IPv6 on ISCSI and Proxmox 8

K

khaled.j.hussein

Active Member
Nov 29, 2017
23
1
43
48
I installed new proxmox node running pve8, and I have cluster of 5 nodes of pve7. I configured ISCSI and multipath to connect to my DELL MD 3820
I joined pve8 node to the cluster, it added and it show all disks on cluster normally

but I see in the journal logs these erroras, I tried to disable IPv6 from interfaces using this
post-up echo 1 > /proc/sys/net/ipv6/conf/eth0/disable_ipv6

but it did not work

Sep 24 09:48:46 pmx10.bisan.com iscsid[221847]: connection-1:0 cannot make a connection to fe80::2a0:98ff:feab:5ab4:3260 (-1,22)
Sep 24 09:48:46 pmx10.bisan.com iscsid[221847]: connection-1:0 cannot make a connection to fe80::2a0:98ff:feab:5aa7:3260 (-1,22)
Sep 24 09:48:46 pmx10.bisan.com iscsid[221847]: connection-1:0 cannot make a connection to fe80::2a0:98ff:feab:5ab3:3260 (-1,22)
Sep 24 09:48:46 pmx10.bisan.com iscsid[221847]: connection-1:0 cannot make a connection to fe80::2a0:98ff:feab:5aa8:3260 (-1,22)

node seems working fine, see all disks on the cluster, but how can I solve these erros

Thanks
 
khaled.j.hussein said:
I installed new proxmox node running pve8, and I have cluster of 5 nodes of pve7. I configured ISCSI and multipath to connect to my DELL MD 3820
I joined pve8 node to the cluster, it added and it show all disks on cluster normally

but I see in the journal logs these erroras, I tried to disable IPv6 from interfaces using this
post-up echo 1 > /proc/sys/net/ipv6/conf/eth0/disable_ipv6

but it did not work

Sep 24 09:48:46 pmx10.bisan.com iscsid[221847]: connection-1:0 cannot make a connection to fe80::2a0:98ff:feab:5ab4:3260 (-1,22)
Sep 24 09:48:46 pmx10.bisan.com iscsid[221847]: connection-1:0 cannot make a connection to fe80::2a0:98ff:feab:5aa7:3260 (-1,22)
Sep 24 09:48:46 pmx10.bisan.com iscsid[221847]: connection-1:0 cannot make a connection to fe80::2a0:98ff:feab:5ab3:3260 (-1,22)
Sep 24 09:48:46 pmx10.bisan.com iscsid[221847]: connection-1:0 cannot make a connection to fe80::2a0:98ff:feab:5aa8:3260 (-1,22)

node seems working fine, see all disks on the cluster, but how can I solve these erros

Thanks
Click to expand...
Hi @khaled.j.hussein
Have you solved this issue?
I think I have the same problem. I have my iSCSI node working well but I have these error in my log every 10 seconds.
It seems related to iPv6. But I don't know how reconfigure this.

Regards
 
Same issue! PVE 9.1.1 + Synology DSM 7.3

I have tried setting
Code: 
node.session.scan = manual
in /etc/iscsi/iscsid.conf and then deleting the IPv6 nodes but that doesn't survive a reboot. Help!
 
Same issue as well: PVE 9.1.9 + Synology DSM 7.3 in HA. I cannot disable IPv6 in DSM as in HA mode IPv6 is not supported but their interfaces still have IPv6 addresses, and these are advertised in the iSCSI discovery. Support calls to Synology so far have said remove the IPv6 nodes from the records, but somewhere in the iscsiadm its refreshing the targets every 2 secs, so can't do that

Edit:
geess, Claude thinks I should disable IPv6 on the iSCSI nic..

Option A — Disable IPv6 on the host's iSCSI-facing NIC (permanent fix)

Find the interface that routes to 172.25.1.4:

bash
Code: 
ip route get 172.25.1.4


Disable IPv6 on that interface only (keeps IPv6 elsewhere on the host intact):


bash
Code: 
sysctl -w net.ipv6.conf.<iface>.disable_ipv6=1

Persist it:


bash
Code: 
cat >> /etc/sysctl.d/99-disable-ipv6-iscsi.conf << 'EOF'
net.ipv6.conf.<iface>.disable_ipv6 = 1
EOF
sysctl -p /etc/sysctl.d/99-disable-ipv6-iscsi.conf


Restart iscsid to clear state:

bash
Code: 
systemctl restart iscsid

With IPv6 disabled on that interface, iscsiadm can still receive the fe80 portal in the SendTargets reply, but it can't bind or add a usable session to it — in practice on PVE 9 / open-iscsi this either drops the record immediately or marks it permanently unreachable rather than persisting as a live node entry. Watch a couple of pvestatd cycles to confirm behavior:

bash
Code: 
watch -n1 'iscsiadm -m node'

Caveat: if this VLAN/bridge carries any other IPv6 traffic you need, this breaks it. Confirm that interface is iSCSI-dedicated first (ip -br addr show <iface>).
Click to expand...
 
Last edited:
I stopped my Synology from advertising IPv6 on the iscsi targets with something like:

Bash: 
nas_ip=192.168.10.10
sed -Ei "s/^(network_portals)=(.*)/\1=$nas_ip:3260/" /usr/syno/etc/iscsi_target.conf

You can use Task Scheduler to run this code at startup.
 
In the end I got Synology Support to disable IPv6 on my interfaces. This should survive reboots and firmware updates according to them
 
@damiengm they didn't tell you what they did? did they just turn off IPv6 completely on the interface, or were they able to somehow selectively disable it on the iSCSI portal only?
 
@luckman212 No they didn't, I'll ask for you, but all our interfaces no longer have any IPv6 (all interfaces have -- next to IPv6 addresses, as before it was a self assigned address), which is fine for us as we have a SA3200D which is natively in HA mode, and are not running any IPv6 network here.
 
@luckman212 here's Synology's reply
Certainly. Since the SA3200D does not provide an IPv6 setting in the DSM interface, we used DSM’s internal network configuration to set IPv6 to Off on the relevant network interfaces.

The setting was applied separately to both controllers because these interface settings are not synchronized automatically between the active and passive controllers. We then performed a switchover and verified that the IPv6 addresses had been removed from both sides.

No iSCSI target, LUN, or stored data configuration was changed.
Click to expand...
 
You must log in or register to reply here.
Top Bottom
Back