VSAN on Proxmox 9 not connecting

[SteveJK]

I have just updated one of my nodes to Proxmox 9.0.3 I have Starwind 8 build 1955

I can connect form that node to the starwind on the proxmox 8.2.6 with no problem

when I try to connect to the starwind on the Proxmox 9.0.3 node i get

Code: Select all

pve2:~# iscsiadm -m discovery -t st -p *.*.*.*
192.168.103.102:3260,-1 iqn.2008-08.com.starwindsoftware:.95-backup
192.168.103.102:3260,-1 iqn.2008-08.com.starwindsoftware:.95-data

pve2:~# iscsiadm -m node -T iqn.2008-08.com.starwindsoftware:.95-data -p *.*.*.*  -l 
iscsiadm: Could not login to [iface: default, target: iqn.2008-08.com.starwindsoftware:-data, portal: *.*.*.*,3260].
iscsiadm: initiator reported error (8 - connection timed out)
iscsiadm: Could not log into all portals

I can't connect to the same target from my other proxmox server

Code: Select all

rpve1:~# iscsiadm -m discovery -t st -p .*.*.*.*
192.168.103.102:3260,-1 iqn.2008-08.com.starwindsoftware:.95-backup
192.168.103.102:3260,-1 iqn.2008-08.com.starwindsoftware:-data

pve1:~# iscsiadm -m node -T iqn.2008-08.com.starwindsoftware:.95-data -p *.*.*.* -l 
Logging in to [iface: default, target: iqn.2008-08.com.starwindsoftware:.95-data, portal: *.*.*.*,3260]
iscsiadm: Could not login to [iface: default, target: iqn.2008-08.com.starwindsoftware:.95-data, portal: *.*.*.*,3260].
iscsiadm: initiator reported error (8 - connection timed out)
iscsiadm: Could not log into all portals

so iscsiadm sees the mount but I get Could not log into all portals

I am able to connect to the the other node just fine

Code: Select all

pve2:~# iscsiadm -m session
tcp: [16] 192.168.103.101:3260,1 iqn.2008-08.com.starwindsoftware:.90-data (non-flash)

I am at a loss, is it Proxmox 9 or my Starwind install?

rrmoved IPs bscause the form made me

[yaroslav (staff)]

Hi,

Proxmox 9.0.3 has not been tested by us yet. If the previous version works well and the connection can be established, there's an issue on the Proxmox side.
Try setting MTU to 1500 on both target and initiator sides.

[SteveJK]

set the mtu to 1500 on the appliance, and the adaptor in proxmox and nope, not working still.
still not able to connect from 8.4 and 9.0

[yaroslav (staff)]

Check the synchronization status of the device on the updated side. Perhaps, the updated side is not synchronized.
You can check StarWind HA devices sync status in the WEB UI LUN view, or using the Windows-based console.

[SteveJK]

it is not syncing it is "Limited availability" and I a alert is popping up "The replication partner is not synchronised"

[yaroslav (staff)]

Alright, that's something. Thanks for checking!
Try pinging the replication IP from the "live" StarWind VSAN VM. If that does not work, and if you are using VIRTIO adapters, try using VMXNET3 instead. Consider changing the adapter type for SYNC on both nodes only. Beware of downtime if you shut down the only living VM!

[SteveJK]

well changing the nic helped I now see "The replication partner is synchronizing" as soon as it is done I will try to add it to a pve node

[yaroslav (staff)]

Many thanks for your cooperation. Fingers crossed!

[SteveJK]

odd thing is I can get the VMXNET3 nic past 1gb

Code: Select all

root@pve2-starwind:~# ethtool enp6s19 
Settings for enp6s19:
	Supported ports: [ TP ]
	Supported link modes:   1000baseT/Full 
	                        10000baseT/Full 
	Supported pause frame use: No
	Supports auto-negotiation: No
	Supported FEC modes: Not reported
	Advertised link modes:  Not reported
	Advertised pause frame use: No
	Advertised auto-negotiation: No
	Advertised FEC modes: Not reported
	Speed: 1000Mb/s
	Duplex: Full
	Port: Twisted Pair
	PHYAD: 0
	Transceiver: internal
	Auto-negotiation: off
	MDI-X: Unknown
	Supports Wake-on: uag
	Wake-on: d
	Link detected: yes

trying to tun on Auto-negotiation

Code: Select all

# ethtool -s enp6s19 autoneg on
Cannot set new settings: Operation not supported
  not setting autoneg

noting I tried will push it past 1g

i have the same set up on pve1 and that one works fine at 10gb

i guess that is what I get for upgrading to soon

[yaroslav (staff)]

Try benchmarking in multiple threads with iperf from inside StarWind CVM.

[SteveJK]

I decided to move everything odd the VE SAN for now and dump the servers,
once this is working on PVE 9 I will try it again

[yaroslav (staff)]

I think the issue was related to Proxmox networking, not StarWind CVM.

[Heracles31]

Upgrading my own cluster here. 2 Nodes, from 8.4 to 9. Everything starting as Healthy.

On one node (B), I stopped everything, including the Starwind appliance (shutdown from Starwind's WebUI). Once everything down, I started Proxmox upgrade process following their recommendations for an inline upgrade using apt.

The apt dist-upgrade command failed, the open-iscsi package being unable to update properly because of the failed iscsi connection.

I re-started the appliance on that node and reconnected the local proxmox to it.

I did another apt dist-upgrade to upgrade the open-iscsi package. It succeeded thanks to the restored iscsi session.

Rebooted the Proxmox node and everything came back online as it should : the Starwind appliance, itself re-syncing with the one on node A and accepting the connection from node A. As usual, I have to manually reconnect the node B to its own appliance after a reboot. It always tries to connect it itself but in the middle of the boot process, long before the appliance has been started itself.

[yaroslav (staff)]

You need to set up the automatic rescanning, as it looks like Proxmox does not reconnect iSCSI targets automatically.
1. install sshpass on each StarWind CVM.
2. Copy the archive contents to /opt/starwind/starwind-virtual-san/drive_c/starwind
3. set up the rescan script. It can be a custom user who can do iscsiadm rescanning.