rdbms: make LoadBalancer::waitForAll() better respect the timeout
authorAaron Schulz <aschulz@wikimedia.org>
Tue, 16 Jan 2018 16:58:59 +0000 (17:58 +0100)
committerAaron Schulz <aschulz@wikimedia.org>
Sun, 21 Jan 2018 05:31:15 +0000 (05:31 +0000)
commitcd58ffbaf58d5676b7405cf17c0dc32c51b3da2c
tree1d7aa9c83621ed42ae6c2f3134c1e18c7937619d
parent79d298e37f8aeb42ccf43155a5be3c16b7688901
rdbms: make LoadBalancer::waitForAll() better respect the timeout

If several replicas (or even all) had replication stuck, then the
timeout would happen for each server, one after another.

Change-Id: Id5431360b9cde7e5dc0115a1f41b9903003f47c4
includes/libs/rdbms/loadbalancer/LoadBalancer.php