bugCERTI - Bugs: bug #23746, RTIG should detect RTIA crash and...

 
 

bug #23746: RTIG should detect RTIA crash and destroy its object instances

Submitter:  Petr Gotthard <gotthardp>
Submitted:  Tue 01 Jul 2008 08:06:54 PM UTC
   
 
Category:  CERTI Severity:  * 4 - Important
Item Group:  Software error Status:  None
Privacy:  Public Assigned to:  None
Open/Closed:  Open Release:  * 3.3.0
Reproducibility:  None Fixed Release:  * Unknown
* Mandatory Fields

Add a New Comment Rich Markup
   

Wed 02 Jul 2008 08:46:35 AM UTC, comment #3: 

I know that RTIA-RTIG is out of scope because it's
a specific design choice, but I think may be there
could be some rationale about expected behavior if
a federate

1- unexpectedly quit the federation
2- resign from federation without unpublishing/deleting Object
   first


In the resign case, we have some indication
see 4.5  (HLA 1.3 document)

 Resign Federation Execution service shall accept an action argument that directs the RTI to perform zero or more of the following actions:
— Release all owned instance attributes for future ownership acquisition. This shall place the instance attributes
     into an unowned state (implying that their values are not being updated), which shall make them eligible for
     ownership by another federate. See Clause 7., Ownership management for a more detailed description.
— Delete all object instances for which the federate has that privilege (implied invocation of the Delete Object
     Instance service).

So a Federate may voluntarily resign from federation
without deleting  object instance or even set instance
attribute to unowned.

I think this would lead to the same state as the crash case?

This does not give us the answer but, I don't really know
what we should do in case of crash.

Eric NOULARD <erk>
Group administrator
Wed 02 Jul 2008 08:30:35 AM UTC, comment #2: 

The RTIA-RTIG communication is not in the scope of the HLA standards, so even failures of the RTIA-RTIG communication are probably also out of scope. I don't expect it's described in the standards. (After a quick search in IEEE 1516 I haven't found anything.)

Petr Gotthard <gotthardp>
Group Member
Wed 02 Jul 2008 08:16:28 AM UTC, comment #1: 

I don't know if we clearly thought
about 'failure' mode in CERTI.

Did you read something about failure secnarii
in HLA documents?

I add Pierre as cc for comment.

Eric NOULARD <erk>
Group administrator
Tue 01 Jul 2008 08:06:54 PM UTC, original submission:  

When a federate/RTIA crashes, the published objects remain in the RTIG, so the next simulation run throws an exception. This causes a lot of inconvenience in unstable environments.

Is there a reason why RTIG should not destroy all objects owned by an RTIA that just has disconnected/crashed without unpublishing? (I think that ownership management should "assure" that after a given RTIA crashes, no one will ever be able to change/destroy its object instances.)

Petr Gotthard <gotthardp>
Group Member

 

(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

Attach Files:
   
   
Comment:
   

No files currently attached

 

Carbon-Copy List
  • -email is unavailable- added by erk (Posted a comment)
  • -email is unavailable- added by erk (Needs comments)
  • -email is unavailable- added by gotthardp (Submitted the item)
  • -email is unavailable- added by gotthardp
  •  

    Follow 3 latest changes.

    Date Changed by Updated Field Previous Value => Replaced by
    2008-07-02 erk Dependencies- bugs #23725 is dependent
    2008-07-02 erk Carbon-Copy- Added siron
    2008-07-01 gotthardp Carbon-Copy- Added -email is unavailable-

    Back to the top

    Powered by Savane 3.13-4b48.
    Corresponding source code