Syncthing start timeout is too short for managers with a lot of data

Description

If the manager contains several hundred MB of data, waiting 20 seconds for syncthing to replicate it during `cfy cluster join` is not enough.

This is somewhat random, because we only wait synchronously to replicate the restservice config files, and those might or might not be replicated first.

Increasing the timeout is the simplest way to work around this, but perhaps we can not require these config files synchronously at all.

Steps to Reproduce

Environment:
OS (CLI), HA cluster, cloud provider
------------------------------------

Steps to reproduce:
------------------
1. Start a cluster
2. Put lots of data into a replicated directory, eg. /opt/manager/resources or /opt/mgmtworker/env/plugins
3. Join another node to the cluster

Expected result:
---------------
The node joins the cluster successfully within about 1 minute

Actual result:
-------------
Occasional timeout. (if the node doesn't join in 2 minutes, then it won't join at all. Errors are logged in the joining node's /var/log/cloudify/cloudify-cluster.log)

Why Propose Close?

None

Status

Assignee

Łukasz Maksymczuk

Reporter

Łukasz Maksymczuk

Labels

None

Severity

Medium

Target Version

4.4

Premium Only

yes

Found In Version

4.4

QA Owner

Lital Hamami

Bug Type

legacy bug

Customer Encountered

Yes

Customer Name

FICO

Release Notes

yes

Priority

None

Epic Link

Sprint

None

Fix versions

Priority

Unprioritized
Configure