2025年度折腾 - All-in-One 家庭服务器 - 序章和目录

Posts

• 2025年度折腾 - All-in-One 家庭服务器 - 第一篇: 硬件以及存储架构
• 2025年度折腾 - All-in-One 家庭服务器 - 第二篇: 网络搭建
• 2025年度折腾 - All-in-One 家庭服务器 - 第三篇: 备份系统

Notes

• About Network
  • [opnsense-access-through-wan]
  • [script-to-update-proxmox-hosts-file-with-ip-addresses-of-running-vms-and-lxcs]
  • [script-to-sync-pve-hosts-entries-to-adguard-home]
  • [script-to-get-ip-addresses-for-all-running-vms-and-lxcs-on-proxmox]
  • dhcp-and-asus-router-and-unbound-on-opnsense
• About Applications
  • docker-on-lxc-proxmox-2025
  • [install-caddy-in-lxc-and-finish-dns-01-challenge]
  • [frigate-in-lxc-with-intel-igpu]
• About SAMBA shares
  • [allow-a-specific-user-group-to-access-a-default-samba-share-on-truenas]
• About VM/Host management and UX
  • [ubuntu-vm-enable-xterm]
  • [change-debian-13-from-using-dhcp-to-static-ip]
• About Data Migration
  • [rsync-to-truenas-via-ssh]
• About ‘Why’
  • [some-issues-with-truenas-as-the-host-os-for-aio]

Snippets

After proxmox installation:

zfs get atime rpool
zfs set atime=off rpool
zfs set xattr=sa rpool
# Do the same for 'all zfs pools'.

• In the UI, for the ‘storage’ to be used for VM disks, make sure we set ‘volblocksize’ to 64k.
• zfs monthly scrub:

systemctl status zfs-scrub-monthly@rpool.timer
 
# if 'disabled' or 'inactive':
systemctl enable --now zfs-scrub-monthly@rpool.timer

After opnsense installation:

• Qemu Agent
  • System → Firmware → Plugins → os-qemu-guest-agent
• Set /var/log to ram disk:
  • System → Settings → Miscellaneous → ‘var/log RAM disk’
    • Not setting ‘/tmp’ to ram disk since I think it will block ‘firmware updates’.

Provision Big QCOW on TrueNAS via NFS

Error I saw:

update VM 104: -scsi1 truenas-pve-hdd-vm-disks:2048,format=qcow2,iothread=on
 
Formatting '/mnt/pve/truenas-pve-hdd-vm-disks/images/104/vm-104-disk-0.qcow2', fmt=qcow2 cluster_size=65536 extended_l2=off preallocation=metadata compression_type=zlib size=2199023255552 lazy_refcounts=off refcount_bits=16
 
TASK ERROR: unable to create image: 'storage-truenas-pve-hdd-vm-disks'-locked command timed out - aborting

Basically, timed out, both when:

• you create the qcow with 2TB size (probably will be triggered anything higher than 64GB)
• or you try to increase your disk gradually with step high than 64GB.

My TrueNAS tank setup:

• 4x 14TB HDD (Segate EXOS and WD Enterprise)
• 1x MX500 SATA SSD for Cache (L2 ARC)
• 64GB RAM
• Anyhow, still too slow for creating/reserving big qcow.

Solution:

• Temporarily set the ‘sync’ flag for your VM disks zfs dataset to off
• Re-run the qm resize 2048G → Should work now.
• Change the ‘sync’ flag back to ‘standard’