1
msg7086 2016-06-09 06:35:03 +08:00
说实话,真正建站并且要求 uptime 的人不应该盯着这个价位看。
|
2
laoyur OP @msg7086 就算是做个人博客,要求便宜的同时也希望尽量稳定,用 RamNode 就是奔着这个目标去的,事实上一直以来 RamNode 的 uptime 确实不低。
他们号称已经给客户发邮件了,不知道什么原因,我的 email 就是收不到,但 ticket 的邮件都能收到的 找客服后看到了他们发的邮件内容: "This message is to clients on LACKVM1. Unfortunately, we have been unable to restore service to this node. We encountered the same RAID card fault state issue today after installing a brand new motherboard and RAID card. The first motherboard that we tried was actually DOA, which added to the delay. We eventually removed the SSDs, installed a second brand new RAID card, and were able to see parts of the partition table. While the host OS is not functional, there is still a possibility of recovering some KVM data. We are overnighting a new server to which we will try to transfer any available data as soon as possible. Any data we recover will likely have some level of corruption, but we will give you access to retrieve it as desired. We will send another email once that is done. In the meantime, anyone who still needs a new VPS can send us a request via support ticket. We again apologize for the inconvenience caused by this very rare event. We appreciate your continued business and support." 看来得彻底死心了,我还是恢复旧数据去吧。最近的几篇博客丢了,我是不是该回忆一下有没有比较重要的,从 Google 快照里给找回来(真滑稽)。 此事虽然让我略失望,但目前来说 RamNode 还算是相对可靠的,以后千万要多做备份才是正道。 |
3
msg7086 2016-06-09 22:41:57 +08:00 1
成本与灾备有直接的关联。
比如 ramnode 有 5 个机房,任何一个机房的机器挂掉,你都得直接跑到机房去处理。飞机票要多少钱。 再要么每个机房都留一个工程师(等于每个机房每月额外开销 10000 美元)。 花更多的钱就意味着有更多的工程师留在机房,有更多的备件会留在机房,有更多的人手去处理事故。 uptime 高不稀奇,有技术又不超兽的话是可以轻松做到的。 (我随便拿台笔记本跑 Windows 都可以做到 uptime 一年) 真正的差别则从发生事故的时候开始。 |