David Turner 5eb33f2df4
Document the reduction in fault detection timeouts (#40200)
The new cluster coordination subsystem introduced in 7.0 will only keep an
unresponsive node in the cluster for 30 seconds, whereas in earlier versions it
might have remained in the cluster for 90 seconds. This commit adds a note to
the migration documentation to that effect.
2019-03-20 15:49:50 +00:00
..