diff --git a/osaf/services/saf/amf/amfd/si.cc b/osaf/services/saf/amf/amfd/si.cc
--- a/osaf/services/saf/amf/amfd/si.cc
+++ b/osaf/services/saf/amf/amfd/si.cc
@@ -874,13 +874,16 @@ static void si_admin_op_cb(SaImmOiHandle
* return an error.
*/
if (si->sg_of_si->sg_fsm_state != AVD_SG_FSM_STABLE) {
- LOG_WA("SI lock of %s failed, SG not stable",
objectName->value);
if ((si->sg_of_si->sg_fsm_state != AVD_SG_FSM_SI_OPER)
||
(si->saAmfSIAdminState !=
SA_AMF_ADMIN_SHUTTING_DOWN) ||
(adm_state != SA_AMF_ADMIN_LOCKED)) {
report_admin_op_error(immOiHandle, invocation,
SA_AIS_ERR_TRY_AGAIN, NULL,
- "'%s' other semantics...",
objectName->value);
+ "SI lock of %s failed, SG not
stable", objectName->value);
goto done;
+ } else {
+ report_admin_op_error(immOiHandle,
si->invocation, SA_AIS_ERR_INTERRUPT, NULL,
+ "'SI lock has been issued
'%s'", objectName->value);
+ si->invocation = 0;
}
}
Please check.
Thanks
-Nagu
> -----Original Message-----
> From: Hans Feldt [mailto:[email protected]]
> Sent: 05 September 2014 15:00
> To: Nagendra Kumar; Hans Nordebäck; Praveen Malviya
> Cc: [email protected]
> Subject: RE: [devel] [PATCH 1 of 1] amfd: respond to shutdown op if lock has
> been issued before it completes [#1021]
>
> Good catch! The proposal is fine with me.
> /Hans
>
> > -----Original Message-----
> > From: Nagendra Kumar [mailto:[email protected]]
> > Sent: den 5 september 2014 08:49
> > To: Hans Feldt; Hans Nordebäck; Praveen Malviya
> > Cc: [email protected]
> > Subject: RE: [devel] [PATCH 1 of 1] amfd: respond to shutdown op if
> > lock has been issued before it completes [#1021]
> >
> > > But bigger question is : If Specs defines this path/flow as valid
> > > and provides sanctity and doesn't consider it as an error. Why Amf should
> return an error?
> > [Hans] Returning prematurely from SHUTDOWN before LOCKED state has
> > been reached is an error to me. AMF can internally wait until the LOCK
> operation is done and then respond to both SHUTDOWN and LOCK.
> >
> > Well, I got answer from CLM B.4.2 Specs section 5.2.4 :
> >
> > SA_AIS_ERR_INTERRUPT - The SA_CLM_ADMIN_SHUTDOWN administrative
> > operation was superseded by a SA_CLM_ADMIN_LOCK or by a
> > SA_CLM_ADMIN_UNLOCK operation.
> >
> > So, do you think we can implement it, it is not mentioned anywhere in Amf ?
> >
> > Thanks
> > -Nagu
> >
> > > -----Original Message-----
> > > From: Hans Feldt [mailto:[email protected]]
> > > Sent: 04 September 2014 18:24
> > > To: Nagendra Kumar; Hans Nordebäck; Praveen Malviya
> > > Cc: [email protected]
> > > Subject: RE: [devel] [PATCH 1 of 1] amfd: respond to shutdown op if
> > > lock has been issued before it completes [#1021]
> > >
> > >
> > >
> > > > -----Original Message-----
> > > > From: Nagendra Kumar [mailto:[email protected]]
> > > > Sent: den 4 september 2014 10:24
> > > > To: Hans Feldt; Hans Nordebäck; Praveen Malviya
> > > > Cc: [email protected]
> > > > Subject: RE: [devel] [PATCH 1 of 1] amfd: respond to shutdown op
> > > > if lock has been issued before it completes [#1021]
> > > >
> > > > > So far as I have seen SMF never works on SI level.
> > > > We did the same thing with SU as part of ticket #582.
> > > >
> > > > > I vote for BAD_OPERATION.
> > > > As per Specs, this error is reported when target entity is locked
> > > > already. Here it
> > > is not locked.
> > > >
> > > > But bigger question is : If Specs defines this path/flow as valid
> > > > and provides sanctity and doesn't consider it as an error. Why Amf
> > > > should return
> > > an error?
> > > [Hans] Returning prematurely from SHUTDOWN before LOCKED state has
> > > been reached is an error to me. AMF can internally wait until the
> > > LOCK operation is done and then respond to both SHUTDOWN and LOCK.
> > >
> > > >
> > > > Thanks
> > > > -Nagu
> > > >
> > > > > -----Original Message-----
> > > > > From: Hans Feldt [mailto:[email protected]]
> > > > > Sent: 04 September 2014 12:53
> > > > > To: Nagendra Kumar; Hans Nordebäck; Praveen Malviya
> > > > > Cc: [email protected]
> > > > > Subject: RE: [devel] [PATCH 1 of 1] amfd: respond to shutdown op
> > > > > if lock has been issued before it completes [#1021]
> > > > >
> > > > > So far as I have seen SMF never works on SI level. The problem
> > > > > here is also more manually triggered, some operator gets tired
> > > > > of waiting for SHUTDOWN and goes for LOCK. SMF would not do that.
> > > > > I vote for BAD_OPERATION.
> > > > > /Hans
> > > > >
> > > > > > -----Original Message-----
> > > > > > From: Nagendra Kumar [mailto:[email protected]]
> > > > > > Sent: den 4 september 2014 08:44
> > > > > > To: Hans Feldt; Hans Nordebäck; Praveen Malviya
> > > > > > Cc: [email protected]
> > > > > > Subject: RE: [devel] [PATCH 1 of 1] amfd: respond to shutdown
> > > > > > op if lock has been issued before it completes [#1021]
> > > > > >
> > > > > > Hi Hans,
> > > > > > I think it is better to return OK because of following reasons:
> > > > > > 1. It avoids Smf upgrade failures. If Amf returns TIME_OUT,
> > > > > > Smf tries again(not sure about the implementation, but as per
> > > > > > Spec, it should try again)
> > > > > and Smf may get TRY_AGAIN or BAD_OP error return code. If Smf
> > > > > get BAD_OP, it will abort upgrade.
> > > > > > 2. If Smf accepts OK in this case and then run another command
> > > > > > like unlock, then worst case, it will get TRY_AGAIN and this
> > > > > > is one of above
> > > case.
> > > > > >
> > > > > > Thanks
> > > > > > -Nagu
> > > > > >
> > > > > > > -----Original Message-----
> > > > > > > From: Hans Feldt [mailto:[email protected]]
> > > > > > > Sent: 03 September 2014 17:54
> > > > > > > To: Hans Feldt; Nagendra Kumar; Hans Nordebäck; Praveen
> > > > > > > Malviya
> > > > > > > Cc: [email protected]
> > > > > > > Subject: RE: [devel] [PATCH 1 of 1] amfd: respond to
> > > > > > > shutdown op if lock has been issued before it completes
> > > > > > > [#1021]
> > > > > > >
> > > > > > > Or from 9.4.6:
> > > > > > >
> > > > > > > SA_AIS_ERR_BAD_OPERATION - The operation was not successful
> > > > > > > because the target entity is locked or locked for instantiation.
> > > > > > >
> > > > > > > > -----Original Message-----
> > > > > > > > From: Hans Feldt [mailto:[email protected]]
> > > > > > > > Sent: den 3 september 2014 14:15
> > > > > > > > To: [email protected]; Hans Nordebäck;
> > > > > > > > [email protected]
> > > > > > > > Cc: [email protected]
> > > > > > > > Subject: Re: [devel] [PATCH 1 of 1] amfd: respond to
> > > > > > > > shutdown op if lock has been issued before it completes
> > > > > > > > [#1021]
> > > > > > > >
> > > > > > > > Do we think OK is an appropriate error code here?
> > > > > > > > Maybe TIMEOUT?
> > > > > > > > /Hans
> > > > > > > >
> > > > > > > > > -----Original Message-----
> > > > > > > > > From: [email protected]
> > > > > > > > > [mailto:[email protected]]
> > > > > > > > > Sent: den 3 september 2014 13:45
> > > > > > > > > To: Hans Feldt; Hans Nordebäck;
> > > > > > > > > [email protected]
> > > > > > > > > Cc: [email protected]
> > > > > > > > > Subject: [PATCH 1 of 1] amfd: respond to shutdown op if
> > > > > > > > > lock has been issued before it completes [#1021]
> > > > > > > > >
> > > > > > > > > osaf/services/saf/amf/amfd/si.cc | 4 ++++
> > > > > > > > > 1 files changed, 4 insertions(+), 0 deletions(-)
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > If shutdown op is undergoing, lock can be issued.
> > > > > > > > > If lock is isued, Amf is not responding to shutdown op.
> > > > > > > > > Amf should ideally respond to shutdown op immediately
> > > > > > > > > after it receives lock op. And after lock op completes,
> > > > > > > > > then Amf should respond to lock op.
> > > > > > > > >
> > > > > > > > > diff --git a/osaf/services/saf/amf/amfd/si.cc
> > > > > > > > > b/osaf/services/saf/amf/amfd/si.cc
> > > > > > > > > --- a/osaf/services/saf/amf/amfd/si.cc
> > > > > > > > > +++ b/osaf/services/saf/amf/amfd/si.cc
> > > > > > > > > @@ -881,6 +881,10 @@ static void
> > > si_admin_op_cb(SaImmOiHandle
> > > > > > > > >
> > > report_admin_op_error(immOiHandle,
> > > > > > > invocation, SA_AIS_ERR_TRY_AGAIN, NULL,
> > > > > > > > > "'%s' other
> > > semantics...",
> > > > > > > objectName->value);
> > > > > > > > > goto done;
> > > > > > > > > + } else {
> > > > > > > > > + rc = SA_AIS_OK;
> > > > > > > > > +
> > > > > > > avd_saImmOiAdminOperationResult(immOiHandle,
> > > > > > > si->invocation, rc);
> > > > > > > > > + si->invocation = 0;
> > > > > > > > > }
> > > > > > > > > }
> > > > > > > > >
> > > > > > > >
> > > > > > > > ----------------------------------------------------------
> > > > > > > > ----
> > > > > > > > ----
> > > > > > > > ----
> > > > > > > > --------
> > > > > > > > Slashdot TV.
> > > > > > > > Video for Nerds. Stuff that matters.
> > > > > > > > http://tv.slashdot.org/
> > > > > > > > _______________________________________________
> > > > > > > > Opensaf-devel mailing list
> > > > > > > > [email protected]
> > > > > > > > https://lists.sourceforge.net/lists/listinfo/opensaf-devel
------------------------------------------------------------------------------
Want excitement?
Manually upgrade your production database.
When you want reliability, choose Perforce
Perforce version control. Predictably reliable.
http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk
_______________________________________________
Opensaf-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensaf-devel
