targetcli sessions detail shows no active sessions

[devedse]

When I run the command:

root@vm-iscsibeast2:~# targetcli sessions detail
(no open sessions)

I however am 100% sure that there's an active iSCSI session where I booted windows from a specific disk.

Here's the full iSCSI configuration:

root@vm-iscsibeast2:~# targetcli ls
o- / ......................................................................................................................... [...]
  o- backstores .............................................................................................................. [...]
  | o- block .................................................................................................. [Storage Objects: 4]
  | | o- disk_devedse1 ........................................ [/dev/iscsi_thick_vg/iscsi_devedse1 (200.0GiB) write-thru activated]
  | | | o- alua ................................................................................................... [ALUA Groups: 1]
  | | |   o- default_tg_pt_gp ....................................................................... [ALUA state: Active/optimized]
  | | o- disk_devedse2 ........................................ [/dev/iscsi_thick_vg/iscsi_devedse2 (250.0GiB) write-thru activated]
  | | | o- alua ................................................................................................... [ALUA Groups: 1]
  | | |   o- default_tg_pt_gp ....................................................................... [ALUA state: Active/optimized]
  | | o- disk_greataa1 ........................................ [/dev/iscsi_thick_vg/iscsi_greataa1 (250.0GiB) write-thru activated]
  | | | o- alua ................................................................................................... [ALUA Groups: 1]
  | | |   o- default_tg_pt_gp ....................................................................... [ALUA state: Active/optimized]
  | | o- disk_wompus1 ........................................... [/dev/iscsi_thick_vg/iscsi_wompus1 (25.0GiB) write-thru activated]
  | |   o- alua ................................................................................................... [ALUA Groups: 1]
  | |     o- default_tg_pt_gp ....................................................................... [ALUA state: Active/optimized]
  | o- fileio ................................................................................................. [Storage Objects: 0]
  | o- pscsi .................................................................................................. [Storage Objects: 0]
  | o- ramdisk ................................................................................................ [Storage Objects: 0]
  o- iscsi ............................................................................................................ [Targets: 4]
  | o- iqn.2024-11.local.vm-iscsibeast:devedse1 .......................................................................... [TPGs: 1]
  | | o- tpg1 .................................................................................................. [gen-acls, no-auth]
  | |   o- acls .......................................................................................................... [ACLs: 0]
  | |   o- luns .......................................................................................................... [LUNs: 1]
  | |   | o- lun0 .................................... [block/disk_devedse1 (/dev/iscsi_thick_vg/iscsi_devedse1) (default_tg_pt_gp)]
  | |   o- portals .................................................................................................... [Portals: 1]
  | |     o- 0.0.0.0:3260 ..................................................................................................... [OK]
  | o- iqn.2024-11.local.vm-iscsibeast:devedse2 .......................................................................... [TPGs: 1]
  | | o- tpg1 .................................................................................................. [gen-acls, no-auth]
  | |   o- acls .......................................................................................................... [ACLs: 0]
  | |   o- luns .......................................................................................................... [LUNs: 1]
  | |   | o- lun0 .................................... [block/disk_devedse2 (/dev/iscsi_thick_vg/iscsi_devedse2) (default_tg_pt_gp)]
  | |   o- portals .................................................................................................... [Portals: 1]
  | |     o- 0.0.0.0:3260 ..................................................................................................... [OK]
  | o- iqn.2024-11.local.vm-iscsibeast:greataa1 .......................................................................... [TPGs: 1]
  | | o- tpg1 .................................................................................................. [gen-acls, no-auth]
  | |   o- acls .......................................................................................................... [ACLs: 0]
  | |   o- luns .......................................................................................................... [LUNs: 1]
  | |   | o- lun0 .................................... [block/disk_greataa1 (/dev/iscsi_thick_vg/iscsi_greataa1) (default_tg_pt_gp)]
  | |   o- portals .................................................................................................... [Portals: 1]
  | |     o- 0.0.0.0:3260 ..................................................................................................... [OK]
  | o- iqn.2024-11.local.vm-iscsibeast:wompus1 ........................................................................... [TPGs: 1]
  |   o- tpg1 .................................................................................................. [gen-acls, no-auth]
  |     o- acls .......................................................................................................... [ACLs: 0]
  |     o- luns .......................................................................................................... [LUNs: 1]
  |     | o- lun0 ...................................... [block/disk_wompus1 (/dev/iscsi_thick_vg/iscsi_wompus1) (default_tg_pt_gp)]
  |     o- portals .................................................................................................... [Portals: 1]
  |       o- 0.0.0.0:3260 ..................................................................................................... [OK]
  o- loopback ......................................................................................................... [Targets: 0]
  o- vhost ............................................................................................................ [Targets: 0]
  o- xen-pvscsi ....................................................................................................... [Targets: 0]

And here's a tcpdump:

root@vm-iscsibeast2:~# tcpdump -i any -nn port 3260
00:06:07.159954 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95397443:95398823, ack 35041, win 65532, length 1380
00:06:07.159963 eth0  Out IP 10.88.21.223.3260 > 10.62.0.50.49695: Flags [.], ack 95398823, win 2505, length 0
00:06:07.159995 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95398823:95401583, ack 35041, win 65532, length 2760
00:06:07.160005 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95401583:95402963, ack 35041, win 65532, length 1380
00:06:07.160020 eth0  Out IP 10.88.21.223.3260 > 10.62.0.50.49695: Flags [.], ack 95402963, win 2505, length 0
00:06:07.160052 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95402963:95405723, ack 35041, win 65532, length 2760
00:06:07.160074 eth0  Out IP 10.88.21.223.3260 > 10.62.0.50.49695: Flags [.], ack 95405723, win 2505, length 0
00:06:07.160103 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95405723:95407103, ack 35041, win 65532, length 1380
00:06:07.160113 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95407103:95411243, ack 35041, win 65532, length 4140
00:06:07.160113 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95411243:95412623, ack 35041, win 65532, length 1380
00:06:07.160126 eth0  Out IP 10.88.21.223.3260 > 10.62.0.50.49695: Flags [.], ack 95412623, win 2505, length 0
00:06:07.160149 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95412623:95415383, ack 35041, win 65532, length 2760
00:06:07.160172 eth0  Out IP 10.88.21.223.3260 > 10.62.0.50.49695: Flags [.], ack 95415383, win 2505, length 0
00:06:07.160204 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95415383:95416763, ack 35041, win 65532, length 1380
00:06:07.160234 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [P.], seq 95416763:95418143, ack 35041, win 65532, length 1380
00:06:07.160244 eth0  Out IP 10.88.21.223.3260 > 10.62.0.50.49695: Flags [.], ack 95418143, win 2505, length 0
00:06:07.160501 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95418143:95419523, ack 35041, win 65532, length 1380
00:06:07.160556 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95419523:95420903, ack 35041, win 65532, length 1380
00:06:07.160556 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95420903:95423663, ack 35041, win 65532, length 2760
00:06:07.160571 eth0  Out IP 10.88.21.223.3260 > 10.62.0.50.49695: Flags [.], ack 95420903, win 2505, length 0
00:06:07.160578 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95423663:95427803, ack 35041, win 65532, length 4140
00:06:07.160578 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95427803:95430563, ack 35041, win 65532, length 2760
00:06:07.160583 eth0  Out IP 10.88.21.223.3260 > 10.62.0.50.49695: Flags [.], ack 95423663, win 2505, length 0
00:06:07.160598 eth0  Out IP 10.88.21.223.3260 > 10.62.0.50.49695: Flags [.], ack 95430563, win 2505, length 0
00:06:07.160685 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95430563:95431943, ack 35041, win 65532, length 1380
00:06:07.160721 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95431943:95433323, ack 35041, win 65532, length 1380
00:06:07.160721 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95433323:95436083, ack 35041, win 65532, length 2760
00:06:07.160731 eth0  Out IP 10.88.21.223.3260 > 10.62.0.50.49695: Flags [.], ack 95433323, win 2505, length 0
00:06:07.160745 eth0  Out IP 10.88.21.223.3260 > 10.62.0.50.49695: Flags [.], ack 95436083, win 2505, length 0
00:06:07.160769 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95436083:95437463, ack 35041, win 65532, length 1380
00:06:07.160789 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95437463:95441603, ack 35041, win 65532, length 4140
00:06:07.160789 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95441603:95444363, ack 35041, win 65532, length 2760
00:06:07.160813 eth0  Out IP 10.88.21.223.3260 > 10.62.0.50.49695: Flags [.], ack 95444363, win 2505, length 0
00:06:07.160831 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95444363:95445743, ack 35041, win 65532, length 1380
00:06:07.160846 eth0  In  IP 10.62.0.50.49695 > 10.88.21.223.3260: Flags [.], seq 95449883:95452643, ack 35041, win 65532, length 2760

[gonzoleeman]

I just tested this on SUSE Tumbleweed, which has a fairly new kernel and target and initiator packages.

It worked just fine for me.

[devedse]

I'm using debian12 for my host. Could that have anything to do with it?

[gonzoleeman]

I'm using debian12 for my host. Could that have anything to do with it?

I have no idea. Debian 12 is a couple of years old? If I was you, I might try to narrow down where the issue is. I'm pretty sure the targetcli code just scans sysfs to find "sessions". That's just a guess, but that'd be the easiest way. If so, maybe the targetcli code is doing the wrong thing?

Targetcli-fb uses rtslib-fb and configshell-fb to do much of it's dirty work. If you know python, you could look in the code to see what it's doing. Or you could try installing the latest upstream of those 3 packages (upgrading just one may or may not work).

Or the problem could be in open-iscsi, but I doubt it. The open-iscsi package creates sessions, but the kernel code is what creates the sysfs entries. So changing/updating the open-iscsi packages probably won't help. What does "iscsiadm -m session" say?

[maurizio-lombardi]

I tested it too on Fedora rawhide and works for me, targetcli 3.0.1.
Can you get the version of targetcli your Debian system is running?

I remember that some time ago I used a Debian-based OS on the RaspberryPI and I was surprised by the fact that it was running a really ancient version of targetcli. Maybe is this the case?

[devedse]

I just checked and apparently I'm using version 2.1.53 on debian 12.

Is there an easy way to upgrade to the latest version?

@gonzoleeman , I think iscsiadm is for the client side of iscsi right? (As that tool is not installed on my debian12 machine)

Also I'm not that experiences in digging through those packages internals.

[maurizio-lombardi]

@devedse Hmm, I tested it on an old RHEL system which ships with 2.1.53 and I couldn't reproduce it:

# targetcli sessions detail
alias: ampere-mtsnow-altra-02-vm sid: 1 type: Normal session-state: LOGGED_IN
name: iqn.1994-05.com.redhat:2237729e73f5 (NOT AUTHENTICATED)
mapped-lun: 0 backstore: fileio/disk mode: rw
address: 10.8.2.214 (TCP) cid: 0 connection-state: LOGGED_IN
# rpm -q targetcli
targetcli-2.1.53-7.el9.noarch

I guess I should try to download Debian 12 and see what happens there...

[devedse]

@maurizio-lombardi , if you could help me with testing it on debian 12 that would be amazing!

This is the script I use to setup the iscsi target on a thickly (is this a word?) provisioned LVM:

#!/bin/bash

set -e

# Configuration
DEFAULT_VG="iscsi_thick_vg"  # Default volume group name for thick provisioning
IQN_PREFIX="iqn.2024-11.local.vm-iscsibeast"

# Function to convert size to bytes for validation
size_to_bytes() {
    local size=$1
    local num=$(echo $size | sed 's/[^0-9.]//g')
    local unit=$(echo $size | sed 's/[0-9.]//g' | tr '[:lower:]' '[:upper:]')
    
    case $unit in
        ""|"B") echo "${num%.*}" ;;
        "K"|"KB") echo $((${num%.*} * 1024)) ;;
        "M"|"MB") echo $((${num%.*} * 1024 * 1024)) ;;
        "G"|"GB") echo $((${num%.*} * 1024 * 1024 * 1024)) ;;
        "T"|"TB") echo $((${num%.*} * 1024 * 1024 * 1024 * 1024)) ;;
        *) echo "0" ;;
    esac
}

USERNAME="$1"
SIZE="$2"
VG="${3:-$DEFAULT_VG}"

SIZE_BYTES=$(size_to_bytes "$SIZE")

LV_NAME="iscsi_${USERNAME}"
LV_PATH="/dev/$VG/$LV_NAME"

VG_FREE=$(vgs --noheadings --units b -o vg_free "$VG" | tr -d ' B')

TARGET_NAME="$IQN_PREFIX:$USERNAME"
SERVER_IP=$(hostname -I | awk '{print $1}')

echo "=== Creating iSCSI Disk for User: ${USERNAME} with LVM thick provisioning ==="

echo "Creating iSCSI target with LVM thick provisioning..."
echo "Username: $USERNAME"
echo "Size: $SIZE (fully allocated)"
echo "Volume Group: $VG"
echo "Logical Volume: $LV_NAME"
echo "Target Name: $TARGET_NAME"

# Create the logical volume with thick provisioning
echo "Step 1: Creating logical volume ${LV_PATH} (${SIZE} fully allocated)"
lvcreate -L "$SIZE" -n "$LV_NAME" "$VG" --wipesignatures y
if [ $? -ne 0 ]; then
    echo "❌ Failed to create logical volume"
    exit 1
fi

echo "Step 2: Creating iSCSI storage object and target using targetcli..."
# Use targetcli to configure the iSCSI target
targetcli <<EOF
/backstores/block create name=disk_${USERNAME} dev=${LV_PATH}
/backstores/block/disk_${USERNAME} set attribute emulate_tpu=1
/iscsi create ${TARGET_NAME}
/iscsi/${TARGET_NAME}/tpg1/luns create /backstores/block/disk_${USERNAME}
/iscsi/${TARGET_NAME}/tpg1 set attribute authentication=0 generate_node_acls=1 demo_mode_write_protect=0
/iscsi/${TARGET_NAME}/tpg1 enable
saveconfig
exit
EOF

if [ $? -ne 0 ]; then
    echo "❌ Failed to create iSCSI target"
    # Cleanup logical volume on failure
    echo "Cleaning up logical volume..."
    lvremove -f "$LV_PATH" 2>/dev/null
    exit 1
fi

[gonzoleeman]

@gonzoleeman , I think iscsiadm is for the client side of iscsi right? (As that tool is not installed on my debian12 machine)

Also I'm not that experiences in digging through those packages internals.

Yes, iscsiadm is part of open-iscsi, which is a Linux initiator. As I mentioned, I did not test with non-Linux.

[devedse]

I could try to setup a Linux iscsi cliënt but I don't see how that would differ from a windows client. (From showing the sessions perspective)

[devedse]

I just did a complete fresh install of debian13 but it still results in the same issue:

root@vm-iscsibeast3:~# targetcli sessions detail
(no open sessions)

root@vm-iscsibeast3:~# targetcli --version
/usr/bin/targetcli version 2.1.53

[devedse]

@maurizio-lombardi @gonzoleeman , is there some guidance on how to compile from source?

[gonzoleeman]

These days installing python has become somewhat of a PITA IMHO. :)

You might be able to get away with something like "python3 -mpip install --prefix /usr/local --verbose ." But you will have to install rtslib-fb, configshell-fb, and targetcli-fb. FWIW the code to actually handle sessions seems to be in rtslib. Also, as a reference, on one of my systems I see only one session listed from "targetcli sessions" even though I had two sessions at the time. So there are obviously some bugs in there?

Note that if you install new stuff in /usr/local you'll leave the original stuff in place. You'll have to run the /usr/local/bin/targetcli version, and ensure it's getting it's python libraries from /usr/local/... as well, using PYTHONPATH.

[devedse]

@gonzoleeman , isn't there a normal "requirements.txt" and some other things that would just allow me to run it without installing it?

[gonzoleeman]

On Aug 19, 2025, at 12:50 PM, Devedse ***@***.***> wrote: devedse left a comment (open-iscsi/targetcli-fb#223) <#223 (comment)> @gonzoleeman <https://github.com/gonzoleeman> , isn't there a normal "requirements.txt" and some other things that would just allow me to run it without installing it?

You can use python in pip mode (I believe) to install it in a “virtual environment”, which is really a subdirectory tree with proper library path setup, as far as I understand. I shudder to say, but you can save a copy of the old files, and then just copy the new ones into places. That’s not recommended, but has been done. :) By the way, I looked through the git commit list for rtslib and targetcli and didn’t see anything related to “sessions” that stood out, and no changes recently at all. Even though your version is old, the changes for “sessions” seems older. — Lee D