[ClusterLabs] Antw: Re: Question on sharing data with DRDB

[Valentin Vidic]

On Thu, Mar 21, 2019 at 08:00:05AM +0100, Ulrich Windl wrote:
> Actually it makes no difference to a non-clustered local disk: If the buffers
> are not flushed, data can get lost if there is a power failure. If you use sync
> writes, the data should be on disk, and I guess with DRBD the data should be on
> the remote site also.

Right, but I was thinking more on the DRBD layer and how it avoids
SplitBrain in that case.

  16.3. The Activity Log
  16.3.1. Purpose

  During a write operation DRBD forwards the write operation to the local
  backing block device, but also sends the data block over the network.
  These two actions occur, for all practical purposes, simultaneously.

Since local and remote writes execute in parallel it is possible that
one of those fails and DRBD needs to undo the other change somehow.
For the local writes it keeps the activity log with recent writes not
yet acknowledged by the remote peer.

-- 
Valentin