iSCSI Initiator > Connect to Target > Log On to Target > Service Unavailable

[InphinitePhractals]

Hi,

I currently have the StarWind Virtual SAN setup implemented using the following instructions:

https://www.starwindsoftware.com/techni ... luster.pdf

Everything had been running great, and resilient, for about 2 months until after a recent restart. Now all of our CSVs won’t synchronize:

1 StarWind CSVs Down.png (7.42 KiB) Viewed 23570 times

Through some research, we discovered this might have been the issue:

https://knowledgebase.starwindsoftware. ... kb4019215/

All of our iSCSI targets, minus the Quorum Witness, were stuck in the reconnecting state. We applied the patch KB4034681 designated to have resolved the issue and restarted. Since the problem had not resolved automagically, I attempted to follow the instructions of disconnecting and reconnecting the stuck targets:

2 iSCSI Inactive Targets.png (23.88 KiB) Viewed 23570 times

However, when I attempt to connect the Inactive target I get a dialog saying “Log On to Target - Service Unavailable.”:

3 Log On to Target - Service Unavailable.png (13.81 KiB) Viewed 23570 times

I tried restarting the "Microsoft iSCSI Initiator Service" service as well as each machine and no dice. Any ideas?

Thank you!

[Ivan (staff)]

Hello InphinitePhractals,
Thank you for interest in the StarWind solution.
As far as I can see, your devices have red exclamation marks in the StarWind management console. I can conclude that your CSV1 and CSV2 devices are "Not Synchronized" (that's why you can't connect them via iSCSI).
Firstly, you would need to discover which node has the most recent data. To discover it, you can compare Application logs and find the node with the lastly synchronized devices.
Follow these steps to resolve this issue:
With Management console
1) Open StarWind Management console on the node which has most recent data.
2) Right-click on the CSV1 device (HAImage2 in your case)
3) Click "Mark as Synchronized"
4) Once you will mark it, StarWind will start the synchronization.

Without Management console
1) Go to - C:\Program Files\StarWind Software\StarWind\StarWindX\Samples\powershell
2) Open GetHASyncState.ps1
3) Modify following next strings regarding to your HAImage name and iqn.

Code: Select all

$deviceName = "HAImage1"
    $partnerTargetName = "iqn.2008-08.com.starwindsoftware:192.168.1.156-partnerha1"

4) script above will show you the state of your device.
5) If it's not synchronized, open syncHADevice.ps1 and modify the name of your device like on the example below on the node which contains the most recent data:

Code: Select all

if ( $device.name -eq "HAImage1" )

6) Run the modified script. Once you will run the script, StarWind will start the synchronization.
Please let us know if you have any additional questions.
Have a great day!

[InphinitePhractals]

That got them back online. You're awesome, thank you!

Another issue I'm running into now. I have the SQL Server role installed, and running simple CRUD statements seems to work great, but when I went to run a PowerShell script that restored several databases, all the CSVs lost synchronicity and started syncing from the beginning again. For 600 GB total volume, it looks like is going to take about 2-3 hours. Any ideas on how to prevent that from happening? Thanks again.

[Ivan (staff)]

Hi InphinitePhractals,
Could you please specify the underlying storage (RAID array level, count of disks, type of disks, RPM if it is HDD)?
Could you, also, collect the logs with our LogCollector (https://knowledgebase.starwindsoftware. ... collector/) and share with us the link for download from any cloud (OneDrive/ google drive/DropBox)?

[InphinitePhractals]

There are two (2) nodes with identical hardware. Each node has two (2) physical drives:

1. A smaller Samsung SSD 840 EVO 120GB that Windows Server 2012, StarWind v8, and SQL Server 2014 are installed on.
2. A larger Seagate Barracuda HDD ST1000M003-1CH162 7200RPM 1TB, no RAID, that hosts the Quorum Witness and two (2) CSVs including the SQL Data and Log file folders.

We also have Hyper-V installed with two (2) VMs that appear to be hosted in C:\ClusterStorage\Volume1 as a “Mounted Volume”? It is not my intent to host any files that need to fail-over on the SSD (C:\), but I’m unable to verify if that’s a shortcut (which it appears to be by the arrow) or if they’re directly hosted there.

DropBox, OneDrive, and Google Drive aren’t permitted here so I’ll be PMing you FTP credentials to download the log ZIP files. If they need to go to someone else, let me know and I’ll send the credentials to them as well.

[Boris (staff)]

InphinitePhractals,

C:\ClusterStorage\Volume1 is merely a shortcut, as physically your VMs are stored on StarWind disks, but not the OS disk.

[InphinitePhractals]

Okay, great. I sent the FTP credentials to both Ivan and you (Boris) containing the StarWind log files for each node.

[Boris (staff)]

InphinitePhractals,
Indeed, I received a PM from you, but there were no creds there. Could you please send them to me?

[InphinitePhractals]

They're at the very bottom of the original message, under the quoted text. I resent to you again just in case.

[Boris (staff)]

Got the creds. WIll keep you posted.