[autotest] Use RetryingAFE for RPCs to shards
RPC call can fail by timeout time to time.
If a RPC sent to a shard from a master fails, DB inconsistency can happen.
Let's use RetryingAFE when sending RPCs to shards.
BUG=chromium:496452
TEST=puppylab. Try to add/remove labels for a sharded host.
It uses RPCs that calls RetryingAFE.
Unit Tests.
frontend.afe.rpc_interface_unittest
frontend.afe.site_rpc_interface_unittest
DEPLOY=apache
Change-Id: I23853c8c0f76d31bb10493759cafdbdc6c003b3f
Reviewed-on: https://chromium-review.googlesource.com/276248
Reviewed-by: Mungyung Ryu <mkryu@google.com>
Tested-by: Mungyung Ryu <mkryu@google.com>
Commit-Queue: Mungyung Ryu <mkryu@google.com>
2 files changed