will share a patch
/Hans
On 7 May 2014 07:47, Hans Feldt wrote:
> The way you have changed the macro has made useable only for pre-instantiable
> SUs since the table you point to is only valid for pre-inst SUs.
>
> We already have some tech debt with this macro:
> // TODO
Summary: v3 amfd: Allow multiple csi addition in single ccb [#750]
Review request for Trac Ticket(s): #750
Peer Reviewer(s): Hans F, Hans N, Praveen
Pull request to: <>
Affected branch(es): All
Development branch: opensaf-4.3.x
Impacted area Impact y/n
--
osaf/services/saf/avsv/avd/avd_csi.c | 77 ++-
osaf/services/saf/avsv/avd/avd_sgproc.c | 14 +
osaf/services/saf/avsv/avd/include/avd_csi.h | 1 +
3 files changed, 55 insertions(+), 37 deletions(-)
Problem:
Amf rejects csi addition if previous csi add
OK,
Here goes the solution.
There is a 8 bit field in each message that is being exchanged between
the services (which contains the message priority, mds prot_ver(mds
protocol and mds version)) when message send is attempted. 6 bits are
allocated for the mds prot_ver, present value for this fiel
osaf/libs/core/mds/include/mds_dt.h | 3 --
osaf/libs/core/mds/mds_dt_tipc.c| 45 +---
2 files changed, 17 insertions(+), 31 deletions(-)
Using connection-less TIPC messaging the largest message that can be sent and
received is 66000 defined by the constant
Summary: MDS TIPC receive buffer v2 (1/4)
Review request for Trac Ticket(s): 654
Peer Reviewer(s): Mahesh
Pull request to: <>
Affected branch(es): default
Development branch: default
Impacted area Impact y/n
Docs
All I am saying that I think you should just log "SI '%s' could not be
assigned to any SU".
The reasons for this has to be logged/found somewhere else.
Thanks,
Hans
On 7 May 2014 07:59, praveen malviya wrote:
>
> On 07-May-14 11:05 AM, Hans Feldt wrote:
>> Since there can be many more reasons why
On 07-May-14 11:05 AM, Hans Feldt wrote:
> Since there can be many more reasons why the SI is not assigned, I don't
> think you should log any of what you have proposed in this part of the code.
> SG, SU can be locked. Node can be locked etc.
This is admin operation on SI. This is the reason why
Ack except for the commit message which just (again) describes the problem
instead of what the patch is changing.
/Hans
> -Original Message-
> From: Minh Hon Chau [mailto:minh.c...@dektech.com.au]
> Sent: den 7 maj 2014 06:39
> To: Hans Nordebäck; Hans Feldt; praveen.malv...@oracle.com;
The way you have changed the macro has made useable only for pre-instantiable
SUs since the table you point to is only valid for pre-inst SUs.
We already have some tech debt with this macro:
// TODO(nagu) remove saAmfSUPreInstantiable check and
move into m_AVD_APP_SU_IS_I
Since there can be many more reasons why the SI is not assigned, I don't think
you should log any of what you have proposed in this part of the code. SG, SU
can be locked. Node can be locked etc.
In other places in the code we could log more when we actually know more things.
/Hans
> -Orig
osaf/services/saf/amf/amfd/si.cc | 5 +
1 files changed, 5 insertions(+), 0 deletions(-)
Problem: A SI is unlocked and it is not assigned to any SU. User
does not have any clue for unassignment.
Reason: Reasons for unassignment can be many.Since admin operation
is invoked on the SI and a S
Summary: amfd: syslog no assignment of SI when it is unlocked [#867]
Review request for Trac Ticket(s): #867
Peer Reviewer(s): Hand F., Nagendra
Pull request to: <>
Affected branch(es): <>
Development branch: <>
Impacted area Impact y/n
osaf/services/saf/amf/amfnd/susm.cc | 13 +
1 files changed, 13 insertions(+), 0 deletions(-)
Problem: If the timer is started, delete the model, load model again,
and kill component in order to escalate to the su restart. Then if the
previous timer event comes during the escalation
Summary: v3 amfnd: saAmfSGCompRestartProb timer is not stopped after the
relevant model had been deleted [#831]
Review request for Trac Ticket(s): 831
Peer Reviewer(s): Hans F, Hans N, Nagendra, Praveen
Pull request to:
Affected branch(es): default
Development branch: default
Surya,
Thank for reiterating arch_word of the MDS feature ,we all in sync.
On 5/6/2014 3:55 PM, SuryaNarayana Garlapati wrote:
>> MDS version unless we get alternate bits/variables used for MDS
>> version.
> [Surya] Thats the reason i am asking for some time.
[AVM] If we get some alternate
Hi Praveen,
I don't think this is necessary because avd_sg_2n_su_chose_asgn()
(and others) checks the readiness state. And the readiness state is only
set to "in-service" in other parts of the code by first testing this
macro which I've modified.
So, if an SU is not instantiated, the
Comments inline...
On 05/06/2014 09:08 AM, Hans Feldt wrote:
> First I don't think we should bring in cluster and application state just
> like this. It is not relevant for this case and we anyway don't support such
> ops.
[Alex] It is relevant in that this macro is a test for SU in-service.
There is a print message there.
fprintf(stderr, "logWriteLogCallbackT FAILED: wrong invocation\n");
I can add if you are referring to a missing message.
Cheers,
Mathi.
- anders.bjornerst...@ericsson.com wrote:
> One comment below.
>
> mathi.naic...@oracle.com wrote:
> > Summary: log: safl
One comment below.
mathi.naic...@oracle.com wrote:
> Summary: log: saflogger to return EXIT_FAILURE when SaLogWriteLogCallbackT
> fails - v2 [#884]
> Review request for Trac Ticket(s): #884
> Peer Reviewer(s): Lennart
> Pull request to: <>
> Affected branch(es): opensaf-4.3.x, 4.4.x, default
> De
Ack from me.
Reviewed the code, but didn't test.
Best regards,
Zoran
-Original Message-
From: Anders Björnerstedt
Sent: den 29 april 2014 16:37
To: reddy.neelaka...@oracle.com; Zoran Milinkovic
Cc: opensaf-devel@lists.sourceforge.net
Subject: [PATCH 0 of 1] Review Request for imm: Allow
osaf/tools/saflog/saflogger/saf_logger.c | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
saflogger is not returning appropriate exit code to the shell when
SaLogWriteLogCallbackT fails either because of an error reported in the
callback or because
of an invalid invocationId.
The pat
tests/logsv/saflogtest.c | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
saflogtest is not returning appropriate exit code to the shell when
SaLogWriteLogCallbackT fails either because of an error reported in the
callback or because
of an invalid invocationId.
The patch exits with E
Summary: log: saflogger to return EXIT_FAILURE when SaLogWriteLogCallbackT
fails - v2 [#884]
Review request for Trac Ticket(s): #884
Peer Reviewer(s): Lennart
Pull request to: <>
Affected branch(es): opensaf-4.3.x, 4.4.x, default
Development branch: <>
Impacted ar
Summary: amfd: return try_again to sg admin unlock-in if any su is in
terminating state [#854]
Review request for Trac Ticket(s): #854
Peer Reviewer(s): Hans F, Hans N, Praveen
Pull request to: <>
Affected branch(es): All
Development branch: Default
Impacted area
osaf/services/saf/amf/amfd/sg.cc | 13 +
1 files changed, 13 insertions(+), 0 deletions(-)
Problem:
SG admin lock-in return immediately.
Next admin unlock-in also succeed even if sus are not terminated yet.
Analysis:
If Amf could return try again to unlock-in admin op, then
another
First I don't think we should bring in cluster and application state just like
this. It is not relevant for this case and we anyway don't support such ops.
Second this macro is out of control and should be changed into a (inline)
function instead.
Haven't I already done that in my refactoring se
In #493 the assignments were given to SU5 without checking its presence
state. But this was not because of some admin operation, but due to the
continuous faults in
in all other SUs which were receiving the assignments. So ideally such a
check of Presence sate should be added in assignment algo
Please check the ticket, it is updated now.
Thanks
-Nagu
> -Original Message-
> From: Hans Feldt [mailto:hans.fe...@ericsson.com]
> Sent: 06 May 2014 17:22
> To: Nagendra Kumar; Hans Nordebäck; Praveen Malviya
> Cc: opensaf-devel@lists.sourceforge.net
> Subject: RE: [PATCH 1 of 1] amfd: i
Hi
This seems Ok but there is one more problem in the same area that should be
fixed as well.
if (cb_invocation != invocation) {
fprintf(stderr, "logWriteLogCallbackT FAILED: wrong
invocation\n");
return errorCode;
}
Should be changed to:
I don't get the use/test case. In the ticket node lock is performed. Node lock
should not affect the mw SU meaning after mw restart the mw SU should be
instantiated. The patch seems to indicate that SU admin ops are done.
Thanks,
Hans
> -Original Message-
> From: nagendr...@oracle.com [m
Summary: amfd: instantiate mw sus when node is joining [#863]
Review request for Trac Ticket(s): #863
Peer Reviewer(s): Hans F, Hans N< praveen
Pull request to: <>
Affected branch(es): All
Development branch: Default
Impacted area Impact y/n
---
osaf/services/saf/amf/amfd/su.cc | 11 +--
1 files changed, 9 insertions(+), 2 deletions(-)
Problem:
When mw su is in locked-in state and opensaf is started, amfnd hangs.
When mw su is unlocked-in, amfnd still doesn't instantiate the mw su.
Analysis:
Amfd doesn't send instantiate messa
Have tested and it seems Ok. ACK
Thanks
Lennart
> -Original Message-
> From: mathi.naic...@oracle.com [mailto:mathi.naic...@oracle.com]
> Sent: den 2 maj 2014 23:45
> To: Lennart Lund
> Cc: opensaf-devel@lists.sourceforge.net
> Subject: [PATCH 0 of 2] Review Request for log: ignore enviro
Summary: log: saflogger to return EXIT_FAILURE when SaLogWriteLogCallbackT
reports error [#884]
Review request for Trac Ticket(s): #884
Peer Reviewer(s): lennart.l...@ericsson.com
Pull request to: <>
Affected branch(es): opensaf-4.3.x, 4.4.x, default
Development branch: <>
---
osaf/tools/saflog/saflogger/saf_logger.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
saflogger is not returning appropriate exit code to the shell when
SaLogWriteLogCallbackT returns error.
The patch exits with EXIT_FAILURE when the SaLogWriteLogCallbackT returns error
diff --git
Before going ahead, Following is the explanation for the arch_word of
the MDS.
Arch word(4bits) is combination of architecture and bit size of the
machine. 3 bits are allocated
for architecture and 1 bit is allocated for bit size.
architecture of value 0 means unspecified.
Message encoding is d
37 matches
Mail list logo