Running out of disk space in a vm

[wbravin]

Hello all

I am running Proxmox 8 with OPNsens, truenas scale and home assistant as VMs

MY truenas scale solution issues a warning that the boot disk is 96% full

In the setup of the truenas VM i allocated 64gb of disk space for the truenas VM

My objective is to increase the size to the truenas currently configured a a (scusi 0)

Option 1: A how do i expand the the current size of the disk for this VM ?
Option 2: How could i replace the current disk with a larger one ?

If i will need to replace the disk, I would see that i would need to reinstall Proxmox and all the VMs

If so, is the a process to save all the current proxmox settings and configuration so i could reload once the new disk and a new instance of proxmox is installed?


Thank you for your help

 

[zzz09700]

If you are not changing anything else besides disks on your server, I assume you can take out the old disk, connect it to another Linux machine, dd the whole disk to another larger diskand then place the larger disk back. Everything should boot up as before and you can enlarge your server's partition/file system when the server is running.

Disclaimer: I did this a few times with Ubuntu based servers but never tried it with PVE. I'm just a random guy who works with servers and I'm not responsible for anything bad that happens because of the procedure above, including full data loss of your old disk.

 

[Dunuin]

MY truenas scale solution issues a warning that the boot disk is 96% full

So a virtual disk stored on some PVE storage? You could expand a virtual disk via PVE webUI. But for the VM to actually make use of that space you will have to extend the partitions and filesystems on the guestOS level too. As TrueNAS uses ZFS, you will probably have to extend the partition that hosts the rpool as well as then telling ZFS to extend the pool via "zpool online -e ...".

Did you think about freeing up space on that TrueNAS boot disk instead? Not sure about Scale, but for Core you could remove old TrueNAS versions (it keeps all the old TrueNAS versions, each using multiple GBs, when updating it) via the webUI to free up space. Moving the "system dataset" to some other pool than rpool might help too.

 

[wbravin]

If you are not changing anything else besides disks on your server, I assume you can take out the old disk, connect it to another Linux machine, dd the whole disk to another larger disk and then place the larger disk back. Everything should boot up as before and you can enlarge your server's partition/file system when the server is running.

Disclaimer: I did this a few times with Ubuntu based servers but never tried it with PVE. I'm just a random guy who works with servers and I'm not responsible for anything bad that happens because of the procedure above, including full data loss of your old disk.

Thank you for your reply

My requirement is to increase the size of the allocate disk to the treunas VM.

I have no urgency or need to change the actual SSD, but since i'm going to do this exercise i thought it best to change the whole disk.

Thank you for your prompt reply

 

[wbravin]

So a virtual disk stored on some PVE storage? You could expand a virtual disk via PVE webUI. But for the VM to actually make use of that space you will have to extend the partitions and filesystems on the guestOS level too. As TrueNAS uses ZFS, you will probably have to extend the partition that hosts the rpool as well as then telling ZFS to extend the pool via "zpool online -e ...".

Did you think about freeing up space on that TrueNAS boot disk instead? Not sure about Scale, but for Core you could remove old TrueNAS versions (it keeps all the old TrueNAS versions, each using multiple GBs, when updating it) via the webUI to free up space. Moving the "system dataset" to some other pool than rpool might help too.

Thank you so much for this reply

I do not really know how to extend the partition of the SSD.

Tuenas To my knowledge is not using the SSD as part of the ZFS pool.

This ssd only holds the installation files. Am i correct in assuming this ?

If so i believe that your second paragraphed is more accurate in describing the requirement.

How would i

1. see the system dataset
2. how would i move the extra content of the system dataset

Thank you all for your prompt replies and help

 

[Dunuin]

Tuenas To my knowledge is not using the SSD as part of the ZFS pool.

Usually it got its own ZFS pool for that.
Run a zpool list -v and zfs list -o space to see whata going on.

This ssd only holds the installation files. Am i correct in assuming this ?

The whole OS and multiple copies of it in case you ever did any update.

 

[wbravin]

Usually it got its own ZFS pool for that.
Run a zpool list -v and zfs list -o space to see whata going on.


The whole OS and multiple copies of it in case you ever did any update.

Thank you for this information

i tried both of these command lies

I'm sorry to say that i'm verry ignorant on this subject

I would presume that as it was mentioned, i probably do have more than one in instance of truenas version on that drive

Can i please be so ignorant and ask you to provide me with the command line i will need to insert in the proxmox shell to
a: connect to this drive
b: to delete the unwanted items

Sorry to impose on you all

Thank you for your patience and teachings

 

[Dunuin]

PVE can't access your VMs. You have to run those on the TrueNAS shell.

 

[wbravin]

PVE can't access your VMs. You have to run those on the TrueNAS shell.

Thank your for you instant reply

My understanding is that the boot disk installed on the server is used to install proxmox. Then when you add VM, the disk allocation attributed to VM on the same disk.

At the moment of the creation of the vm is on the same disk.

Currently I have

local pve at 10.1 %
local-lvm pve at 23.2%


Therefore I see that my disk is use at about 50% at the most

I allocated to truenas 64 gb, OPNSense 32 gb and home assistant 32GB

Therefore I can identify 2 issues

1 i miss-allocated disk space to the VMs. I may have over allocated disk space to HA and OPNsense
2 my allocation to truenas has garbage and most likely need more space.

This is why i thought that the cleaning of this drive to free space would be able to do in proxmox shell

I understand the disk reallocation must be done at the creation of the VM. This will probably mean a complete reinstall of the Vms correct?

If so I will then proceed to replace the ssd with a 250gb ssd Which i already have.

Seeing the vms that i am running, how much Space should i allow to the vms?

Can i add a second ssd of the same size and mirror the installation?

Is there a way to avoid this?



Thank you for your patience and time you are spending on this

 

[Dunuin]

MY truenas scale solution issues a warning that the boot disk is 96% full

You said your TrueNAS is complaing about a full boot disk. Not that PVE is complaining about storages getting full.
A VM doesn't know anything about the physical disks and their usage/utilization of the PVE node.

No matter if you decide to free up space on the virtual disk or to increase the size of the virtual disk, both has nothing to do with PVE and has to be done inside TrueNAS. Only thing PVE could do is extend the virtual disk but TrueNAS won't be able to use that additional space without extending the partitions and ZFS pool from within TrueNAS.

PVE only provides the "virtual hardware/computer". What to do with that and how to use it is totally up to the guestOS and isolated from PVE.
So for TrueNAS related questions you will probably find better help when asking in the TrueNAS community forums.

 

[wbravin]

You said your TrueNAS is complaing about a full boot disk. Not that PVE is complaining about storages getting full.
A VM doesn't know anything about the physical disks and their usage/utilization of the PVE node.

No matter if you decide to free up space on the virtual disk or to increase the size of the virtual disk, both has nothing to do with PVE and has to be done inside TrueNAS. Only thing PVE could do is extend the virtual disk but TrueNAS won't be able to use that additional space without extending the partitions and ZFS pool from within TrueNAS.

PVE only provides the "virtual hardware/computer". What to do with that and how to use it is totally up to the guestOS and isolated from PVE.
So for TrueNAS related questions you will probably find better help when asking in the TrueNAS community forums.

Thank you just the same

I did post this issue to trunas forum.

Just to be clear this ssd is not part of the truenas zfs this is a disk used as boot for all VMs on proxmox

this is what is causing my confusion. I thought that since was the common boot disk it would be managed bvu proxmox

 

[Dunuin]

Still not clear to me what you mean. Are you talking about TrueNAS or PVE?
TrueNAS isn't booting from the same disk as PVE. How you described it, TrueNAS boots from that 64GB virtual disks stored on the VM storage on PVE.
If TrueNAS complains about running out of space that 64GB virtual disk is full and according to your outputs the PVE storages (local and local-lvm) got plenty of space left.