How to Troubleshoot iSCSI Connection Problems on Linux

Updated: 30 September 2026 · Applies to: iscsi-initiator-utils 6.2.1 (open-iscsi) on AlmaLinux and Rocky Linux 9 and 10, RHEL 9 and 10; open-iscsi on Ubuntu 24.04 and 26.04 and Debian 13

When an iSCSI disk does not appear, disappears, or is slow, work from the network up to the session and the disk. The commands below are the same on all current distributions (the package is iscsi-initiator-utils on AlmaLinux, Rocky Linux and RHEL, and open-iscsi on Ubuntu and Debian). For the basics of iscsiadm, see understanding the iscsiadm utility.

1. Can the server reach the target?

ping -c 3 192.168.50.10
nc -zv 192.168.50.10 3260

iSCSI uses TCP port 3260. If the port is closed, check the storage system, the firewalls on both sides, and that you use the storage network interface. With several interfaces in the same subnet, see iscsiadm discovery timeout with two or more network interfaces.

2. Is the initiator name allowed?

cat /etc/iscsi/initiatorname.iscsi

The storage system must allow exactly this IQN (and, if used, the CHAP user and password configured in /etc/iscsi/iscsid.conf or on the node). A login that fails with "authorization failure" almost always means the IQN or CHAP settings do not match.

3. Discover and log in again

iscsiadm -m discovery -t sendtargets -p 192.168.50.10
iscsiadm -m node -T iqn.2026-01.com.example:storage.lun1 -p 192.168.50.10 --login

4. Look at the session in detail

iscsiadm -m session -P 3

This shows each session's state (LOGGED_IN), the network interface used, the timeouts, and at the end the SCSI disks attached to it (Attached scsi disk sdc). A session without an attached disk usually means the target has no LUN mapped to this initiator.

5. Read the logs

journalctl -u iscsid -n 50 --no-pager
journalctl -k | grep -i -E "iscsi|connection[0-9]+:"

Messages such as detected conn error and repeated reconnects point to network problems: packet loss, an MTU mismatch (if you use jumbo frames, every device on the path must use the same MTU), or an overloaded storage system.

6. Check performance

iostat -x 5 sdc          # package sysstat: latency (r_await, w_await) and utilisation

High wait times with low throughput suggest the network or the storage system; test the path with ping -s 8972 -M do 192.168.50.10 for jumbo frames.

7. Multipath

With two paths to the storage, use multipath -ll to check that both are active. A failed path shows as failed faulty; the disk keeps working over the other path until you fix it.

Make logins survive a reboot

iscsiadm -m node -T iqn.2026-01.com.example:storage.lun1 -p 192.168.50.10 --op update -n node.startup -v automatic
systemctl enable iscsid.socket iscsi.service          # AlmaLinux, Rocky Linux, RHEL
systemctl enable iscsid.socket open-iscsi.service     # Ubuntu, Debian

Mount iSCSI file systems in /etc/fstab with _netdev, so they wait for the network.

Official documentation: open-iscsi and RHEL 9: configuring an iSCSI initiator.