Hi Alex
I can reproduce the coredump by doing “immcfg -f AppConfig-2N.xml” (the
amf_demo sample). It looks better with the patch.
Thanks
Gary
From: Alex Jones
Organization: Ribbon
Date: Saturday, 4 August 2018 at 12:59 am
To: Gary Lee , ,
,
Cc:
Subject: Re: [devel] [PATCH 0/1
free'd
On 3/8/18, 11:25 am, "Gary Lee" wrote:
Hi Alex
I haven’t had a chance to look at it, but I did run our regression tests
with the patch.
amfd is segfaulting regularly, with backtraces like the attachment.
Tha
Hi Alex
I haven’t had a chance to look at it, but I did run our regression tests with
the patch.
amfd is segfaulting regularly, with backtraces like the attachment.
Thanks
Gary
From: Alex Jones
Organization: Ribbon
Date: Thursday, 2 August 2018 at 3:52 am
To: , ,
,
Cc:
---
src/osaf/consensus/consensus.cc | 17 +
src/osaf/consensus/consensus.h | 3 +++
2 files changed, 20 insertions(+)
diff --git a/src/osaf/consensus/consensus.cc b/src/osaf/consensus/consensus.cc
index 800b776e6..2a8e9bb1c 100644
--- a/src/osaf/consensus/consensus.cc
+++
Sometimes the 'watch' command in the KV plugin will not return
a takeover request, if the KV store does not respond in time.
Then rded would try to read the takeover request in the main
thread after receiving a takeover request notification.
This can cause rded to not respond to AMF callbacks in
revision 5c9104ebf210c390b6157341f802720086a8
Author: Gary Lee
Date: Wed, 1 Aug 2018 04:43:58 +
rded: perform KV store operations outside main thread [#2905]
Sometimes the 'watch' command in the KV plugin will not return
a takeover request, if the KV store does not respond in time.
Then rded would t
---
src/amf/amfd/clm.cc | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/amf/amfd/clm.cc b/src/amf/amfd/clm.cc
index e113a65f9..1e67ff389 100644
--- a/src/amf/amfd/clm.cc
+++ b/src/amf/amfd/clm.cc
@@ -631,7 +631,7 @@ AvdJobDequeueResultT ClmTrackStart::exec(AVD_CL_CB*
revision e463518e3d09eb6936c074a269cee24e7a1672e7
Author: Gary Lee
Date: Wed, 1 Aug 2018 03:42:14 +
amfd: remove redundant const_cast [#2907]
Complete diffstat:
--
src/amf/amfd/clm.cc | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
Testing Commands:
-
*** LIST THE COMMAND
2018 04:07 AM, Gary Lee wrote:
> If there is a queued update on a particular object's attribute,
> also queue further 'sync' updates so we don't end up with an
> inconsistent value.
> ---
> src/amf/amfd/imm.cc | 35 +--
>
Hi Nagu
Do you have any comments on this? It seems OK to me, but I know you've
worked on similar scenarios with TIPC flickering before, where reboot is
issued from the PL side.
Thanks
Gary
On 09/07/18 16:37, thuan.tran wrote:
There is a abnormal state that AMFND on remote node keep sending
Hi Minh
ack (review only). Perhaps std::getline() could have been used, but it
doesn't really matter.
Thanks
Gary
On 24/07/18 08:59, Minh Hon Chau wrote:
Hi,
I would like to push the patch by tomorrow if there's no comment on it.
Thanks
Minh
On 20/07/18 13:33, Minh Chau wrote:
If a
Hi Minh
Ack (review only).
It should be OK to call unsetenv all the time though?
man page:
The unsetenv() function deletes the variable name from the
environment. If name does not exist in the environment,
then the function succeeds, and the environment is unchanged.
On
Hi Minh
Ack (review)
Thanks
Gary
On 13/7/18, 9:27 pm, "Minh Chau" wrote:
In scenario of removing assignment, amfnd calls
avnd_su_si_remove(cb, tmp->su, tmp);
Inside avnd_su_si_remove(), the @tmp points to an object AVND_SU_SI_REC
which could be deleted. amfnd continues use
Hi Nagu
The AMF agent library appears to do that. I guess extra care needs to be
taken to remove such "cb locks" before MDS API calls are made.
Gary
On 12/07/18 14:08, nagen...@hasolutions.in wrote:
Hi Thuan,
Since MDS thread is RT and application thread can be non-RT, so I
think there
Hi Minh
ack with 2 minor comments (see below)
Thanks
Gary
On 03/07/18 17:49, Minh Chau wrote:
When THREAD_TRACE_BUFFER enables, it uses by default a list 10240 trace string
as
a circular buffer. There is a concern that how much extra memory of all threads
may use.
This patch makes
Hi Thuan
Looks OK to me, but I would remove the assert. We shouldn't do that in library
code.
Thanks
Gary
On 10/7/18, 6:52 pm, "thuan.tran" wrote:
Sometimes, callback agent dispatch and fail at saSmfReponse()
because cbk list is empty, agent by somehow handle evt before increase
Hi Vu
Ack (review only)
On 2/7/18, 4:23 pm, "Vu Minh Nguyen" wrote:
Valgrind reports mismatched free() / delete / delete [].
---
src/imm/immnd/ImmModel.cc | 32 ++--
1 file changed, 18 insertions(+), 14 deletions(-)
diff --git
revision 0b6fd070e19d15b85dde045e29d9aba1d0862af7
Author: Gary Lee
Date: Fri, 6 Jul 2018 02:02:46 +
amfd: check for queued updates on attribute [#2884]
If there is a queued update on a particular object's attribute,
also queue further 'sync' updates so we don't end up with an
inconsistent value.
Complet
If there is a queued update on a particular object's attribute,
also queue further 'sync' updates so we don't end up with an
inconsistent value.
---
src/amf/amfd/imm.cc | 35 +--
src/amf/amfd/imm.h | 9 ++---
2 files changed, 35 insertions(+), 9 deletions(-)
Hi Minh
You are the right. That line is redundant. I will remove it.
Thanks
Gary
On 05/07/18 09:45, Minh Hon Chau wrote:
Hi Gary,
ack with a minor comment with [M]
Thanks,
Minh
On 04/07/18 11:15, Gary Lee wrote:
---
src/amf/amfd/cb.h | 3 ++-
src/amf/amfd/clm.cc | 4 ++--
src
Hi
I will push this next Monday if no one has comments.
Thanks
Gary
On 25/6/18, 1:10 pm, "Gary Lee" wrote:
* increase timeout in etcd3 plugin
* reduce retry frequency on KV store. This can overload the
KV store even more when the KV store is already timing out
* w
Hi Minh
Ack (review only)
Gary
On 3/7/18, 9:26 pm, "Minh Chau" wrote:
Currently amfd already check @avd_imm_status before setting fd,
amfd should also check @avd_imm_status before calling saImmOiDispatch
---
src/amf/amfd/main.cc | 3 ++-
1 file changed, 2 insertions(+),
---
src/amf/amfd/cb.h | 3 ++-
src/amf/amfd/clm.cc | 4 ++--
src/amf/amfd/clm.h | 4 ++--
src/amf/amfd/imm.cc | 33 -
src/amf/amfd/imm.h | 18 +-
src/amf/amfd/ntf.cc | 2 +-
6 files changed, 40 insertions(+), 24 deletions(-)
diff --git
revision 47decce67b854afa8404202704ade356ceb2a8b2
Author: Gary Lee
Date: Wed, 4 Jul 2018 01:08:34 +
amfd: put sync jobs into queue if IMM is busy [#2863]
Complete diffstat:
--
src/amf/amfd/cb.h | 3 ++-
src/amf/amfd/clm.cc | 4 ++--
src/amf/amfd/clm.h | 4 ++--
src/amf/amfd/i
Hi Minh
Ack with minor comments.
"This option enables the tracing buffered in every threads of OpenSAF services"
- perhaps replace with "This option enables per-thread (a circular buffer
per-thread) tracing of an OpenSAF service".
Maybe you can explicitly state: add 'export
Ack with minor comments (review only)
Use bool instead of SaBoolT.
Variables names should follow the Google Style Guide
https://google.github.io/styleguide/cppguide.html#Variable_Names
So is_changed is probably better.
On 22/6/18, 11:43 am, "thang.nguyen" wrote:
Currently changing
revision a7b0102bdf586c82bccaf42727f5cfc5864c05f0
Author: Gary Lee
Date: Mon, 25 Jun 2018 02:59:54 +
osaf: reduce load on keyvalue store [#2883]
* increase timeout in etcd3 plugin
* reduce retry frequency on KV store. This can overload the
KV store even more when the KV store is already timing out
* when 'watch ke
* increase timeout in etcd3 plugin
* reduce retry frequency on KV store. This can overload the
KV store even more when the KV store is already timing out
* when 'watch key' returns, try to use the value returned instead of
unnecessarily reading it again
---
src/osaf/consensus/consensus.cc
Hi Thang
Is it possible to update the value, eg. comp_type->saAmfCtDefDisableRestart,
outside the loop going through the nodes?
I think with the change, it will send unnecessary updates if you set it to the
same value?
/Gary
-Original Message-
From: "thang.nguyen"
Date: Tuesday, 19
Hi Minh
OK, I'll try to rewrite the patch.
/Gary
On 13/06/18 14:35, Minh Hon Chau wrote:
Hi Gary
Can we reuse @avd_imm_status instead?
Thanks
Minh
On 05/06/18 13:39, Gary Lee wrote:
---
src/amf/amfd/cb.h | 7 +++
src/amf/amfd/clm.cc | 4 ++--
src/amf/amfd/clm.h | 4
Hi
The commit msg should read:
If trace is enabled (via sigusr2) after the last TRACE_ENTER() statement is
run, but
*before* ~Trace() is called, then _file is left as nullptr when
logtrace_output() is called in ~Trace(). This will segfault.
On 08/06/18 13:12, Gary Lee wrote:
If trace
If trace is enabled after the last TRACE() statement is run, but
*before* ~Trace() is called, then _file is left as nullptr when
logtrace_output() is called in ~Trace(). This will segfault.
---
src/base/logtrace.h | 10 +++---
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git
revision 9ac0ec916e55c52ec2484f69ff6a23c725842eb7
Author: Gary Lee
Date: Fri, 8 Jun 2018 12:52:51 +1000
base: ensure _file is always initialised [#2872]
If trace is enabled after the last TRACE() statement is run, but
*before* ~Trace() is called, then _file is left as nullptr when
logtrace_output() is called
---
src/amf/amfd/cb.h | 7 +++
src/amf/amfd/clm.cc | 4 ++--
src/amf/amfd/clm.h | 4 ++--
src/amf/amfd/imm.cc | 36 +++-
src/amf/amfd/imm.h | 18 +-
src/amf/amfd/ntf.cc | 2 +-
6 files changed, 44 insertions(+), 27 deletions(-)
diff
revision 93f9129e651418ecf6e7ddf7c12d2a0b6bad0dab
Author: Gary Lee
Date: Tue, 5 Jun 2018 13:30:18 +1000
amfd: put sync jobs into queue if IMM is busy [#2863]
Complete diffstat:
--
src/amf/amfd/cb.h | 7 +++
src/amf/amfd/clm.cc | 4 ++--
src/amf/amfd/clm.h | 4 ++--
src/amf/amfd/i
Hi Hans
Ack
Thanks
Gary
On 16/5/18, 5:27 pm, "Hans Nordeback" wrote:
---
src/base/daemon.c | 52 ++---
tools/cluster_sim_uml/build_uml | 1 +
2 files changed, 49 insertions(+), 4 deletions(-)
nly change is addition of a single line:
+ election_end_time_ = base::kTimespecMax;
This is to ensure we don't launch another thread,
before SetRole.
revision 047b0545824d8a6118a98b933ea8ed89ebab4a3a
Author: Gary Lee <gary@dektech.com.au>
Date: Thu, 24 May 2018 14:12:46 +1000
rded
Currently, the consensus code relating to node promotion
is run from the main thread. We can improve rded's
responsiveness by moving this code into another thread.
---
src/rde/rded/rde_cb.h| 3 +-
src/rde/rded/rde_main.cc | 6 +++-
src/rde/rded/role.cc | 83
Currently, the consensus code relating to node promotion
is run from the main thread. We can improve rded's
responsiveness by moving this code into another thread.
---
src/rde/rded/rde_cb.h| 3 +-
src/rde/rded/rde_main.cc | 6 +++-
src/rde/rded/role.cc | 82
n
OpenSAF servicesy
Core libraries n
Samples n
Tests n
Other n
Comments (indicate scope for each "y" above):
-
revision a12571713545342e03bafce4cfd6c0b4ab28094b
Author
In etcd3.plugin, if the unlock transaction times out
for some reason, unlock() would return 1, indicating it is owned by
someone else when it isn't because lock_owner was not
being called correctly.
Also PromoteThisNode() would indicate a successful lock
attempt when it failed, because rc was
n
OpenSAF servicesy
Core libraries y
Samples n
Tests n
Other n
Comments (indicate scope for each "y" above):
-
revision 0855c361a9c736bfdac4dfd7f5c834a338963a3b
Au
rded should not automatically include itself in the cluster member list.
Instead it should rely solely on AMFND service up, so that the count
is consistent across nodes.
Also adjust some split-brain prevention related values. More time
is required to ensure we should have an accurate view of
revision 0018bdeb4df610289560be30c57d505b77af7705
Author: Gary Lee <gary@dektech.com.au>
Date: Wed, 25 Apr 2018 18:18:25 +1000
rded: prevent unnecessary takeover [#2843]
rded should not automatically include itself in the cluster member list.
Instead it should rely solely on AMFND service up, so that the cou
Sorry, it should read:
Note: Patches 1 to 5 are identical to V3.
Patch 6 (osaf: remove timestamp from takeover request) is the diff
between V3 and V4.
On 13/04/18 21:50, Gary Lee wrote:
Summary: split-brain: select active SC from largest network partition V4 [#2795]
Review request for Ticket
* update create() in the plugins to include a timeout parameter
* remove timestamp from the takeover request and utilise the
built-in timeout functionality in the KV store
---
src/osaf/consensus/consensus.cc | 165 ---
src/osaf/consensus/consensus.h
- add create_key function (fails if key already exists)
- add setkey_match_prev function (set value if previous value matches)
- add missing quotes
- add etcd3.plugin
---
src/osaf/consensus/plugins/etcd.plugin | 86 +++-
src/osaf/consensus/plugins/etcd3.plugin | 366
---
src/amf/amfd/role.cc | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/amf/amfd/role.cc b/src/amf/amfd/role.cc
index c8aa9cf1f..790983ee7 100644
--- a/src/amf/amfd/role.cc
+++ b/src/amf/amfd/role.cc
@@ -1217,7 +1217,7 @@ uint32_t amfd_switch_stdby_actv(AVD_CL_CB *cb) {
---
src/fm/fmd/fm_cb.h| 2 +-
src/fm/fmd/fm_main.cc | 26 +-
src/fm/fmd/fm_mds.cc | 2 ++
src/fm/fmd/fm_rda.cc | 27 ++-
4 files changed, 30 insertions(+), 27 deletions(-)
diff --git a/src/fm/fmd/fm_cb.h b/src/fm/fmd/fm_cb.h
index
- add 3 new internal message:
RDE_MSG_NODE_UP
RDE_MSG_NODE_DOWN
RDE_MSG_TAKEOVER_REQUEST_CALLBACK
- subscribe to AMFND service up events to keep track of the number
of cluster members
- listen for takeover requests in KV store
---
src/rde/rded/rde_cb.h| 12 ++--
- add create and set (if previous value matches) functions to KeyValue class
- add Consensus::MonitorTakeoverRequest() function for use by RDE to answer
takeover requests
- add Consensus::CreateTakeoverRequest() - before a SC is promoted to active,
it will
create a takeover request in the KV
Note: Patches 1 to 4 are identical to V3.
Patch 5 (osaf: remove timestamp from takeover request) is the diff
between V3 and V4.
revision c8ca1da49c61d207cc7c1fa706f35fb3950b492d
Author: Gary Lee <gary@dektech.com.au>
Date: Fri, 13 Apr 2018 21:40:24 +1000
osaf: remove timesta
/18 16:16, Hans Nordebäck wrote:
Hi,
On 04/12/2018 04:15 PM, Gary Lee wrote:
Hi
On 12/04/18 23:34, Anders Widell wrote:
Ack with comments:
* There is no need to use "const" when passing function arguments by
value. E.g. the argument "const uint64_t cluster_size"
matically deletes the takeover request when it expires? That
way we would not require synchronized clocks.
[GL] Good idea. I did question why I hadn't use TTL/lease once I had
finished the ticket. :-) Will see what I can do!
regards,
Anders Widell
On 04/11/2018 09:35 AM, Gary Lee wrote:
Summary:
- add 3 new internal message:
RDE_MSG_NODE_UP
RDE_MSG_NODE_DOWN
RDE_MSG_TAKEOVER_REQUEST_CALLBACK
- subscribe to AMFND service up events to keep track of the number
of cluster members
- listen for takeover requests in KV store
---
src/rde/rded/rde_cb.h| 12 ++--
- add create_key function (fails if key already exists)
- add setkey_match_prev function (set value if previous value matches)
- add missing quotes
- add etcd3.plugin
---
src/osaf/consensus/plugins/etcd.plugin | 86 +++-
src/osaf/consensus/plugins/etcd3.plugin | 366
revision c7bc78656d5de11f6147727bd8612274fb6e438f
Author: Gary Lee <gary@dektech.com.au>
Date: Wed, 11 Apr 2018 17:16:46 +1000
rded: adapt to new Consensus API [#2795]
- add 3 new internal message:
RDE_MSG_NODE_UP
RDE_MSG_NODE_DOWN
RDE_MSG_TAKEOVER_REQUEST_CALLBACK
- subscribe to AMFND service up even
---
src/amf/amfd/role.cc | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/amf/amfd/role.cc b/src/amf/amfd/role.cc
index c8aa9cf1f..790983ee7 100644
--- a/src/amf/amfd/role.cc
+++ b/src/amf/amfd/role.cc
@@ -1217,7 +1217,7 @@ uint32_t amfd_switch_stdby_actv(AVD_CL_CB *cb) {
---
src/fm/fmd/fm_cb.h| 2 +-
src/fm/fmd/fm_main.cc | 26 +-
src/fm/fmd/fm_mds.cc | 2 ++
src/fm/fmd/fm_rda.cc | 27 ++-
4 files changed, 30 insertions(+), 27 deletions(-)
diff --git a/src/fm/fmd/fm_cb.h b/src/fm/fmd/fm_cb.h
index
- add create and set (if previous value matches) functions to KeyValue class
- add Consensus::MonitorTakeoverRequest() function for use by RDE to answer
takeover requests
- add Consensus::CreateTakeoverRequest() - before a SC is promoted to active,
it will
create a takeover request in the KV
- add create_key function (fails if key already exists)
- add setkey_match_prev function (set value if previous value matches)
- add missing quotes
- add etcd3.plugin
---
src/osaf/consensus/plugins/etcd.plugin | 86 +++-
src/osaf/consensus/plugins/etcd3.plugin | 355
- add 3 new internal message:
RDE_MSG_NODE_UP
RDE_MSG_NODE_DOWN
RDE_MSG_TAKEOVER_REQUEST_CALLBACK
- subscribe to AMFND service up events to keep track of the number
of cluster members
- listen for takeover requests in KV store
---
src/rde/rded/rde_cb.h| 12 ++--
---
src/fm/fmd/fm_main.cc | 26 +-
src/fm/fmd/fm_mds.cc | 2 ++
src/fm/fmd/fm_rda.cc | 15 +--
3 files changed, 20 insertions(+), 23 deletions(-)
diff --git a/src/fm/fmd/fm_main.cc b/src/fm/fmd/fm_main.cc
index 73c9b9ccd..3371ec5e8 100644
---
- add create and set (if previous value matches) functions to KeyValue class
- add Consensus::MonitorTakeoverRequest() function for use by RDE to answer
takeover requests
- add Consensus::CreateTakeoverRequest() - before a SC is promoted to active,
it will
create a takeover request in the KV
---
src/amf/amfd/role.cc | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/amf/amfd/role.cc b/src/amf/amfd/role.cc
index c8aa9cf1f..790983ee7 100644
--- a/src/amf/amfd/role.cc
+++ b/src/amf/amfd/role.cc
@@ -1217,7 +1217,7 @@ uint32_t amfd_switch_stdby_actv(AVD_CL_CB *cb) {
uot; above):
-
Changes from V1:
Delete takeover request *after* obtaining lock
Add etcd3.plugin
revision b23363e798e8f02eae550f520a1525afe11362bc
Author: Gary Lee <gary@dektech.com.au>
Date: Fri, 6 Apr 2018 16:07:41 +1000
rded: adapt to new Consensus API [#2795]
- add 3 new
---
src/amf/amfd/role.cc | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/amf/amfd/role.cc b/src/amf/amfd/role.cc
index c8aa9cf1f..790983ee7 100644
--- a/src/amf/amfd/role.cc
+++ b/src/amf/amfd/role.cc
@@ -1217,7 +1217,7 @@ uint32_t amfd_switch_stdby_actv(AVD_CL_CB *cb) {
- add 3 new internal message:
RDE_MSG_NODE_UP
RDE_MSG_NODE_DOWN
RDE_MSG_TAKEOVER_REQUEST_CALLBACK
- subscribe to AMFND service up events to keep track of the number
of cluster members
- listen for takeover requests in KV store
---
src/rde/rded/rde_cb.h| 12 ++---
---
src/fm/fmd/fm_main.cc | 26 +-
src/fm/fmd/fm_mds.cc | 2 ++
src/fm/fmd/fm_rda.cc | 15 +--
3 files changed, 20 insertions(+), 23 deletions(-)
diff --git a/src/fm/fmd/fm_main.cc b/src/fm/fmd/fm_main.cc
index 73c9b9ccd..3371ec5e8 100644
---
revision 65b5489d7209987a7246b94419d1358235d15abb
Author: Gary Lee <gary@dektech.com.au>
Date: Wed, 4 Apr 2018 16:18:14 +1000
rded: adapt to new Consensus API [#2795]
- add 3 new internal message:
RDE_MSG_NODE_UP
RDE_MSG_NODE_DOWN
RDE_MSG_TAKEOVER_REQUEST_CALLBACK
- subscribe to AMFND service up even
- add create_key function (fails if key already exists)
- add setkey_match_prev function (set value if previous value matches)
- add missing quotes
---
src/osaf/consensus/plugins/etcd.plugin | 86 ++
1 file changed, 77 insertions(+), 9 deletions(-)
diff --git
- add create and set (if previous value matches) functions to KeyValue class
- add Consensus::MonitorTakeoverRequest() function for use by RDE to answer
takeover requests
- add Consensus::CreateTakeoverRequest() - before a SC is promoted to active,
it will
create a takeover request in the KV
to analyze the flows in MDS code to figure out the root cause, but it
is of vain.
Can you please give me the reproducible scenario, I will try out by adding
debug logs
Thanks,
Syam.
-Original Message-
From: Ravi Sekhar Reddy Konda
Sent: Monday, March 19, 2018 10:28 AM
To: Gary Lee <g
urn NCSCC_RC_REQ_TIMOUT from the MDS and
> let the Application handle it
>
> Thanks,
> Ravi
>
> -Original Message-
> From: Gary Lee [mailto:gary@dektech.com.au]
> Sent: Thursday, March 01, 2018 11:02 AM
> To: hans.nordeb...@ericsson.com; ravisekhar.ko
ack from me (review)
On 16/03/18 16:14, Minh Hon Chau wrote:
Hi Hans, I will update the PR.
Hi Gary, Ravi,
Do we have any comments?
Thanks,
Minh
On 16/03/18 01:51, Hans Nordebäck wrote:
Hi Minh, ack review only. Good if the AMF Programmer's Reference also
be updated regarding this, (si
d lock as last operation in opensafd stop?
/HansN
On 03/09/2018 06:57 AM, Gary Lee wrote:
If we have a 'tied election' and split-brain prevention is enabled,
then the 'old active' is fenced, or the 'old active' will self-reboot
when it is notified a new node is active.
We need to disable this redundan
with minor comment: Return type from strlen() is size_t, and
return type from send() is ssize_t. So a more type-correct way to
implement this is outlined in my inlined comments in the code below.
regards,
Anders Widell
On 03/09/2018 02:44 AM, Gary Lee wrote:
---
src/rde/agent/rda_papi.cc | 8
If we have a 'tied election' and split-brain prevention is enabled,
then the 'old active' is fenced, or the 'old active' will self-reboot
when it is notified a new node is active.
We need to disable this redundant check in fmd. Otherwise, the 'new active'
will also reboot, along with the 'old
uot; above):
-
revision 49325b25b08d47015be2aa892fe5500b50928788
Author: Gary Lee <gary@dektech.com.au>
Date: Fri, 9 Mar 2018 16:51:17 +1100
fmd: avoid conflict with split-brain prevention if two nodes are elected [#2801]
If we have a 'tied election' and split-brain preve
revision e7601d2a04d99ecaa1022880370ead597187647f
Author: Gary Lee <gary@dektech.com.au>
Date: Fri, 9 Mar 2018 12:37:01 +1100
rde: avoid SIGPIPE in send functions [#2800]
Complete diffstat:
--
src/rde/agent/rda_papi.cc | 8 +---
src/rde/rded/rde_rda.cc | 8 +---
2 files changed, 10
---
src/rde/agent/rda_papi.cc | 8 +---
src/rde/rded/rde_rda.cc | 8 +---
2 files changed, 10 insertions(+), 6 deletions(-)
diff --git a/src/rde/agent/rda_papi.cc b/src/rde/agent/rda_papi.cc
index a07d2b06f..57ad57761 100644
--- a/src/rde/agent/rda_papi.cc
+++
:09, Anders Widell wrote:
Ack with comment: I think the file should be called README.etcd
instead of etcd.readme. Shouldn't it be added to the Makefile as well?
regards,
Anders Widell
On 03/07/2018 04:42 AM, Gary Lee wrote:
---
src/osaf/consensus/plugins/etcd.readme | 32
---
src/osaf/consensus/plugins/etcd.plugin | 25 -
1 file changed, 16 insertions(+), 9 deletions(-)
diff --git a/src/osaf/consensus/plugins/etcd.plugin
b/src/osaf/consensus/plugins/etcd.plugin
index f8d3c7f25..586059b32 100644
--- a/src/osaf/consensus/plugins/etcd.plugin
revision 8ba1cf4a10e1756b50ce39beb717f496a2b36040
Author: Gary Lee <gary@dektech.com.au>
Date: Wed, 7 Mar 2018 15:11:56 +1100
osaf: use local etcd instance only [#2797]
Complete diffstat:
--
src/osaf/consensus/plugins/etcd.plugin | 25 -
1 file changed, 16 insertions(+), 9 deleti
---
src/osaf/consensus/plugins/etcd.readme | 32
1 file changed, 32 insertions(+)
create mode 100644 src/osaf/consensus/plugins/etcd.readme
diff --git a/src/osaf/consensus/plugins/etcd.readme
b/src/osaf/consensus/plugins/etcd.readme
new file mode 100644
index
revision eba49de4b62bb2274fd5bd97cf7395456fa587e3
Author: Gary Lee <gary@dektech.com.au>
Date: Wed, 7 Mar 2018 14:24:24 +1100
osaf: add example config for etcd [#2784]
Added Files:
src/osaf/consensus/plugins/etcd.readme
Complete diffstat:
--
src/osaf/consensus/plugins/
Hi
Perhaps you just need to run ldconfig.
Gary
On 06/03/18 08:19, Feng Xie wrote:
Hi,
I am new in using OpenSAF. I encountered an error in running the latest
OpenSaf software in a Ubuntu Linux VM. I would appreciate if somebody can
provide some hint on this issue in specific and some
Callers of ava_mds_send() assume *o_msg is not null, if
the return code is NCSCC_RC_SUCCESS.
---
src/amf/agent/ava_mds.cc | 4
1 file changed, 4 insertions(+)
diff --git a/src/amf/agent/ava_mds.cc b/src/amf/agent/ava_mds.cc
index 440885332..cd139365d 100644
--- a/src/amf/agent/ava_mds.cc
revision 91c24d80f69c283d7107cd98fbcc6dffd3c7639e
Author: Gary Lee <gary@dektech.com.au>
Date: Thu, 1 Mar 2018 14:27:25 +1100
amf: do not dereference null pointer [#2791]
Callers of ava_mds_send() assume *o_msg is not null, if
the return code is NCSCC_RC_SUCCESS.
Complete diffstat:
--
src/amf/agen
---
src/amf/amfnd/err.cc | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/src/amf/amfnd/err.cc b/src/amf/amfnd/err.cc
index a0529b96c..1d6eb3757 100644
--- a/src/amf/amfnd/err.cc
+++ b/src/amf/amfnd/err.cc
@@ -458,9 +458,6 @@ uint32_t avnd_err_process(AVND_CB *cb,
Hi Ravi
ack
Thanks
On 22/02/18 21:37, ravi-sekhar wrote:
---
src/amf/amfd/sgproc.cc | 8 +---
1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/src/amf/amfd/sgproc.cc b/src/amf/amfd/sgproc.cc
index 610c205..2bee875 100644
--- a/src/amf/amfd/sgproc.cc
+++
revision 88b55873c7e747092a660df6ed8c5f7366e5a599
Author: Gary Lee <gary@dektech.com.au>
Date: Mon, 19 Feb 2018 11:42:17 +1100
amfnd: remove duplicate log entry [#2783]
Complete diffstat:
--
src/amf/amfnd/err.cc | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
Testing Commands:
--
Ack, review only.
Will push on Nguyen's behalf.
On Fri, 16 Feb 2018 at 7:29 pm, Ravi Sekhar Reddy Konda <
ravisekhar.ko...@oracle.com> wrote:
> Hi Nguyen,
>
> Ack, code review only
>
> Regards,
> Ravi
>
> -Original Message-
> From: Nguyen Luu [mailto:nguyen.tk@dektech.com.au]
>
d9331cc7c15a4
Author: Gary Lee <gary@dektech.com.au>
Date: Tue, 13 Feb 2018 14:53:24 +1100
fmd: prevent data races between MDS and main threads [#2763]
Complete diffstat:
--
src/fm/fmd/fm_cb.h| 89 ++-
src/fm/
---
src/fm/fmd/fm_cb.h| 89 ++-
src/fm/fmd/fm_main.cc | 38 +++---
src/fm/fmd/fm_mds.cc | 35 +++-
src/fm/fmd/fm_mds.h | 2 ++
src/fm/fmd/fm_mem.h | 8 -
5 files changed, 86 insertions(+), 86
Hi Ravi/Anders
AndersW> This is slightly out of scope since there are many RAFT
implementations, but I agree it could be a good idea to provide a sample
configuration for etcd along with the sample etcd plugin.
I will try to provide a sample plugin for an external etcd server,
Hi Anders
Thanks, please see [GL].
On 08/02/18 02:06, Anders Widell wrote:
Ack with minor comments inline, marked AndersW>.
regards,
Anders Widell
On 02/06/2018 03:43 AM, Gary Lee wrote:
---
src/fm/fmd/fm_cb.h | 89
++-
src/fm/
Hi Nguyen
Some comments below marked with [GL].
Thanks
Gary
On 6/2/18, 6:35 pm, "Nguyen Luu" wrote:
Valid environment variable should have the format 'var=value'.
AMF currently does not validate this format during CREATE CCBs
for comptype and
Hi Minh
Ack (review)
Thanks
Gary
On 6/2/18, 5:15 pm, "Minh Chau" wrote:
When two errors happen to component and both escalates to a su
failover. For the first su failover recovery, amfd will send
new assignment to su. However, the second error happens
revision d51ded1a789b252e7f75c96737276069b8806748
Author: Gary Lee <gary@dektech.com.au>
Date: Tue, 6 Feb 2018 13:37:35 +1100
fmd: prevent data races between MDS and main threads [#2763]
Complete diffstat:
--
src/fm/fmd/fm_cb.h| 89 ++-
201 - 300 of 904 matches
Mail list logo