"restart required : yes" is never set to "no" after network/VPC restart

[andrijapanicsb]

ISSUE TYPE

• Improvement request
  (perhaps a bug?)

COMPONENT NAME

https://github.com/apache/cloudstack/blob/master/server/src/main/java/com/cloud/network/vpc/VpcManagerImpl.java#L1719-L1764

CLOUDSTACK VERSION

4.13

CONFIGURATION/SUMMARY/ETC

When a network is showing as "Restart required: yes" in the GUI (field "restart_required"="1" in cloud.networks table) - there is no way to "fix" this (set it back to "no")

Confirmed by manually setting "1" in the DB for multiple Shared, non-VPC Isolated networks and VPC Isolated networks.
Restart (no cleanup) doesn't work
Restart (with cleanup) doesn't work
The only way seems to be to hack the DB to set

The code provided above does set this for the VPC itself, not for it's networks.

[yadvr]

#3803

[andrijapanicsb]

Not completely addressed by #3803 - when you restart VPC with cleanup, it does NOT mark the field as 0 in DB (per @weizhouapache testing), thus all VPC networks still show as requiring a restart. Thus reopening the issue.

[Spaceman1984]

I have tested on 4.13.1 with isolated, shared and L2 networks and a VPC network. In all instances, the restart required flag was reset with and without cleanup. @andrijapanicsb this doesn't seem like an issue anymore.

[andrijapanicsb]

It can't be gone by its own? See this comment from Wei #3803 (comment)

Can you trace any commit that would have fixed the thing?

[Spaceman1984]

Ok @andrijapanicsb, I didn't test restarting the VPC. My test only included restarting the network in the VPC. I'll look into it.

[Spaceman1984]

Restarting the VPC did not restart the network. I'll work on a fix.