Ack from me.
Good work!
I have not tested the patch, only reviewed the code.
Most important is to test that this patch does not break anything (regression
testing).
Hopefully Neelakanta can do some of that.
The actual functionality provided by the patch is a better/clearer error
indication
tow
osaf/services/saf/immsv/immnd/immnd_evt.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/osaf/services/saf/immsv/immnd/immnd_evt.c
b/osaf/services/saf/immsv/immnd/immnd_evt.c
--- a/osaf/services/saf/immsv/immnd/immnd_evt.c
+++ b/osaf/services/saf/immsv/immnd/immnd_evt.c
Summary: IMM: fix memory leak in IMM [#834]
Review request for Trac Ticket(s): 834
Peer Reviewer(s): Neelakanta
Pull request to: Zoran
Affected branch(es): opensaf-4.3.x, opensaf-4.4.x, default(4.5)
Development branch: default(4.5)
Impacted area Impact y/n
--
Ack from me.
/AndersBj
-Original Message-
From: Zoran Milinkovic [mailto:zoran.milinko...@ericsson.com]
Sent: den 3 april 2014 11:34
To: reddy.neelaka...@oracle.com
Cc: opensaf-devel@lists.sourceforge.net
Subject: [devel] [PATCH 1 of 1] IMM: fix memory leak in IMM [#834]
osaf/services
I think this patch is a bit controversial. We need to define a "protocol" for
the text carried in the error strings to
allow for future changes. For example we should consider source identification,
maybe we should have key/value pairs in
the string?
Remember that the string might end up in fr
Ack
/Hans
On 03/27/2014 02:13 AM, mathi.naic...@oracle.com wrote:
> osaf/services/saf/clmsv/clms/clms_evt.c | 48
> ++--
> 1 files changed, 27 insertions(+), 21 deletions(-)
>
>
> When a node tries to join with a node name that is not configured
> or just gets simp
This patch only affects the ccb-operation callbacks for ccb-
create/delete/modify.
It does not append anything to a user error string since the callback never
reached the user. It adds an error string in the imm server where there
previously could
be no string, now clarifying what the cause of t
Anders Björnerstedt wrote:
> This patch only affects the ccb-operation callbacks for ccb-
> create/delete/modify.
> It does not append anything to a user error string since the callback never
> reached the user. It adds an error string in the imm server where there
> previously could
> be no stri
Comments:
1. First of all, the mds sub part version should not be stored/used. You
should depend on the message format version as SI rank is being embedded
in the new message(which is being sent from AVD).
If you have a requirement for usage of message version, you can get it
at runtime when y
See inline
Thanks,
Hans
On 03/27/2014 02:13 AM, mathi.naic...@oracle.com wrote:
> osaf/services/saf/clmsv/clms/clms_evt.c | 12 ++--
> osaf/services/saf/clmsv/nodeagent/main.c | 7 +++
> 2 files changed, 13 insertions(+), 6 deletions(-)
>
>
> When a node join request from an u
On the controller I get an extra "send failed" log message that is ugly:
Apr 3 13:40:01 SC-1 local0.notice osafclmd[417]: NO CLM NodeName: 'PL-6' is
not a configured cluster node.
Apr 3 13:40:01 SC-1 local0.notice osafclmd[417]: NO /etc/opensaf/node_name
should contain the rdn value of configu
Thanks, that's a good comment. Need to think what the best way forward is. The
change will most likely affect the rest of patch series.
Skickat från min Sony Xperia™-smartphone
SuryaNarayana Garlapati skrev
Comments:
1. First of all, the mds sub part version should not be stored/used
There is no functional need for a retry when it is not a communication problem
there.
The default number of re-spawns has been 3 all along. Yes, we can increase it
to a bigger value.
It is up to the system integrator to decide what to do when OpenSAF startup
fails.
Thanks,
Mathi.
- hans.fe
Ok, Will do that.
- Mathi.
- hans.fe...@ericsson.com wrote:
> See inline
> Thanks,
> Hans
>
> On 03/27/2014 02:13 AM, mathi.naic...@oracle.com wrote:
> > osaf/services/saf/clmsv/clms/clms_evt.c | 12 ++--
> > osaf/services/saf/clmsv/nodeagent/main.c | 7 +++
> > 2 files c
osaf/services/saf/immsv/immnd/ImmModel.cc | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
when the RT Object is deleted in a PBE enabled cluster which is subscribed to
special applier,
the delete callback is not sent. with this patch the delete callback will be
sent
diff --git a/osa
Summary: IMM:sending delete callback for PBE enabled RTobj delete [#835]
Review request for Trac Ticket(s): 835
Peer Reviewer(s): AndersBj
Affected branch(es): 4.3.x, 4.4.x, default
Development branch: default
Impacted area Impact y/n
Hi AndersBj/zoran,
comments inline.
/Neel.
On Thursday 27 March 2014 06:47 PM, Zoran Milinkovic wrote:
> Hi Neelakanta,
>
> Generally, ack from me, with a comment to Anders's comment.
>
> I think you can use "state->attrDefinitions" to determine if a class is
> persistent runtime or not.
> "stat
Hi AndersBj,
I forgot, we have to search first if the special applier is there or
not, the following the new patch( no need to change the "if condition"
as in published patch).
diff --git a/osaf/services/saf/immsv/immnd/ImmModel.cc
b/osaf/services/saf/immsv/immnd/ImmModel.cc
--- a/osaf/servi
Hi Neel,
The proposed patch will work logically, but it is not as efficient as it could
be.
By changing
if(*spApplConnPtr) {
To:
if(spApplConnPtr) {
Is the same as:
if(true) {
That is the condition will always be true because the pointer itself is of
course always
osaf/services/saf/immsv/immnd/ImmModel.cc | 15 +--
1 files changed, 9 insertions(+), 6 deletions(-)
Changes to the server code for immsv done in relation to adding support
for 2PBE in OpenSAF 4.4, broke the logic that generates callbacks to
special appliers for PRTO deletes. The pr
Summary: IMM: Special applier callbacks for PRTO deletes are re-established
[#835]
Review request for Trac Ticket(s): 835
Peer Reviewer(s): Neel
Pull request to:
Affected branch(es): 4.4; defaultr(4.5)
Development branch:
Impacted area Impact y/n
Hi,
I interpreted your answer such that there are two alternatives:
1) rely on mds sub part version but then I need to handle NEW-ACTIVE or
2) use message format version instead
Seems like you had a preference for 2?
Any alternative seems to require quite a few more changes in amfnd.
Alt1 would n
Fine let's see that as a possible future change. Remember removing nid
has been discussed, then we will have no retry at all without my
proposed change.
About the log "NO proc_initialize_msg: send failed", do you think it
comes from the fact that the payload start is not correct? clmna
seemed to r
Hi Hans,
Comments/answers inline.
Regards
Surya
On Friday 04 April 2014 11:16 AM, Hans Feldt wrote:
> Hi,
> I interpreted your answer such that there are two alternatives:
> 1) rely on mds sub part version but then I need to handle NEW-ACTIVE or
> 2) use message format version instead
>
> Seems li
See inline
Thanks,
Hans
> -Original Message-
> From: SuryaNarayana Garlapati [mailto:suryanarayana.garlap...@oracle.com]
> Sent: den 4 april 2014 08:29
> To: Hans Feldt
> Cc: Hans Feldt; Hans Nordebäck; praveen malviya; nagendr...@oracle.com;
> opensaf-devel@lists.sourceforge.net
> Subjec
25 matches
Mail list logo