Hot spare support in ScaleIO

Hot spare space is reserved across all nodes. By default it should reserve enough space for a single node failure within a Protection Domain. This can be increased if you desire to have added protection.

Reserving HS space across all nodes results to very fast rebuilds. Think of a 100 node Protection Domain, if one node goes down 99 will work to rebuild the data. It's a very parallel process.

Thanks first,could you help answer the following question?

how many nodes can fail at the same time ?

You can lose one node in each Protection Domain without losing access to data. If you define Fault Sets (which are groups of nodes inside a PD that do not protect each other) you can lose a Fault Set (which consists of several node) in each PD simultaneously.

Hello experts

About rebuild ,there are two questions that needs your answer.

Q1>

If one SDS server shutdown (or reboot),rebuild process will start immediately ? or it has one timeout time,for example 30 second ,rebuild will start after timeout

Q2>

Shutdown SDS server,and rebuild process finished. power on SDS server,will the devices in the SDS become available space ?

Thanks in advance.

Hi,

The answers to your questions are below:

Q1) Rebuild will occur immediately.

Q2) The chunks no longer in use will become free space.