Virtual disk Info

[cwiggs]

Hello!

 

I'm looking at creating a new Xpenology VM in proxmox 7 and am looking into what the best virtual disk setup I should use. 

 

Currently my DSM 6 install is using JBOD and when I need to expand I just add a new virtual disk to the VM and add that new vdisk to the JBOD "array", however I would rather just increase the vdisk in proxmox and then expand it in DSM.

 

Searching this forum and I haven't found much info so maybe if I ask the questions here someone can answer.  Here are my questions:

 

• Is there a benefit to using SCSI vs SATA for your storage disks?
• If you are already using ZFS in proxmox there isn't a need to us RAID or BRTFS in DSM so which type of Raid/filesystem should I use?
• How can I expand a virtual disk in DSM after things are installed to it without data lose?
• The vdisks live on an SSD, should I check the SSD emulation in proxmox?
• Why there is so much storage loss with two 16G vdisks in JBOD mode.  I would think it would be slightly less than 32G, but It seems to be ~10G usable, why?

[cwiggs]

Looking through the synology docs and it doesn't seem like there is a good way to grow a virtual disk via the DSM gui.  Not being able to grow the vdisk makes sense since the official DSM systems don't have vdisks for the main DSM.

 

There is some talk on the forum about how to grow a vdisk (usually with esxi) but they are look older.  These tests are on DVA3221 7.1.1-42962 with 1 vdisk in a "basic" storage pool.

 

Here is some info growing a virtual disk:

 

# Check which volumes we are dealing with (in this e.g it's a basic storage pool)

df | fgrep volume1
/dev/mapper/cachedev_0  21543440  107312  21317344   1% /volume1


# Looks like we are using LVM, lets check the physical volume
sudo pvdisplay

  --- Physical volume ---
  PV Name               /dev/md2
  VG Name               vg1
  PV Size               21.77 GiB / not usable 3.00 MiB
  Allocatable           yes
  PE Size               4.00 MiB
  Total PE              5573
  Free PE               194
  Allocated PE          5379
  PV UUID               TmLYpo-38Ky-3Ycm-hUqn-n2M7-u19B-ooOJpo


# Now we can check the "raid" array
sudo mdadm --detail /dev/md2
/dev/md2:
        Version : 1.2
  Creation Time : Wed Dec 28 14:25:33 2022
     Raid Level : raid1
     Array Size : 22830080 (21.77 GiB 23.38 GB)
  Used Dev Size : 22830080 (21.77 GiB 23.38 GB)
   Raid Devices : 1
  Total Devices : 1
    Persistence : Superblock is persistent

    Update Time : Thu Dec 29 09:35:04 2022
          State : clean
 Active Devices : 1
Working Devices : 1
 Failed Devices : 0
  Spare Devices : 0

           Name : nas04:2  (local to host nas04)
           UUID : d0a76bee:7701a3b8:9b3649e9:9f0dafc2
         Events : 13

    Number   Major   Minor   RaidDevice State
       0       8       19        0      active sync   /dev/sdb3


# used to stop DSM service (syno_poweroff_task used to be used)
sudo synostgvolume --unmount -p /volume1

# Resize the parition
parted /dev/sdb resizepart 3 100%

# Then resize the raid array
mdadm --grow /dev/md2 --size=max

# Not sure if we need to reboot, but idk how to bring the services and volume back online.
reboot


# Once DSM is back up you can login go to storage manager > storage pool and expand it there, however we will stick with the cli.
# Expand the physical volume first
sudo pvresize /dev/md2

# Then extend the logical volume
sudo lvextend -l +100%FREE /dev/vg1/volume_1


# Next resize the actual partition (ext4)
# I got an error saying the device was busy, I tried using synostgvolume again but i still wasn't able to resize the partition.
# trying to resize via the GUI also gave an error and asked me to submit a support ticket.
sudo resize2fs -f /dev/vg1/volume_1

 

Edited December 29, 2022 by cwiggs

[cwiggs]

After failing to be able to resize a virtual disk that was using "basic" raid level and ext4 I tried with "basic" raid level and brtfs.

 

The interesting thing I've found is the `brtfs` command seems to be missing in the cli on 7.1.1, however I was able to expand the filesystem using the gui, here is how:

 

# Not 100% sure if we need this command
sudo synostgvolume --unmount -p volume1

parted /dev/sda resizepart 3 100%

mdadm --grow /dev/md2 --size=max

reboot

Then after it reboots go into DSM > storage manager > storage pool > expand

 

I assume we could do the same thing with ext4 but I haven't tried it.

[cwiggs]

More findings: the JBOD type in DSM is just a "linear" mdadm array when you ssh in and check.  It seems you can't grow a mdadm array that is linear:

 

sudo mdadm --grow /dev/md2 --size=max
mdadm: Cannot set device size in this type of array.

 

So with JBOD mode the only way to add storage space would be to attach another vdisk and add it to the JBOD array.  That is currently how I do it with DSM 6 but IMO it isn't ideal.

[cwiggs]

More findings: I rebuilt the storage pool as a "basic" array with ext4 as the filesystem.  Here is then what I did to grow the vdisk:

 

sudo parted /dev/sdb resizepart 3 100%

sudo mdadm --grow /dev/md2 --size=max

sudo reboot

After a reboot login to DSM > Storage Manager > Storage pool > Expand

 

So it looks like you don't need to take any DSM service offline (synostgvolume --unmount -p volume1) in order to grow a virtual disk.  It would be nice to be able to do all this in the CLI or all of it in the GUI but it looks like for now that isn't possible.