Starwind vSAN Free - v8 Upgrade

[BenM]

Guys,

I have a 2 node cluster using Starwind vSAN Free to provide the CSV. Its been in and working largely without problem since 2018 - version 8.0.0.12585

I have a short maintenance window so I have started the upgrade process.

1. Migrated all roles to one node
2. Wait for Syncronisation to be complete (the cluster was quiescent for 4 days - no users at all, just background processes so not much data written to the csv at all)
3. Installed the new Startwind v8.0.0.<latest> on quiet node.

OK so this is where things go a little unexpected...

1. Installation process completes as expected.
2. Starwind console no longer works in read only mode (as it did previously; I am aware that I have the free license... the console used to work in RO mode after the 30 days) so getting status of the HA volumes is slightly more tricky (Powershell)
3. When I get the status of the Starwind HA volume it is worryingly unpopulated...

Code: Select all

Name                    : HAImage2
DeviceType              : HA Image
DeviceId                : 0x0000021DC4C79300
File                    : My Computer\D\CSV1\CSV1_HA.swdsk
TargetName              : iqn.2008-08.com.starwindsoftware:sl-hvbwac05.learning.local-csv1
TargetId                : 0x0000021DC4CB5180
Size                    : 0
CacheMode               : wt
CacheSize               : 5120
CacheBlockExpiryPeriod  : 5000
Exists                  : True
DeviceLUN               : 
IsSnapshotsSupported    : False
Snapshots               : 
SectorSize              : 
State                   : 1
Parent                  : 
Partners                : System.__ComObject
SynchronizationChannels : 
HeartbeatChannels       : 
SyncStatus              : 
SyncPercent             : 
IsWaitingAutosync       : False
SyncTrafficShare        : 
NodeType                : 0
FailoverStrategy        : 0
MaintenanceMode         : 0

Obviously i need to make sure the device is OK and synchronising (I don't think it is at the moment as there is no disk activity and the device status indicates that it isn't syncing)
I don't want to break anything any further - so I haven't rebooted the node (I know that the windows standard response to a problem doesn't help in this instance).
I do have recent backups of the SW configuratuion files (it will take a while to get hold of them) and of course the other node is up anyway.

Everything is untouched.

Can anyone lend a hand?

Regards

[BenM]

A little more information:

The upgraded host iSCSI initiator shows the local connections to the csv and witness as "reconnecting"
The connection to the CSV via the other cluser host is "connected"

In loghtof this, would the solution be to disconnect the existing sessions and discover/connect the local node again in iSCSI initiator?

[yaroslav (staff)]

Greetings,

(as it did previously; I am aware that I have the free license... the console used to work in RO mode after the 30 days)

Some small things I'd like to mention here. A free license does not expire it grants no opportunity to manage the system. You still can connect the server to the Management Console and monitor the HA device state from there. The trial license expires in 30 days and you lose access to the targets. They are not read-only, they become unavailable.

Please share the screenshots from the Management Console that showcase the HA devices' state.

The upgraded host iSCSI initiator shows the local connections to the csv and witness as "reconnecting"

The witness can be either not synchronized. Does it show "reconnecting" on both sides?

The connection to the CSV via the other cluser host is "connected"

Please share the screenshot from iSCSI initiator.

If you have access to GUI, i.e., use the trial version, please contact us at support@starwind.com.

[BenM]

Hi Yaroslav,

I will send the screenshots by email to support@starwind.com

For anyone else who is interested...

UI Screenshots show that on the 8.0.0.12585 node heartbeat and sync channels are down and csv1 is not synchronised. Physical network connections are OK (ping between nodes works fine).
iSCSI on 8.0.0.12585 shows remote witness inactive, remote CSV reconnecting, local witness and CSV connected.
iSCSI om 8.0.0.<latest> shows local CSV and Witness reconnecting, Remote csv connetced, remote witness inactive.

I only commented about the Management UI because the behaviour has changed between 8.0.0.12585 (allows connection to server and displays status/logs etc when using the vSAN Free tier) and the latest (will not connect to server without a licence key - from either node).

B.

[yaroslav (staff)]

Yes, it looks like the key expired. Apply a new one. Keep me posted, please.

[BenM]

Hi Yaroslav,

New license key applied. All appears well.

Thanks as always to the Starwind team - great product and great service.

B.

[yaroslav (staff)]

Greetings,

Thanks for your update. Please keep in mind that the trial version will turn off the storage in 30 days from now. Let me know if more assistance is needed.