Where iohyve VM files are stored?

[Borja]

When setting up iohyve you set a zvol for storing files. I only have one volume on my freenas instance where there are some critical files and im worried that set this volume for iohyve could be a risk for this files. I have to be sure before set up.
also i read about this setup iohyve enabling it in rc.conf, so changes become active only after rebooting. Where i can find some guide to run manually without rebooting?

Thanks.

 

[Artion]

 

[Borja]

View attachment 19017

And what volume you specified in pool argument when you set up iohyve? SPACE?

Thank you!

 

[Artion]

Sorry, the screenshot is referred to FN 11 and to the VM service in the GUI.

For the 9.10.2-Ux, when you specify your pool, in my example SPACE, iohyve creates a set of datasets in that pool and links them as /mnt/iohyve... as in this screenshot:

 

[Borja]

Sorry, the screenshot is referred to FN 11 and to the VM service in the GUI.

For the 9.10.2-Ux, when you specify your pool, in my example SPACE, iohyve creates a set of datasets in that pool and links them as /mnt/iohyve... as in this screenshot:

View attachment 19018

I see, thank you. So, you think the rest of data in volumen (in your case SPACE) is not at risk?

 

[Artion]

is not at risk

Well, as iohyve uses a different dataset it does not write or change other then whats in there. If this is what you mean as 'risk'. Or you mean other kind of risks?

 

[Borja]

Well, as iohyve uses a different dataset it does not write or change other then whats in there. If this is what you mean as 'risk'. Or you mean other kind of risks?

Yes, it what i mean. My SMB share is the volume directly. In your case like you sharing /mnt/SPACE/ so people are storing their files there. I'm worried that in the process of creating the dataset in this volume, some data could be destroyed.

 

[Artion]

In this case is better, IMHO, to create a separate dataset for your samba share, as the 'Ludo' one in the screenshot, move all the share data to that dataset and change the smb share to point to it. Restarting the samba service will update the setting.

PS:
1. you can create as sub datasets as you want. It is good practice to have different datasets for different purposes.
2. iohyve have not to destroy any data anyway, but as it's datasets will be exposed to everyone, someone can corrupt it.
3. you can choose also to add one or two SSD to the system, create a new pool, and use it as your iohyve pool. This way your VMs will not effect the performance of your data pool.

 

[Borja]

In this case is better, IMHO, to create a separate dataset for your samba share, as the 'Ludo' one in the screenshot, move all the share data to that dataset and change the smb share to point to it. Restarting the samba service will update the setting.

PS:
1. you can create as sub datasets as you want. It is good practice to have different datasets for different purposes.
2. iohyve have not to destroy any data anyway, but as it's datasets will be exposed to everyone, someone can corrupt it.
3. you can choose also to add one or two SSD to the system, create a new pool, and use it as your iohyve pool. This way your VMs will not effect the performance of your data pool.

Thank you very much for your clarification, very useful. I'm gonna make a backup, because create a dataset and move data is too risky now. About 8TB of data. And after that let iohyve create the dataset. Also, the idea of adding other SSD is very good.