscsiadm -m node -U all
Logging out of session [sid: 1, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.131,3260]
Logging out of session [sid: 2, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.132,3260]
Logout of [sid: 1, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.131,3260] successful.
Logout of [sid: 2, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.132,3260] successful.
root@linux1:~# iscsiadm -m node -L all
Logging in to [iface: default, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.131,3260]
Logging in to [iface: default, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.132,3260]
Login to [iface: default, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.131,3260] successful.
Login to [iface: default, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.132,3260] successful.
root@linux1:~# lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS
sda 8:0 0 25G 0 disk
├─sda1 8:1 0 1M 0 part
├─sda2 8:2 0 2G 0 part /boot
└─sda3 8:3 0 23G 0 part
└─ubuntu--vg-ubuntu--lv 252:0 0 11.5G 0 lvm /
sr0 11:0 1 1024M 0 rom
root@linux1:~# cat /etc/iscsi/initiatorname.iscsi
## DO NOT EDIT OR REMOVE THIS FILE!
## If you remove this file, the iSCSI daemon will not start.
## If you change the InitiatorName, existing access control lists
## may reject this initiator. The InitiatorName must be unique
## for each iSCSI initiator. Do NOT duplicate iSCSI InitiatorNames.
InitiatorName=iqn.2004-10.com.ubuntu:01:5a7e1a2b3cxdThe logins succeed, yet lsblk shows no sdb, sdc or mpath device. iSCSI login only proves the network path; the igroup decides which LUNs the host may see.
✅ Checkpoint: Logins succeed but lsblk shows no NETAPP disk.
2Find it from the storage side
In the cluster shell:
lun igroup show -vserver svm1 -igroup ig_linux1 -instance
vserver iscsi initiator show -vserver svm1You’ll get this:
saat-cluster::> lun igroup show -vserver svm1 -igroup ig_linux1 -instance
Vserver Name: svm1
Igroup Name: ig_linux1
Protocol: iscsi
OS Type: linux
Portset Binding Igroup: -
Igroup UUID: 9cb5f1f0-3cfe-5eb6-6aab-dd1205720888
Initiators: iqn.2004-10.com.ubuntu:01:5a7e1a2b3c4d (not logged in)
saat-cluster::> vserver iscsi initiator show -vserver svm1
Tpgroup Initiator Igroup
Vserver Name Name Name
---------- ------------- ---------------------------------------------- ------------
svm1 svm1_iscsi1 iqn.2004-10.com.ubuntu:01:5a7e1a2b3cxd -
svm1 svm1_iscsi2 iqn.2004-10.com.ubuntu:01:5a7e1a2b3cxd -
svm1 svm1_iscsi1 iqn.1994-05.com.redhat:8f3d2c1b9a7e ig_rhel1
svm1 svm1_iscsi2 iqn.1994-05.com.redhat:8f3d2c1b9a7e ig_rhel1
4 entries were displayed.The igroup’s IQN (...3c4d) is not logged in, while an unknown initiator (...3cxd) is logged in with no igroup. Put the two strings side by side and the typo jumps out.
✅ Checkpoint: You can point at the one character that differs.
3Fix it on the host
Fix the file the host reads, not the igroup: the igroup holds the name the server team gave you. Then restart iscsid and log in again:
sed -i 's/3cxd/3c4d/' /etc/iscsi/initiatorname.iscsi
systemctl restart iscsid
iscsiadm -m node -U all
iscsiadm -m node -L all
multipath -ll
mount -a
df -h /mnt/web
simctl fix host-iqn-wrongYou should see this:
root@linux1:~# sed -i 's/3cxd/3c4d/' /etc/iscsi/initiatorname.iscsi
root@linux1:~# systemctl restart iscsid
root@linux1:~# iscsiadm -m node -U all
Logging out of session [sid: 1, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.131,3260]
Logging out of session [sid: 2, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.132,3260]
Logout of [sid: 1, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.131,3260] successful.
Logout of [sid: 2, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.132,3260] successful.
root@linux1:~# iscsiadm -m node -L all
Logging in to [iface: default, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.131,3260]
Logging in to [iface: default, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.132,3260]
Login to [iface: default, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.131,3260] successful.
Login to [iface: default, target: iqn.1992-08.com.netapp:sn.c84b2d40ac3f32ca940fcb7e13e6d933:vs.3, portal: 192.168.56.132,3260] successful.
root@linux1:~# multipath -ll
3600a09806e7553666d67586b74577143 dm-3 NETAPP,LUN C-Mode
size=10G features='3 queue_if_no_path pg_init_retries 50' hwhandler='1 alua' wp=rw
|-+- policy='service-time 0' prio=50 status=active
| `- 30:0:0:0 sdb 8:16 active ready running
`-+- policy='service-time 0' prio=10 status=enabled
`- 31:0:0:0 sdc 8:32 active ready running
root@linux1:~# mount -a
root@linux1:~# df -h /mnt/web
Filesystem Size Used Avail Use% Mounted on
/dev/mapper/3600a09806e7553666d67586b74577143 10G 118M 9.9G 2% /mnt/web
root@linux1:~# simctl fix host-iqn-wrong
simctl: fixed host-iqn-wrong.The disk is back with both paths and the file system mounts from fstab. The last line just tidies the lab.
✅ Checkpoint: multipath -ll shows two paths again and /mnt/web is mounted.
If the host’s IQN is the one that’s right and the igroup has the typo, fix the igroup instead with lun igroup add and lun igroup remove. Either way, compare both sides before you change anything.
This happened at work
Patch night in Mangaluru
A cooperative bank’s back office in Mangaluru ran its overnight batch jobs on a mix of RHEL and Ubuntu servers, all with LUNs from ONTAP over iSCSI. The Ubuntu servers had been built by a new contractor a few months earlier, and everything had worked since.
Then came the quarterly patch night. The RHEL servers rebooted and carried on.
Two Ubuntu servers came up without their data disks, and a third sat at an emergency mode prompt because its fstab line had no nofail. The batch job that sends files to the clearing house started 40 minutes late.
The storage team’s first check showed the initiators “not logged in”, so they knew it was host side within minutes. The node records were on manual startup, Ubuntu’s default. The team fixed iscsid.conf on every Ubuntu host, added _netdev,nofail to the fstab lines, and made “reboot once and check the disk” a sign-off step for every new SAN host.
Common mistakes and troubleshooting
| Mistake | What happens | Do this instead |
|---|---|---|
| Leaving node.startup on manual (Ubuntu default) | Disks vanish after a reboot | Set automatic in iscsid.conf or on each node record |
| No _netdev in fstab | The mount runs before the network is up and fails | Always use _netdev for iSCSI file systems |
| No nofail in fstab | A missing LUN can stop the boot in emergency mode | Add nofail so the server still boots |
| Logging in to only one portal | One path, no protection if a node or port fails | Log in to all discovered portals; check two prio groups |
| Mounting /dev/sdb directly on a multipath host | The device name changes, or you bypass multipath | Mount /dev/mapper devices or by LABEL/UUID |
| Cloning a VM with its initiatorname file | Two hosts share an IQN and fight over sessions | Generate a new IQN on every clone before first login |
Interview questions
Click a question to open it. Each one is written the way an interviewer would ask it. Under it you will see why they ask it and one way you could answer, in your own words.
Fresher level
1. Where do you find the iSCSI initiator name on a Linux server?
Why they ask this: They want to be sure you can collect the one value the storage team always needs.
How you might answer: It’s in /etc/iscsi/initiatorname.iscsi, on the InitiatorName line. I send that exact string to the storage team, copied rather than retyped, so the igroup matches.
2. Walk me through connecting a Linux server to a new iSCSI LUN.
Why they ask this: They’re checking you know the steps in order and don’t skip the parts that matter after a reboot.
How you might answer: I install open-iscsi and multipath-tools, give the IQN to storage, then run iscsiadm -m discovery -t sendtargets against one LIF and log in to all portals. After a rescan I check multipath -ll, make a file system on the mapper device, and add it to fstab with _netdev,nofail.
Last, I set node.startup to automatic and reboot once to prove it comes back.
2–5 years experience
3. After a reboot an Ubuntu server has lost its iSCSI disk, but the storage side looks fine. What do you check?
Why they ask this: They want to see that you know the most common host-side cause.
How you might answer: First iscsiadm -m session. If there are no sessions, I look at node.startup in the node records, because Ubuntu defaults to manual. I’d set it to automatic, log in, mount, and check the fstab line has _netdev and nofail.
4. Why would you use multipath even when the host has only one network card?
Why they ask this: They’re testing whether you understand that paths come from the target side too.
How you might answer: An ONTAP SVM has LIFs on both nodes, so even one host NIC gives one path per node. Without multipath the host would see two sd devices for one LUN, and if the owning node fails it loses the disk. dm-multipath joins them and handles failover and ALUA.
Senior level
5. You’re writing the Linux SAN build standard for 200 servers. What goes in it?
Why they ask this: They want to see that you think about repeatability, automation and checks, not only commands.
How you might answer: Packages, a standard iscsid.conf with automatic startup, a basic multipath.conf, and fstab rules with labels or UUIDs plus _netdev,nofail. I’d put it in Ansible so every server gets the same config.
Each build ends with proof: two sessions, two priority groups, and a test reboot recorded in the ticket.
6. A team clones Linux VMs from a template that already has iSCSI configured. What risk do you raise?
Why they ask this: They’re checking awareness of identity problems in cloned systems.
How you might answer: Every clone keeps the same IQN, so ONTAP sees several hosts with one name. Sessions get kicked, and an igroup meant for one server lets all clones see the LUNs. I’d make the template generate a fresh IQN on first boot and keep discovery and node records out of the template.
Quick quiz
Q1. Which file holds a Linux host’s iSCSI initiator name?
- /etc/iscsi/iscsid.conf
- /etc/iscsi/initiatorname.iscsi
- /etc/multipath.conf
- /etc/fstab
Show answer
B — the InitiatorName line in /etc/iscsi/initiatorname.iscsi.
Q2. What does iscsiadm -m discovery -t sendtargets -p IP do?
- Logs in to the target
- Asks one portal for all targets and portals and saves them
- Creates a LUN
- Formats the disk
Show answer
B — discovery lists and saves the targets and portals; login is a separate step.
Q3. Why do iSCSI disks often vanish after reboot on Ubuntu?
- Ubuntu can’t use iSCSI
- node.startup defaults to manual
- multipath is missing
- The LUN goes offline
Show answer
B — Ubuntu ships node.startup = manual, so it doesn’t log in at boot.
Q4. Which fstab options suit an iSCSI file system?
- defaults only
- _netdev,nofail
- noauto,ro
- sync,dirsync
Show answer
B — _netdev waits for the network and nofail keeps a missing disk from stopping the boot.
Q5. On ONTAP, which command lists the iSCSI sessions from hosts?
- lun show
- vserver iscsi session show
- network port show
- volume show
Show answer
B — vserver iscsi session show lists sessions per LIF.
Quick recap
- Give storage the IQN from /etc/iscsi/initiatorname.iscsi.
- Discover from one LIF, log in to every portal, rescan.
- Use dm-multipath and mount mapper devices or labels, not raw sd names.
- Set node.startup to automatic and mount with _netdev,nofail.
- Reboot once during the build; check vserver iscsi session show if a disk goes missing.
Next lesson
Lesson 26: iSCSI on Windows Server → You’ll connect a Windows Server VM with the iSCSI Initiator and MPIO, and bring a new disk online safely.
Further reading: ONTAP iSCSI Linux
Want to go deeper into ONTAP iSCSI Linux? Read NetApp’s ONTAP concepts guide for the SAN side, and your distribution’s open-iscsi and multipath documentation for host details. You can also see the full plan for this course on the NetApp ONTAP Storage Administration course page.