作者: qxwo 时间: 2015-7-22 14:32
This email is being sent to all customers on VPS Node Chord.
For the past few weeks, this server has been suffering from stability and performance troubles, which has been caused by a sub-par RAID configuration. To put it simply, instead of running on the normal 4 hard drives, we've been running on just 3. One of the hard drives was marked as failed, so we promptly got on to our data center to replace it. The drive was replaced, and the RAID array was still marked as Degraded, or "incomplete" and the replaced drive was not showing up in the BIOS. We have been working with our data center for quite some time on possible solutions, though they were not entirely sure what to suggest. After going through various support queues and talking with different people, it was decided that the backplane and breakout cable would be replaced. We (Ethernet Servers) were willing to give this a go, and schedule downtime accordingly. Over the past 24 hours, this maintenance has been becoming more and more unstable. It crashed last night at 8PM BST, and we managed to bring it back online. It then went down at around 1AM, just a few hours later. At this point, we asked our data center technicians to replace the suspected faulty components, and they did this. This meant we weren't able to schedule maintenance. The server has come back online, and the fourth drive is now being detected. For the last few hours, the RAID array has been rebuilding and when the server comes back online, it should be fully up-to-date strength with all four drives running in the normal RAID 10 array.
We are very sorry for the troubles caused. This problem has taken a lot longer to diagnose than we would have liked.
We estimate that all services will be back online within 1 hour.