Author: Glori

按照之前配置成功的筆記進行了配置(kelch.ini)，但是啓動後任然一直報錯：

“数据库没解锁” — 從字面一直怀疑是需要解锁或开通某些授权才能使用，但是实际在Kelch工程师指导下，配置OLEDB方式連接Oracle（只保存一套連接設置[REMOTEOLEDBORA]），启动也不报错：

之前這套設置也有試過，但一直是設置ConnType=2(根據注釋的提示–實際是誤導了的)。这里虽然启用的是OLEDB ORA，但实际配置ConnType=1 (ODBC – for Oracle)才能通過–猜想是接口程序默认Oracle始终使用ConnType=1

但这样配置后顺利启动后，任然无法得到TDM的对刀界面:

等待咨询Kelch德国工程师排查问题，未完待续。。。

1. 安裝OMV時如果沒有可用的網絡，OMV會讓你指定對應的Rounte IP用於聯網。將VirtualBox的網絡設置為NAT后順利繼續安裝。從上圖可知安裝時會抓取或更新安裝的組件包。
2. 雖然定位不同，OMV和Proxmox的界面風格完全不同。OMV更傾向於家用，簡潔，直觀，不需要多少文檔，開箱能用。
3. 是不是該嘗試下FreeNas(TrueNas)來和OMV比較一下？

在准备正式启用Proxmox之前，需要对ZFS, Raid-1，备份等操作作验证操作时，可以使用Virtualbox工具熟悉与测试。

为了简化设置， 所有的网卡都可以在VirtbualBox设置为”Host Only Adapter”。以下是Virtualbox的网络连通性表（源自Virtualbox Manual）：

如果是单PVE+单PBS(Proxmox Backup Server)的测试场景，VM1<->VM2在Host-only模式下完全够用。

Pre-conditions:

/dev/sda -- existing good disk
/dev/sdb -- exiting bad disk to be replaced
/dev/sdc -- new disk

Reference

https://hosting-tutorials.co.uk/tutorials/linux/replacing-a-zfs-root-disk-with-proxmox
https://pve.proxmox.com/wiki/Host_Bootloader#sysboot_proxmox_boot_setup
https://pve.proxmox.com/wiki/ZFS_on_Linux#_installation_as_root_file_system

Steps

Replace disk on root zfs raid-1

zpool offline rpool /dev/disk/by-id/ata-VBOX_HARDDISK_VB8b4872d1-c28eed93 # Offline the bad disk
fdisk -l # Check new disk is presenting
sgdisk /dev/sda -R /dev/sdc # Copy partition table from sda to sdc
sgdisk -G /dev/sdc	# Same - Randomize the disk guids  #>sgdisk --randomize-guids /dev/sdc	# Randomize the disk guids
#grub-install /dev/sdc --NOT work on new Proxmox version, check following new command(not verified)
#check partitions of sdc first -- there're 3 partitions now
proxmox-boot-tool format /dev/sdc2
proxmox-boot-tool init /dev/sdc2 # to be verified
#proxmox-boot-tool init /dev/sdc2 grub # may be work on old version of proxmox
proxmox-boot-tool refresh # add/update new UUID to /etc/kernel/proxmox-boot-uuids
proxmox-boot-tool clean # clean unused UUID in /etc/kernel/proxmox-boot-uuids
reboot	#make sure system can boot normally
# Resize disk(parted) at this point
zpool replace -f rpool /dev/disk/by-id/ata-VBOX_HARDDISK_VB8b4872d1-c28eed93 /dev/disk/by-id/ata-VBOX_HARDDISK_VBdb5f453c-7a4145b7-part3 # replace bad disk with new disk's partition 3, DO NOT rebook before resilvering was done.

# Reference:
# Resize disk on root zfs raid-1
# 1st Try
zpool offline rpool ata-VBOX_HARDDISK_VB8b4872d1-c28eed93
fdisk /dev/sdc
	d
	3		#delete /dev/sdc3 partition
	n
	3
	(Enter)
	(Enter)	#create new partition /dev/sdc3 with all rest size
	t
	157		#change sdc3 partition type to "Solaris /usr & Apple ZFS"
	p
	w
zpool online rpool ata-VBOX_HARDDISK_VB8b4872d1-c28eed93
#wait until reslizer done and do the same on another raid-1 disk.
# 2nd Try
# Maybe this is the correct way to resize the disk:
# https://blog.doussan.info/posts/how-to-expand-a-zfs-pool-partitions/
parted /dev/sda
	print
	resizepart 3 100%
	print
	q

Notes:
1. Feel like bootable ZFS raid-1 (zpool) should contains 3 partitions: sdc1 sdc2 is used for boot, sdc3 is actually data partition.
2. Additional commands to perform for partition table copy and bootable partition setup
3. There's no easy way to replace the bootable raid-1/mirror or on GUI -- maybe I can try another open source solution(OMV) to see if that's better?

Note or Summary

1. Feel like bootable ZFS raid-1 (zpool) should contain 3 partitions: such as sdc1 sdc2 is used for boot, sdc3 is actually data partition.
2. Additional commands to perform for partition table copy and bootable partition setup
3. There’s no easy way to replace the bootable raid-1/mirror or on GUI — maybe I can try another open source solution(OMV) to see if that’s better?