Reboot machine running Node leads to issues with consensus

This issue is create to split different problems from INDY-1256

ozheregelya
Environment:
indy-node 1.3.364 (master)
AWS QALive pool (20 nodes) with 260,490 txns in ledger.

Steps to Reproduce:
1. Make sure that all nodes are in consensus and that all nodes have the same primary.
2. Reboot instance with primary node (Node10).

Actual Results:
ViewChange was not happened on one node (Node15). For the rest nodes primary was changed (to Node11). Pool is still in consensus.
After restart of problematic node (Node15) primary on this node was changed (to Node11), problematic node successfully completed catch-up and continued to work properly.

Logs: https://drive.google.com/file/d/1CBZYr2pMs1dapRqVbKSvn9RgVIb5quas/view?usp=sharing
Journalctl: https://drive.google.com/file/d/1_Nd4Exqe2Bhh1_OHh0l5c0y56obsGZMb/view?usp=sharing

Additional Information:
Similar case was tried by NIkita Zhigunenko and Vladimir Shishkin on another pools with stable version. ViewChange was successfully completed on all of the nodes in case of simple reboot.

zhigunenko.dsr
Environment:
indy-node 1.3.364 (master)
AWS QALive pool (20 nodes)

Steps to Reproduce:
1. Make sure that all nodes are in consensus and that all nodes have the same primary.
2. execute on primary node (Node11) this "execute sudo apt-get update && sudo apt-get upgrade && sudo reboot"
Actual results:
Ex-primary node cannot catch up new txns. Rest of nodes are consider ex-primary as unreachable
Expected results:
Ex-primary successfully catch up and write new txns.

Logs: https://drive.google.com/open?id=1YxjUlWcSxpXylq1Mu09dIY9kcyqJ5WV3
Thu Apr 12 07:31:20 2018 - approximate reboot time