By us creating a problem I mena this:
SAF has declared a principle that runtime data is not controlled by
transactions/CCBs, except when runtime
object are deleted as children of config objects in a CCB.
This iconflicts with the SMF principle that all objects (including all runtime
state ?!) shall be recreated by a rollback.
Well who created that SMF principle?
Was it SAF that created two contradictory principles ?
Or did we create an SMF principle that contradicts with the first (asymetric)
SAF principle.
I think (can not be sure) that SAF intended that an uppgrade rollback should
recreate only the connfig state.
Any re-creation (or not) of runtime state would be up to the OI.
In other words I think SAF did not create contradictory principles here, but
can not be sure.
Thus I think we created the contradiction.
In the end, insisting on re-creation of all runtime state will be impossible,
whatecer we decide.
So there is no solution to our created problem.
I hate creating problems that can not be solved.
/AndersBj
________________________________
From: Anders Björnerstedt [mailto:[email protected]]
Sent: den 28 januari 2014 09:28
To: [opensaf:tickets] ; [email protected]
Subject: Re: [tickets] [opensaf:tickets] #723 SMF: deletion of object shall be
possible if sub tree contain runtime children only
I have to state that I dont agree with the all encompassing SMF principle that
all objects (including all runtime state) must be recreated
by an SMF rollback.
Runtime data is not directly controlled by SMF but by the OIs, in this case
mainly by the OIs that manage the config objects that
are the actual targets of the upgrade or rollback. It is up to these config
OIs to maintain the runtime state in *relation* to the config
state. The SMF should focus on upgrading or rollbacking too/from any give
config state and not at all try to recreate the runtime state.
The exception is deletion of runtime objects that are children of config
objects to be removed.
Such objects are typically removed *indirectly* by a CCB. An applications sets
admin-owner for the subtree to be removed
whjich will cover also any unknown config or runtime objects.
To maintain transactional consistency, all related objects should preferrably
be created and removed in the same CCB.
Deletes of runtime objects that are children of config objects deleted in the
transaction will also be deleted (or NOT)
atomically with the transaction.
Creates or updates of runtime objects on the other hand are never atomic with
any transaction since they are not covered
by transactions. This is where the asymetry enters the picture and it is an
asymetry that SAF has designed to be there.
Thus is is impossible to avoid.
So this "decision" is just a decision to create a problem there really did not
exist a problem untill we decided that we
should create a problem.
/AndersBj
________________________________
From: "Ingvar Bergström" [mailto:[email protected]]
Sent: den 28 januari 2014 07:45
To: [email protected]
Subject: [tickets] [opensaf:tickets] #723 SMF: deletion of object shall be
possible if sub tree contain runtime children only
Current decision is to solve this by some other mechanism not yet decided.
Deletion of subtree runtime objects will break the SMF principle to be able to
recreate all objects at rollback since the runtime objects cannot be recreated
by SMF.
Workaround is to call e.g. command "immcfg" from the campaign to manipulate
such objects when needed.
________________________________
[tickets:#723]<http://sourceforge.net/p/opensaf/tickets/723/> SMF: deletion of
object shall be possible if sub tree contain runtime children only
Status: wontfix
Created: Thu Jan 16, 2014 07:41 AM UTC by Ingvar Bergström
Last Updated: Thu Jan 16, 2014 10:06 AM UTC
Owner: Ingvar Bergström
SMF normally requires objects to be deleted one by one bottom up i.e. an object
to be deleted must not have any children of any class, runtime or config.
An exception is now introduced:
It shall be possible to delete an object and its sub tree objects in one delete
operation if the children objects are all of runtime category only.
________________________________
Sent from sourceforge.net because [email protected] is
subscribed to https://sourceforge.net/p/opensaf/tickets/
To unsubscribe from further messages, a project admin can change settings at
https://sourceforge.net/p/opensaf/admin/tickets/options. Or, if this is a
mailing list, you can unsubscribe from the mailing list.
------------------------------------------------------------------------------
WatchGuard Dimension instantly turns raw network data into actionable
security intelligence. It gives you real-time visual feedback on key
security issues and trends. Skip the complicated setup - simply import
a virtual appliance and go from zero to informed in seconds.
http://pubads.g.doubleclick.net/gampad/clk?id=123612991&iu=/4140/ostg.clktrk
_______________________________________________
Opensaf-tickets mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensaf-tickets
