Migrate Existing Node with Windows 10 Powershell

Hey All,

I’ve looked for steps to move my node to another hard-drive in windows using powershell. I’ve read the FAQ (which is not very informative — I found clearer steps in the forum, however, in Ubuntu.).

The clearest steps are here

Now I would like to do the same thing but in windows. Can I use the same commands?

1 Like

Yes, you can. But not in Powershell, in the Ubuntu shell of WSL.

  1. Enable WSL
  2. Install Ubuntu from the Microsoft Store
  3. Install/Enable SSH server on the new server (Windows can have a SSH service too)
  4. Use the same commands from the Ubuntu shell

For Cmd/Powershell there is another option:

  1. copy identity to a new server
robocopy /MIR "$env:AppData\Storj\Identity\storagenode" \\server2\storj\identity
  1. run robocopy while the node running
robocopy /MIR y:\storj\storage \\server2\storj\storage
  1. Run it a few more times.
  2. When the difference is negligibly - stop and disable the storagenode service (or stop and remove the container in case of docker)
  3. Run robocopy /MIR one more time
  4. Start the node on the new server
  5. Uninstall the storagenode on the old server (in case of Windows GUI)
5 Likes

Thank you Alexey,

These instructions are clear. I’ll try them now.

Do I also copy the other files in the main directory, e.g., revocations.db, kademlia ect.?

Cheers,
Adam

Yes. All data should be copied, including databases. Just copy the whole folder with data

Hey Alexey, It seems that for one file i get an access denied. Thus I can’t completely clone my database to another drive. It is strange since last week this was successful… I also tried running powershell as admin without success. Any ideas?

This piece is uploading to your node right now. The Windows doesn’t allow to touch files open for write by another process

1 Like

What do I do if the robocopy stops after 10% saying file is damaged and can not be red? Do I have to start a new node? Or can I delete that file and try again?image

You should not do either first or the second. Just let it copy. This file could be used right now.
If you delete it, your node will fail audit for that piece.

After i repair the hard drive it continues a bit more. Is there a way to skip the file when the copy process halts? If I restart the robocopy without repairing it just stops on the same file again.

The old hard drive is an old external 2tb. I am suspended on one satelite at the moment, but was suspended on others too last week.

The external hard drive takes some time to spin up if it hasn’t had activity for some time, so hopefully it is just that, but I’m suspecting it will fail soon.

Edit: thanks for the reply :blush:

I have migrate the node with FreeFileSync and i don’t have problems. Use this software.

1 Like

In such case you can stop the storagenode service to let the file free from usage. If the file will not be accessible even with administrator’s rights, then you should check your disk with a scandisk.

Hello! Two weeks ago I launched my first node running Windows 10. It contains one SSD with operating system and storagenode.exe installed. And there is RAID10 4TB, which is used only for data of storj. Tell me, who knows, can I replace this raid10 with another hdd 4TB simply by copying the data? And how it better to do?

Hello @Alian,
Welcome to the forum!

https://documentation.storj.io/resources/faq/migrate-my-node

1 Like

raid 1 can be split into two working drives also know as a mirror because they are just the same data, raid 10 is basically a stripe across two or more raid 1

so yes you can copy the data onto a single disk, tho the disk ofc needs to have more space than the raid10 array holds, and you should expect it to take quite a while… tho raid10 is pretty fast, much faster than the single drive you will be copying to.

1 Like

Hey sorry to bother but I have a question. I store my node data in. VHDX just to be able to segment things. I’m migrating my node to a new system. Should I mount the drive and copy it each file separately or copy the entire VHDX. Note the node will be offline during this process.

I’m worried about a possibility of corruption with the massive single VHDX as robocopy doesn’t check if it has copied the file correctly

As you say robocopy doesn’t have a verification option built-in. However if you are only copying the single vhdx then robocopy with it’s multi-thread copy capability is not really required. You should be able to xcopy or even Windows copy, the dismounted vhdx, as both have file verification options ( /v).

Ok cool will try thanks.