Log for
電訊台
攞esxi行type 1好定係linux行qemu-kvm好
KVM 其實都可以算 Type1 Hypervisor
又有 config file 唔洗手打 qemu-kvm 啲指令
好啦就當我打唔起 用proxmox我不如用opennebula/openstack
尤其是 Openstack Electron 你係屋企根本冇需要
同埋 Openstack 會逼住你做一次 NAT 然後又要 Overlay Network
其實internet of things對networking大唔大影響？
accelerating the exhausting of ipv4
because people are stupid and forgot what is ipv6
IPv4 isn't really exhausted
It's just IANA havr finished assigned all the blocks to the RIRs
Looking at my IPv6 blocks
My
/32 IPv6 block is literally free
APNIC giveaway present when you request IPv4 subnets
APNIC is hard to request IP now
CentOS 4.5, 150G hd, fsck 左5個鐘都沒完😓
老舊電訊系統，佢地話升級唔到（其實係佢地唔想花時間攪呢套舊野）
Single user mode fsck緊，scan完就無事，只係時間長
有用veeam backup呀，但backup都係已經炒左既os，一開機都迫你fsck
即係個file system一直炒左好耐，只要你一reboot就會發現
New link to Japan is completed.
關西............關西援交系列幾好睇............
我發現一樣野, 我加大個VM resource, fsck快好多, 早知早D加大.....>3<
如果有問題我道歉, 因為我之前咁樣都好似無咩問題, 是咪有野改左? 我以後唔亂講野.
（我最近搬屋可能又要幫新屋買返啲 networking gear）
不過 Telegram 可以 restrict 到 readonly
玩女唔同骨 , Agent CC 返大陸玩女，玩完 cable 就去
Migrate個VM去個SSD Storage度, 加到16CPU, 16G Ram, fsck快左100倍(不過仲係算好慢)
我沒有說你是婊,只是如果你是一個interface,那你的access list一定是permit any
另一部機resource夠, 係個度fsck快好多.
我有諗過好唔好整部機放喺隻 7604 隔離做 route server
係裏面揀哂啲 best route 先畀翻部 cisco 做 routing
突然變 IP 都可以 restore connection
Which route server is better ?
我諗住用 bird，因為捷克
nix.cz 同埋其他 IX 都有用緊
I also want to build a route server for looking glass
bird is a server or is a OS ?
因為你 google 都冇乜教你點樣 set bird
Any windows base route server ?
really ?? what is the server name ?
Or just deploy a Cisco to be a route server.
反而係rras 個sstp cert error 係關iis 事果下有d on9
其中一個問題係 ICMP response 個 source IP 唔係跟 RFC spec 用 Interface IP 係跟 default route 黎搵 interface IP
除非 patch kernel，只係 patch 幾行加多個 if 咁
IRQ 啊 網卡 driver 啊 queues 啊 嗰啲
跟住你會發現有啲問題真係要賴落個 Kernel 到
（就好似 ICMP response 唔跟 spec 呢件事）