> I recently changed rhnServerNetwork due IPv6 and it was quite straight.
>
> Now I focused on rhnServerInterface and I want to give you chance to
> argue with me, before I do the change. :)
>
> Today the table looks like:
> http://spacewalk.redhat.com/documentation/schema-doc/table-RHNSERVERNETINTE
> RFACE.html
>
> It has columns:
> * SERVER_ID NUMBER(38)
> * NAME VARCHAR2(32)
> IP_ADDR VARCHAR2(64)
> NETMASK VARCHAR2(64)
> BROADCAST VARCHAR2(64)
> HW_ADDR VARCHAR2(18)
> MODULE VARCHAR2(128)
> CREATED DATE
> MODIFIED DATE
>
> "*" marks primary key
>
> Previusly clients send over XMLRPC this data:
>
> intDict[interface] = {'hwaddr':hwaddr,
> 'ipaddr':ipaddr,
> 'netmask':netmask,
> 'broadcast':broadcast,
> 'module': module
> }
>
> for some time (since 2010-11-20) it send:
>
> intDict[interface] = {'hwaddr':hwaddr,
> 'ipaddr':ipaddr,
> 'netmask':netmask,
> 'broadcast':broadcast,
> 'module': module,
> 'ipv6': ip6_list}
>
> where ip6_list is:
>
> ip6_list.append({
> 'scope': ip6.scope,
> 'addr': ip6.address,
> 'netmask': ip6.netmask
> })
>
> The ipv6 key/value pair is ignored by server for now.
> The IPv6 stuff will not fit into current table so I come with following
> solution:
>
> Simplify table rhnServerInterface to columns:
> * SERVER_ID NUMBER(38)
> * NAME VARCHAR2(32)
> HW_ADDR VARCHAR2(18)
> MODULE VARCHAR2(128)
> CREATED DATE
> MODIFIED DATE
>
> Create new table rhnServerInterfaceIPv4 with columns:
> * SERVER_ID NUMBER(38)
> * NAME VARCHAR2(32)
> IP_ADDR VARCHAR2(64)
> NETMASK VARCHAR2(64)
> BROADCAST VARCHAR2(64)
> CREATED DATE
> MODIFIED DATE
> Migration of existing data from rhnServerInterface to this new table is
> easy and should not be problem.
>
> Create new table rhnServerInterfaceIPv6 with columns:
> * SERVER_ID NUMBER(38)
> * NAME VARCHAR2(32)
> SCOPE VARCHAR2(19)
> ADDR VARCHAR2(45)
> NETMASK VARCHAR2(49)
> CREATED DATE
> MODIFIED DATE
>
> SCOPE can be (according rfc2373):
> scop is a 4-bit multicast scope value used to limit the scope of
> the multicast group. The values are:
>
> 0 reserved
> 1 node-local scope
> 2 link-local scope
> 3 (unassigned)
> 4 (unassigned)
> 5 site-local scope
> 6 (unassigned)
> 7 (unassigned)
> 8 organization-local scope
> 9 (unassigned)
> A (unassigned)
> B (unassigned)
> C (unassigned)
>
> longest is organization-local with 19 chars, but I seen more different
> strings already ("local", "universe" etc.).
> So I tend to think that more general VARCHAR(128) would be better here.
>
> ADDR - longest IPv6 address can be 45 character wide. While we may
> normalize the address and make it shorter. Question is whether we want
> it normalize here or exactly as we get it over wire.
>
> NETMASK - it is addr plus "/128" as longest prefix.
Today, we discussed another approach with Miroslav:
rhnServerNetInterface:
ID <-- this will be table's primary key
SERVER_ID
NAME
HW_ADDR
MODULE
CREATED
MODIFIED
rhnServerNetAddress4 (or some other / better name):
INTERFACE_ID <-- this will reference rhnServerNetInterface(id)
ADDRESS
NETMASK
BROADCAST
rhnServerNetAddress6 (or some other / better name):
INTERFACE_ID <-- this will reference rhnServerNetInterface(id)
ADDRESS
SCOPE
NETMASK
Plus we may have several other constraints in place, such as:
unique rhnServerNetAddress4(interface_id, address)
unique rhnServerNetAddress6(interface_id, address, scope)
Thoughts, concerns, suggestions?
-Milan Zázrivec
_______________________________________________
Spacewalk-devel mailing list
Spacewalk-devel@redhat.com
https://www.redhat.com/mailman/listinfo/spacewalk-devel
