Description of problem:
As per requestor, Grant Gainey
"
A lot of customers (see https://bugzilla.redhat.com/show_bug.cgi?id=42928) requested that RH add a way to mark errata as requiring a reboot.

RCM added the capability into errata-tool under https://bugzilla.redhat.com/show_bug.cgi?id=869677 , in March.

Satellite got a bunch of commits from SUSE that take advantage of it, which will be released as part of Satellite5.6.0 in September.

However - currently, Hosted doesn't recognize reboot_requested as a valid errata-keyword, and apparently just ignores it when supplied. This means there's no way for it to get from RCM down to Satellite.
"

In this bug, we need to fix the overlying issue of RHN Classic not recognizing 'reboot_requested' errata keyword.

Version-Release number of selected component (if applicable):
N/A

How reproducible:
Every time when reboot_suggested attribute on errata advisories

Steps to reproduce:
1. Push an errata which includes reboot_suggested attribute

  
Actual results:
RHN does not recognize 'reboot_requested' errata keyword, therefore systems that are applicable to receive an errata update that has the 'reboot_requested' errata keyword will not reboot as needed.

Expected results:
RHN recognizes 'reboot_requested' errata keyword so that systems that are Registered to RHN that are applicable to receive an Errata update with this attribute will reboot.

Additional info:

This was requested by Grant Gainey in INC0049406

Other information from Grant that was tabled in INC0049406,
"
The net, as I understand it, is that RCM errata-tooling knows to add the
"reboot_requested" keyword into an errata as of March. However, Hosted
doesn't recognize that as a 'legal' keyword, and so it gets stripped out
when an errata is pushed to Hosted.

Satellite has code/UI that responds to the presence of that keyword - but if
Hosted removes it, Satellite will never see it.

Somewhere in Hosted, in the errata-tool publishing flow, there is a
check for valid-keyword, and "reboot_requested" needs to be added so
that it makes it into Hosted, into the rhnErrataKeyword table. I don't
believe there are any other changes we need.

It looks like maybe the code here

rhn-hosted/backend/src/server/importlib/backend.py

is what's involved? See "def __processObjectCollection__" at ~619. I haven't
ever been in that code myself, so I'm kinda guessing here, sorry :(

Work with dgregor in RCM, it's the communication between errata-tool and Hosted
that needs to be modified.
"