IPv6 Distributed Cloud: sudo cmd return very slow on subcloud worker nodes

[Peng Peng]

Brief Description
-----------------
sudo cmd return in more than 2 mins on subcloud compute nodes, with msg "sudo: ldap_sasl_bind_s(): Can't contact LDAP server"

Severity
--------
Major

Steps to Reproduce
------------------
Install subcloud with compute node
run sudo cmd on compute

Expected Behavior
------------------
return in a reasonable time

Actual Behavior
----------------
return very slow with error msg (minutes)

Reproducibility
---------------
Reproducible

System Configuration
--------------------
DC system

Lab-name: WCP_80-85

Branch/Pull Time/Commit
-----------------------
stx3.0 designer load as of 2020-01-28_18-49-15

Last Pass
---------
unknow

Timestamp/Logs
--------------
 sudo ls
sudo: ldap_sasl_bind_s(): Can't contact LDAP server
Password:
compute-0:~$

Test Activity
-------------
Regression Testing

[Peng Peng (ppeng)]

Log @
https://files.starlingx.kube.cengn.ca/launchpad/1861424

[Yang Liu (yliu12)]

Comments from Andy Ning:
When sudo on any nodes of the subcloud, the system authentication will check against ldap as part of the process. ldap is running on system controller so there must be a route to reach the system controller.

Workaround is to add a route on worker nodes, example:
system host-route-add compute-0 mgmt0 fd01:1::0 64 fd01:2::1

[Ghada Khalil (gkhalil)]

stx.4.0 / medium priority - workaround exists; only applicable to non-AIO subclouds

[Bart Wensley (bartwensley)]

This is not a bug. The installation instructions for DC require the user to add a route for each host in the subcloud to reach the system controller.