家用NAS/Server/Homelab集中討論區(8)

單隻3.5” enclosure好撚low tech，隻隻都伏，同時間隻隻都沖唔伏

其實用咗咁多年有一隻Orico 行247都冇壞
不過我用唔壞又唔代表啲咩

Postfix本身我喺Proxmox有用開，佢會同我定期講下啲harddisk點，有咩error都即刻有email notifications ，我覺行NAS哩點好緊要

我postfix都係裝係openwrt到
openwrt乜q都有 真係好鬼正

raid唔係backup

sfp+ port
買xgspon stick
然後唔知點搞 係modem抽一堆id 寫返落ponstick入面
好處只係用少部機

我都未有興趣實行

用緊ugreen，一隻22t hdd

近排試緊kopia+backblaze b2做backup
2x GB都要成粒鐘係咪正常佢又冇asia dc

同埋發現proxmox lxc一mount住嘢
take backup就會卡住除非用stop mode

我係用suspend mode
https://pve.proxmox.com/wiki/Backup_and_Restore

    This mode uses rsync to copy the container data to a temporary location (see option --tmpdir). Then the container is suspended and a second rsync copies changed files. After that, the container is started (resumed) again. This results in minimal downtime, but needs additional space to hold the container copy.

    When the container is on a local file system and the target storage of the backup is an NFS/CIFS server, you should set --tmpdir to reside on a local file system too, as this will result in a many fold performance improvement. Use of a local tmpdir is also required if you want to backup a local container using ACLs in suspend mode if the backup storage is an NFS server.

要自己ssh入去define個tmpdir (any local directory is ok)

我全部LXC都係咁backup, 其中兩個LXC mount左device (Intel GPU同LIRC), backup去Synology NAS over SMB, so far冇問題

我LXC backup都係用suspend mode，都有assign GPU，但我好似又唔洗自己set tmp dir 。我見佢做親backup 都會自己寫堆哩落/var/tmp定/tmp，應該default就set左係咁

mount嘅係2隻passthrough咗嘅hdd(ntfs-3g)
試過snapshot suspend十幾分鐘自動interrupt
stop就十零秒搞掂

umount咗就全部冇問題

試下tick呢個？都唔work既話就唔知了

我唔係mount point嚟
係device passthrough /dev/sdx之後再喺lxc mount

其實正常做法係點

why not use mount point?

bind mount
https://pve.proxmox.com/wiki/Linux_Container#_bind_mount_points

proxmox storage點樣加現有嘅碟

即係喺proxmox host mount完隻碟
再喺lxc起個bind mount咁？

類似，你會喺host mount左隻碟先，個效能其本上係bare metal。你亦可以用Proxmox嘅ZFS功能，總之就係俾左隻storage Proxmox host mount 左先。

你應該可以成個disk俾(教學第一個option)，我自己就做開分folder咁俾(第二個option)。
第二個option之下，你可以首先喺host嘅mount point create一啲folder，例如
/mnt/disk1/lxc100storage

然後你可以喺host嘅shell打：
pct set 100 -mp0 /mnt/disk1/lxc100storage,mp=/mnt/storage

之後你著左隻LXC 就會見到多左後面個folder，喺LXC嘅/mnt/storage

題外話，
咁做嘅一個好處係，你可以用host本身去控制各個LXC對唔同folder/file嘅access。例如我有隻LXC係Samba用來upload相，我set左每晚半夜4點會將/mnt/disk1/lxc100storage裏面嘅.arw 檔chown 落第二個LXC嘅user ID，但用group去(類似，唔記得係咪exactly咁)俾返個read access原本lxc100 個root account (uid ~ 1001001)。咁我就整到個簡單嘅WORM storage，有咩要改我可以用host/boot起另一個LXC去搞。

冇理解錯3個option係
1. Storage Backed Mount Points
-> 而家隻碟上面有嘢唔想轉fs冇晒data
2. Bind Mount Points
-> 咁即係好似上面講host mount完lxc又bind mount, 睇落官方比較推薦呢個approach
3. Device Mount Points
-> 而家我用緊嘅approach, 佢個offers the same performance and a lot more features唔知講乜咩features

但2同3佢都講明The contents of bind mount points/device mount points are not backed up when using vzdump, 幾奇怪
再研究吓先感謝

佢backup個意思係你喺LXC入面見到嘅bind Mount folder，backup成隻LXC嗰陣佢Proxmox會ignore左佢，唔backup裏面嘅嘢。

你可以當bind mount某程度上類似Samba咁，不過實際上係單一個OS底下用folder/uid/gid 去分shared folder。

我一路都只用option 2，其他兩樣唔熟，等其他人補充

3都可以經mount cmd/fstab file較uid/gid/umask的 實用上應該唔大分別
反而著眼嘅係既然vzdump睇唔到mount point
點解我unmount咗先take到backup

見有台灣人1825+改野
用硬碟 : TOSHIBA 20TB

https://github.com/007revad/Synology_HDD_db/blob/main/2025_plus_models.md