Why can't I extend a device/image?

[wallewek]

Perhaps my previous query got missed because I responded to another thread, or I asked too many questions
(see https://forums.starwindsoftware.com/vie ... ice#p27003).

Using the Management Console, when I right-click on an HA image, the "Extend Size of HA (High Availability) Device..." action is listed but grayed out.

Why is it grayed out, and how do I fix it?

I tried ExtendDevice.ps1, but it just exited with "200 Failed: operation cannot be completed" and took the OTHER cluster node offline! I had to reboot it to restore operation. I don't have ExtendImageFile script.

This is a new two-host Hyper-V HA setup, running version 11404. Should I upgrade to 11456, and if so, how can I do so without re-registering and doing a full reinstall?

-- Ken

[Sergey (staff)]

Hello, Ken. I advise you to update to the latest 11456 build. In order to do that, please quit StarWind Management Console, download the latest stable release build from this link: https://www.starwindsoftware.com/tmplin ... ind-v8.exe and launch the installation file. No need to register it again, as you will update within StarWind version 8. Check this step-by-step guide:
1. Check that all StarWind devices have status “Synchronized” on both nodes;
2. Move all Cluster resources from the node where StarWind will be updated;
3. Install the latest StarWind version;
4. Wait until StarWind service starts and the synchronization process begins;
5. Wait until the synchronization process is over;
6. Check that all StarWind devices have status “Synchronized” on both nodes;
7. Move all Cluster resources from the node where StarWind has been updated to another one;
8. Repeat steps 3, 4, 5 and 6 for another node.
After that please try to extend device again.

[wallewek]

Thank you Sergey,

FWIW, the download URL link to the .exe in this thread is 404, looks like a copy/paste issue. The one on the other thread ( https://forums.starwindsoftware.com/vie ... 679#p27016 ) seems fine.

-- Ken

[wallewek]

I followed your excellent, clear instructions, Sergey, and had no problem upgrading to version 11456.

However, I still seem to have exactly the same issue -- the Extend option is grayed out and unavailable. I've attached a screenshot. Can you see anything wrong?

-- Ken

[Sergey (staff)]

Dear Ken. Now the issue is clear for me. LSFS device cannot be extended at the moment. This will be implemented in one of the upcoming updates, not in the nearest one though.

[wallewek]

Ah! Thank you for catching that, Sergey.

I had forgotten that I used lsfs filesystem (had to halt work on StarWind cluster for a bit after initial setup).
Since this HA cluster will be used purely for hyper-v hosting with SATA drives and SSD cache, lsfs seems appropriate. Although I'm wondering if I should have enabled deduplication. Oh, well.

One question, about lsfs recommended space allocation:

I understand lsfs wants lots of spare space, but I'm unclear at what storage level this should be allocated.

Is this saying that the StarWind device - image size settings should be set to only a fraction (say, 30%) of the host server physical disk partition on which is it located?

Or is it referring to physical file allocation at the host level, where I see a whole lot of .spspx files? And do I just need to avoid over-filling the physical host partition? That actually makes more sense to me.

If it's the former, in this case, the 500 GB image is about all I should allocate on the underlying 1.5 TB hard drive partition? It looks odd at the host filesystem level: that 500 GB lsfs image folder is only using about 35 GB in actual file space on the 1.5 TB partition, the rest is empty! It's so tempting to allocate another image there.

Whereas if it's the latter, I have LOTS of room to allocate more storage, and I could easily put another image in the same host partition with room to spare.

Could you clarify that please? Thanks!


-- Ken

[wallewek]

Quick note: I found the StarWind LSFS FAQ, which says "LSFS will consume 2.5 to 3 times more space than specified in the StarWind management console."

So that seems to clarify that a 500 GB LSFS image definition is about all I should put on a 1.5 TB drive.

This cluster doesn't need maximum performance. It needs low cost capacity and reliability. I'm starting to think I should tear down the storage configuration and switch back to thick provisioning.

-- Ken

[Sergey (staff)]

The main purpose of LSFS is to increase performance on spindle HDD drives at the expense of storage capacity. LSFS is thin provisioned. These .spspx files are snapshot files, and new .spspx files will be created when you write something to HA storage.
Please, note, if you experience the situation, when there is no space for new snapshots, created by LSFS, your HA storage will operate in read-only mode.
Let me guide you to this blog post, where the idea of LSFS is described in details: https://www.starwindsoftware.com/blog/l ... a-nutshell
In your case, as your cluster doesn't need maximum performance, but needs low-cost capacity and reliability, a thick provisioned disk can be a better choice.