revision 0280c451d369880a25869a1e01fbd9ddc2b036a5
Author: thang.d.nguyen
Date: Sat, 25 May 2024 17:59:46 +0700
clm: fix error printed in syslog [#3353]
When the opensafd starts failed. The systemd tries to restart
opensafd service. The CLMD did not update nodeup state correctly
and it causes
When the opensafd starts failed. The systemd tries to restart
opensafd service. The CLMD did not update nodeup state correctly
and it causes the error message.
The fix is to correct that state.
---
src/clm/clmd/clms_evt.cc | 35 ++-
src/clm/clmd/clms_evt.h | 2
uot; above):
-
revision e63ed820cc55055dd7519dd4adb9067144dae55a
Author: thang.d.nguyen
Date: Mon, 6 May 2024 19:34:33 +0700
osaf: supports build with gcc/g++ 13 [#3352]
Fix some minor errors while build/compile source code
with gcc/g++ 13.
Complete diffstat:
--
src/imm/immnd/
Fix some minor errors while build/compile source code
with gcc/g++ 13.
---
src/imm/immnd/ImmModel.h | 1 +
src/osaf/consensus/consensus_env.h | 1 +
src/smf/smfnd/smfnd_amf.c | 2 +-
3 files changed, 3 insertions(+), 1 deletion(-)
diff --git a/src/imm/immnd/ImmModel.h
revision e01bbd5f649452acf4df8046fc8fcacd30bac792
Author: thang.d.nguyen
Date: Mon, 15 Jan 2024 17:23:45 +0700
osaf: supports compile/build with gcc/g++ 12 [#3346]
Fix error when compiling/building Opensaf using
gcc/g++ 12.
Complete diffstat:
--
src/amf/amfd/node.h| 5 -
src
Fix error when compiling/building Opensaf using
gcc/g++ 12.
---
src/amf/amfd/node.h| 5 +++-
src/dtm/dtmnd/dtm_node_db.cc | 2 +-
src/imm/immnd/immnd_proc.c | 2 +-
src/osaf/consensus/key_value.cc| 1 +
src/smf/smfd/SmfProcedureThread.cc | 43
With 2N app model with NPI SU in PI SU, the PI components
is terminated following by lock command on SU. The SU will
be in uninstantiated presence state. And there is no action
can repair the SU. This issue due to during lock the SU,
the NPI component terminated while the SU in restarting.
And it
uot; above):
-
revision 444dac6f8da5c65517c719a5a0a16340db55f3b4
Author: thang.d.nguyen
Date: Sat, 16 Sep 2023 08:04:16 +0700
amf: fix SU in unhealthy state [#3343]
With 2N app model with NPI SU in PI SU, the PI components
is terminated following by lo
When SC fail-overs, the new AMFD active will send
AVSV_D2N_DATA_VERIFY_MSG message event to AMFND.
When AMFND processes this event, it also send the
protection group (PG) start message to new active
AMFD. In case MDS has not updated its tables by
this time, then AMFND still send to old active
uot; above):
-
revision 7cade9c80a6a4370034198f65da628a47b3a34e2
Author: thang.d.nguyen
Date: Wed, 16 Aug 2023 19:40:26 +0700
amf: fix mismatch message id during sc fail-over [#3342]
When SC fail-overs, the new AMFD active will send
AVSV_D2N_DATA_VERIFY_MSG message event to AMFND.
When AMFND
The global smfd_dest (in cd struct) shall be restored with MDS event
NCSMDS_UP and NCSMDS_NEW_ACTIVE (is set to 0 with event NCSMDS_DOWN).
---
src/smf/smfnd/smfnd_mds.c | 9 +
1 file changed, 9 insertions(+)
diff --git a/src/smf/smfnd/smfnd_mds.c b/src/smf/smfnd/smfnd_mds.c
index
revision 2d600b27d045fcaafeb66e0e59d61cc64cce4842
Author: thang.d.nguyen
Date: Thu, 13 Jul 2023 05:54:32 +0700
smf: Fix handling of MDS events in smfnd [#3340]
The global smfd_dest (in cd struct) shall be restored with MDS event
NCSMDS_UP and NCSMDS_NEW_ACTIVE (is set to 0 with event NCSMDS_DOWN).
Complet
uot;y" above):
-
revision 41a0e8e3525f7635f928ffb6740bb378ee1c0fbd
Author: thang.d.nguyen
Date: Tue, 7 Mar 2023 11:29:47 +0700
build: fix HIGH severity in codechecker [#3327]
Fix HIGH severity in codechecker report.
Complet
Fix HIGH severity in codechecker report.
---
src/amf/amfd/ndmsg.cc | 4 ++--
src/amf/amfd/ndproc.cc | 4 ++--
src/amf/amfd/timer.cc | 4 ++--
src/amf/amfnd/amfnd.cc | 4 ++--
src/amf/amfnd/mds.cc
uot;y" above):
-
revision d97c9840bc9c5a6106a29f2609311f7c56e7e57c
Author: thang.d.nguyen
Date: Thu, 2 Mar 2023 05:30:24 +0700
build: fix HIGH severity in codechecker [#3327]
Fix HIGH severity in codechecker report.
Complet
Fix HIGH severity in codechecker report.
---
src/amf/amfd/ndmsg.cc | 4 ++--
src/amf/amfd/ndproc.cc | 4 ++--
src/amf/amfd/timer.cc | 4 ++--
src/amf/amfnd/amfnd.cc | 4 ++--
src/amf/amfnd/mds.cc
uot; above):
-
revision e4fdf7487e7a3fb01d7c79c6d604c435ee8c9652
Author: thang.d.nguyen
Date: Wed, 18 Jan 2023 09:35:52 +0700
amf: reboot to recover from mismatch msg id [#3040]
There is a case that mismatch messge id between
AMFND and active AMFD. It causes the system unhealthy
(partial assigne
There is a case that mismatch messge id between
AMFND and active AMFD. It causes the system unhealthy
(partial assigned).
Reboot the node that message mismatched to recover
the system.
---
src/amf/amfd/sgproc.cc | 7 +++
1 file changed, 7 insertions(+)
diff --git a/src/amf/amfd/sgproc.cc
Remove others/world permissions for mds.log, nid.log,
osaf.log and trace files.
---
src/base/log_writer.cc | 2 +-
src/nid/nodeinit.cc| 5 +
2 files changed, 6 insertions(+), 1 deletion(-)
diff --git a/src/base/log_writer.cc b/src/base/log_writer.cc
index b68566e5c..48bd4bab8 100644
---
uot; above):
-
revision 8c65fb51c6117829f5e36fd04c5aec2f11c2cac0
Author: thang.d.nguyen
Date: Mon, 1 Aug 2022 10:43:46 +0700
base: remove others/world permissions for mds.log, nid.log, osaf.log and trace
files [#3318]
Remove others/world permissions f
uot;y" above):
-
revision 5af7657fbb5051bab9703cc74cfd59150f830e31
Author: thang.d.nguyen
Date: Wed, 30 Mar 2022 13:36:42 +0700
ckpt: fix memleak in API test [#3307]
Fix memleak in ckpt API test.
revision 85c10e503a097004fe7c09631c8fee130b574
Fix error to support compile code with gcc/g++ 10.
---
src/ckpt/agent/cpa_cb.h | 2 +-
src/ckpt/ckptd/cpd_amf.c | 1 -
src/ckpt/ckptd/cpd_init.h | 2 +-
src/evt/agent/eda.h | 2 +-
Fix memleak in ckpt API test.
---
src/ckpt/apitest/ckpttest.c | 11 ++-
src/ckpt/apitest/test_cpa.c | 33 ---
src/ckpt/apitest/test_cpa_util.c | 17 +---
src/ckpt/apitest/test_cpsv.h | 4
src/ckpt/apitest/test_cpsv_conf.h |
Fix memleak in ckpt API test.
---
src/ckpt/apitest/ckpttest.c | 12 ++-
src/ckpt/apitest/test_cpa.c | 33 ---
src/ckpt/apitest/test_cpa_util.c | 17 +---
src/ckpt/apitest/test_cpsv.h | 4
src/ckpt/apitest/test_cpsv_conf.h |
Fix error to support compile code with gcc/g++ 10.
---
src/ckpt/agent/cpa_cb.h | 2 +-
src/ckpt/ckptd/cpd_amf.c | 1 -
src/ckpt/ckptd/cpd_init.h | 2 +-
src/evt/agent/eda.h | 2 +-
uot; above):
-
revision d77833fd07f78b2bd305bc2ed9fa34ffe261bb07
Author: thang.d.nguyen
Date: Tue, 29 Mar 2022 13:27:18 +0700
ckpt: fix memleak in API test [#3307]
Fix memleak in ckpt API test.
revision 8cc8816716b64c4a9f3af957d7363d39fa6b5
uot; above):
-
revision a76363965f5ddb7cdd7f0dcb7c3c7fc949a07a23
Author: thang.d.nguyen
Date: Fri, 25 Mar 2022 09:39:58 +0700
ckpt: fix memleak in API test [#3307]
Fix memleak in API test.
revision 8dacd570f5b805661b747b0e8d834ad991a2315d
Author: than
Fix error to support compile code with gcc/g++ 10.
---
src/ckpt/agent/cpa_cb.h | 2 +-
src/ckpt/ckptd/cpd_amf.c | 1 -
src/ckpt/ckptd/cpd_init.h | 2 +-
src/evt/agent/eda.h | 2 +-
Fix memleak in API test.
---
src/ckpt/apitest/test_cpa.c | 3 +++
src/ckpt/apitest/test_cpa_util.c | 17 ++---
src/ckpt/apitest/test_cpsv_conf.h | 3 ++-
3 files changed, 19 insertions(+), 4 deletions(-)
diff --git a/src/ckpt/apitest/test_cpa.c b/src/ckpt/apitest/test_cpa.c
uot; above):
-
revision b3c201ea329d1c6370af694728249943cb7f0d95
Author: thang.d.nguyen
Date: Wed, 16 Mar 2022 09:26:59 +0700
osaf: fix memleak in API test [#3307]
Fix some memleak in API test.
revision 430bedc07eb65ab90a1f8894950aae61a9cb6
Fix some memleak in API test.
---
src/ckpt/apitest/test_cpa.c | 2 ++
src/ckpt/apitest/test_cpa_util.c | 13 ++---
2 files changed, 12 insertions(+), 3 deletions(-)
diff --git a/src/ckpt/apitest/test_cpa.c b/src/ckpt/apitest/test_cpa.c
index 6c37e91d5..0093b91ea 100644
---
Fix error to support compile code with gcc/g++ 10.
---
src/ckpt/agent/cpa_cb.h | 2 +-
src/ckpt/apitest/test_cpsv_conf.h | 2 +-
src/ckpt/ckptd/cpd_amf.c | 1 -
src/ckpt/ckptd/cpd_init.h | 2 +-
src/evt/agent/eda.h
uot; above):
-
revision 38726264a752b8646030c83ed046388f64bac8b7
Author: thang.d.nguyen
Date: Tue, 15 Mar 2022 18:39:26 +0700
osaf: support compile with gcc/g++ 10 [#3307]
- Fix error to support gcc/g++ 10.
- Fix memleak in api test.
Complet
- Fix error to support gcc/g++ 10.
- Fix memleak in api test.
---
src/ckpt/agent/cpa_cb.h | 2 +-
src/ckpt/apitest/test_cpa.c | 2 +
src/ckpt/apitest/test_cpa_util.c | 13 ++-
src/ckpt/apitest/test_cpsv_conf.h | 2 +-
uot; above):
-
revision 277ed75fc7781dc53c01d9d6aabd1fccfcbec2cf
Author: thang.d.nguyen
Date: Fri, 4 Mar 2022 14:57:19 +0700
amf: reboot to recovery PL in split-brain [#3309]
The connection between the standby SC and that PL was dropped
(disconnect the reconnect ), but that PL still connected with
The connection between the standby SC and that PL was dropped
(disconnect the reconnect ), but that PL still connected with
the active SC. It led the standby SC considered that PL absented
regardless the connection was established after that. During failover,
the standby SC will notify all
During SC failover, message sent on ACTIVE AMFD can not be
checked point to AMFD on STANDBY SC. But the AMFND still
increase receive/send msg id count. Then STANDBY SC takes
ACTIVE and mismatch message id b/w AMFND and new active AMFD.
Solution is to make msg id count alignment b/w AMFD/AMFND
in
"y" above):
-
revision 4062588fae381ecf46b91ee7b7a5e4ab2e776210
Author: thang.d.nguyen
Date: Mon, 21 Feb 2022 20:32:29 +0700
amf: fix unexpected node reboot during failover [#3308]
During SC failover, message sent on ACTIVE AMFD can not
"y" above):
-
revision ef12a46fe03f87d34f7029fe0aa7d67093369e88
Author: thang.d.nguyen
Date: Mon, 21 Feb 2022 13:15:05 +0700
amf: fix unexpected node reboot during failover [#3308]
During SC failover, message sent on ACTIVE AMFD can not
During SC failover, message sent on ACTIVE AMFD can not be
checked point to AMFD on STANDBY SC. But the AMFND still
increase receive/send msg id count. Then STANDBY SC takes
ACTIVE and mismatch message id b/w AMFND and new active AMFD.
Solution is to make msg id count alignment b/w AMFD/AMFND
in
During SC failover, message sent on ACTIVE AMFD can not be
checked point to AMFD on STANDBY SC. But the AMFND still
increase receive/send msg id count. Then STANDBY SC takes
ACTIVE and mismatch message id b/w AMFND and new active AMFD.
Solution is to make msg id count alignment b/w AMFD/AMFND
in
uot; above):
-
revision 011c2e97e6055302d64d3b7d3afb15319f03a022
Author: thang.d.nguyen
Date: Mon, 21 Feb 2022 08:53:32 +0700
amf: fix unexpected node reboot during failover [#3308]
During SC failover, message sent on ACTIVE AMFD can not be
checked point to AMFD on
In mergeBundleRefRollingToSingleStep(), one alias is used to
store m_swAddList then it is reused to store m_swRemoveList.
So it causes the data in m_swAddList overwritten.
The fix is to separate alias to store them.
---
src/smf/smfd/SmfUpgradeProcedure.cc | 9 +
1 file changed, 5
revision 6d6c37d41ccd511d4417757eab930985a35399cf
Author: thang.d.nguyen
Date: Fri, 8 Oct 2021 07:51:05 +0700
smf: correct merge bundle rolling to single step [#3290]
In mergeBundleRefRollingToSingleStep(), one alias is used to
store m_swAddList then it is reused to store m_swRemoveList.
So it causes the data in m
revision 93938ef5d72e4648f7fababf19abd0eb2818e9be
Author: thang.d.nguyen
Date: Thu, 30 Sep 2021 16:22:07 +0700
smf: make more robustness in BISU upgrade [#3286]
When all procedures executed, the complete action invokes and
clean/remove versioned types. At the time commit campaign,
reboot active SC to test robustness of SMF.
When all procedures executed, the complete action invokes and
clean/remove versioned types. At the time commit campaign,
reboot active SC to test robustness of SMF. Standby SC takes
active role, SMFD on new active start procedure threads and
it failed in addStepModifications due to these versioned
revision 9f957e20f6db4b32e88893e7ecf85a20b31631b1
Author: thang.d.nguyen
Date: Mon, 27 Sep 2021 09:04:49 +0700
smf: make more robustness in BISU upgrade [#3286]
When all procedures executed, the complete action invokes
and clean/remove versioned types. At the time commit campaign,
reboot active SC to test robustness of SMF.
When all procedures executed, the complete action invokes
and clean/remove versioned types. At the time commit campaign,
reboot active SC to test robustness of SMF. Standby SC takes
active role, SMFD on new active start procedure threads and it
failed in addStepModifications due to these versioned
In large cluster size(lager than 36 nodes), and many
components reside on each node. The runtime of nodes
(AdminState and OperationalState) are take time to
update in IMM and it causes the application get wrong
state of node in IMM instead AMF already update its
data base.
This enhancement is to
revision 030211d89cd4412be4140de9a62394fb5fe7bd13
Author: thang.d.nguyen
Date: Thu, 23 Sep 2021 16:46:10 +0700
amf: update runtime attributes of node to IMM in sync [#3285]
In large cluster size(lager than 36 nodes), and many
components reside on each node. The runtime of nodes
(AdminState and OperationalState) are
Only retry on SA_AIS_ERR_FAILED_OPERATION if the
the ccb was resoure aborted and not in error state.
---
python/pyosaf/utils/__init__.py | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/python/pyosaf/utils/__init__.py b/python/pyosaf/utils/__init__.py
index
revision 6d86094bcfbbd0e09e72859a90c9b7d7675a810d
Author: thang.d.nguyen
Date: Mon, 9 Aug 2021 06:53:58 +0700
pyosaf: correct retry on failed operation [#3276]
Only retry on SA_AIS_ERR_FAILED_OPERATION if the
the ccb was resoure aborted and not in error state.
Complete diffstat:
--
python/pyosaf/utils/__ini
revision 463fafd7b9a203458033f6c41109bcb2f1e19063
Author: thang.d.nguyen
Date: Fri, 23 Jul 2021 09:47:39 +0700
amfnd: fix crash during shutdown [#3275]
Node shutdown is followed by node failovered. The NPI component
was uninstantiated after all assignment removed.
In function processes component uninstantiated event in
termina
Node shutdown is followed by node failovered. The NPI component
was uninstantiated after all assignment removed.
In function processes component uninstantiated event in
terminating state. It tries to access the csi then cause
the crash. It should ignore if csi of NPI comp removed
during the
CSICOMP_STATE_INFO_MSG event msg will be pushed
into the queue to processed later. But at the time
processing the node has already left. In this case
the creation comp-csi record must be ignored to avoid
amfd crashed.
---
src/amf/amfd/csi.cc | 6 ++
1 file changed, 6 insertions(+)
diff --git
revision 854a8e03042d6a53a45b903262f5197a52a87525
Author: thang.d.nguyen
Date: Wed, 23 Jun 2021 09:48:02 +0700
amf: ignore re-create comp-csi assignment if node left [#3269]
CSICOMP_STATE_INFO_MSG event msg will be pushed
into the queue to processed later. But at the time
processing the node has al
The initialize value can be not zero, so no need to
assert this check. Just log this as a warning.
---
src/imm/agent/imma_init.cc | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/src/imm/agent/imma_init.cc b/src/imm/agent/imma_init.cc
index 8d3b120b6..bde657158 100644
---
revision 8a5dd83879f2665f800885fcc1dec2e6deb2f2b4
Author: thang.d.nguyen
Date: Thu, 27 May 2021 11:55:14 +0700
imm: not assert when checking initial value in IMMA [#3265]
The initialize value can be not zero, so no need to
assert this check. Just log this as a warning.
Complete diffstat:
-
In one step upgrade, during the lock nodegroup. The timeout can happen
and it causes the upgrade failed. By retrying if the return code of
saImmOmAdminOperationInvoke_2() is SA_AIS_ERR_NO_OP, the lock is
considered as successfully.
---
src/smf/smfd/SmfAdminState.cc | 12
1 file
revision cfe6e81da1531fd8120e3a151d2d1f8c6aa64b2b
Author: thang.d.nguyen
Date: Tue, 18 May 2021 10:17:00 +0700
smf: enhance smf to handle timeout in one step upgrade [#3262]
In one step upgrade, during the lock nodegroup. The timeout can happen
and it causes the upgrade failed. By retrying if the r
In one step upgrade, during the lock nodegroup. The timeout can happen
and it causes the upgrade failed. By retrying if the return code of
saImmOmAdminOperationInvoke_2() is SA_AIS_ERR_NO_OP, the lock is
considered as successfully.
---
src/smf/smfd/SmfAdminState.cc | 13 -
1 file
revision 63272ea1359832e349bc3a904e6bceea20d150e8
Author: thang.d.nguyen
Date: Tue, 18 May 2021 09:00:39 +0700
smf: enhance smf to handle timeout in one step upgrade [#3262]
In one step upgrade, during the lock nodegroup. The timeout can happen
and it causes the upgrade failed. By retrying if the r
revision d7829d5d8866aecee18b37827b4e0ce230ff92c3
Author: thang.d.nguyen
Date: Fri, 14 May 2021 09:49:05 +0700
smf: enhance smf to handle timeout in one step upgrade [#3262]
In one step upgrade, during the lock nodegroup. The timeout can happen
and it causes the upgrade failed. By retrying if the r
In one step upgrade, during the lock nodegroup. The timeout can happen
and it causes the upgrade failed. By retrying if the return code of
saImmOmAdminOperationInvoke_2() is SA_AIS_ERR_NO_OP, the lock is
considered as successfully.
---
src/smf/smfd/SmfAdminState.cc | 16
1 file
- Use non-blocking in connect new socket.
- Dtm does not exit when connect failure.
---
src/dtm/dtmnd/dtm_node.h | 1 +
src/dtm/dtmnd/dtm_node_sockets.cc | 108 +++---
2 files changed, 100 insertions(+), 9 deletions(-)
diff --git a/src/dtm/dtmnd/dtm_node.h
revision ba4c64013b85ba88a585be927eecd9c559009e13
Author: thang.d.nguyen
Date: Thu, 22 Apr 2021 15:38:36 +0700
dtm: correct handling connection failure [#2777]
- Use non-blocking in connect new socket.
- Dtm does not exit when connect failure.
Complete diffstat:
--
src/dtm/dtmnd/dtm_node.h | 1
1705690a0a
Author: thang.d.nguyen
Date: Mon, 19 Apr 2021 14:20:37 +0700
amf: update node member in clm_track_cb [#3257]
During scale-out, node is stopped, 4 events received in order below.
- avd_mds_avnd_up_evh (amfnd up) then clm_track_cb(node join).
At this time node state is still in absen
During scale-out, node is stopped, 4 events received in order below.
- avd_mds_avnd_up_evh (amfnd up) then clm_track_cb(node join).
At this time node state is still in absent but node member is set to true.
Node member is set in clm_track_cb.
- avd_mds_avnd_down_evh (amfnd down) then
- Use non-blocking in connect new socket.
- Dtm does not exit when connect failure.
---
src/dtm/dtmnd/dtm_node.h | 1 +
src/dtm/dtmnd/dtm_node_sockets.cc | 105 +++---
2 files changed, 97 insertions(+), 9 deletions(-)
diff --git a/src/dtm/dtmnd/dtm_node.h
revision cad91231c614579dc0e0c199fb42e80121e762c9
Author: thang.d.nguyen
Date: Wed, 31 Mar 2021 13:24:30 +0700
dtm: correct handling connection failure [#2777]
- Use non-blocking in connect new socket.
- Dtm does not exit when connect failure.
Complete diffstat:
--
src/dtm/dtmnd/dtm_node.h | 1
The IMM model modifyer tool is move from existing locally in SMF
to osaf tool it becomes a global tool.
---
opensaf.spec.in | 88 +++
src/osaf/Makefile.am | 132 ++-
.../immtools}/imm_modify_config/README| 0
revision faa45a561ef206946ce18dd3569b9ff482cf955e
Author: thang.d.nguyen
Date: Tue, 23 Mar 2021 08:48:22 +0700
osaf: Move the immtool for modifying the IMM model to osaf [#3068]
The IMM model modifyer tool is move from existing locally in SMF
to osaf tool it becomes a global tool.
Removed Files:
--
revision 2ef4671be2bdb9a3c3e83fefc892630527f732de
Author: thang.d.nguyen
Date: Tue, 16 Mar 2021 08:38:43 +0700
dtm: correct handling connection failure [#2777]
- Use non-blocking in connect new socket.
- Dtm does not exit when connect failure.
Complete diffstat:
--
src/dtm/dtmnd/dtm_node.h | 1
- Use non-blocking in connect new socket.
- Dtm does not exit when connect failure.
---
src/dtm/dtmnd/dtm_node.h | 1 +
src/dtm/dtmnd/dtm_node_sockets.cc | 105 +++---
src/dtm/dtmnd/dtmd.conf | 2 +-
3 files changed, 98 insertions(+), 10 deletions(-)
There is rare case that thread adminOperationAsyncThread
done before the main thread continue. It caused the main
thread stucked forever.
The fix is to correct this issue.
---
src/smf/smfd/SmfUtils.cc | 9 +++--
1 file changed, 7 insertions(+), 2 deletions(-)
diff --git
revision 0ddccc789b0820232e94cce2a612882cab37ef6a
Author: thang.d.nguyen
Date: Mon, 18 Jan 2021 10:12:25 +0700
smd: avoid smf stucked during admin op in parallel [#3248]
There is rare case that thread adminOperationAsyncThread
done before the main thread continue. It caused the main
thread stucked forever
37aaf84464
Author: thang.d.nguyen
Date: Tue, 10 Nov 2020 14:49:18 +0700
amf: correct behavior SU restart [#3233]
During standby SU restarts, active SU is failover. The standby SU
need re-assignment standby then take over active assignment.
This is to correct the issue in the ticket #3207.
During standby SU restarts, active SU is failover. The standby SU
need re-assignment standby then take over active assignment.
This is to correct the issue in the ticket #3207.
---
src/amf/amfnd/comp.cc | 3 +--
src/amf/amfnd/susm.cc | 55 +--
2 files
aa52bea0e1
Author: thang.d.nguyen
Date: Tue, 10 Nov 2020 05:38:49 +0700
amf: fix assignment loop for NWA SU [#3233]
During SU restart, component is instantiated failed. It makes
component failover. In avnd_su_si_oper_done(), the buffered msg
need removing if the SU is in disable oper state.
During SU restart, component is instantiated failed. It makes
component failover. In avnd_su_si_oper_done() the initial
buffer assignments invoked and it caused the loop of assigning to
QUIESED.
Only initial buffer assignments if the SU is ENABLE.
---
src/amf/amfnd/susm.cc | 3 ++-
1 file
revision 06e337ea774d938f390f074208b13803ac053e99
Author: thang.d.nguyen
Date: Mon, 9 Nov 2020 11:46:41 +0700
amf: fix crashes during restart NWA SU [#3233]
During SU restart, component is instantiated failed. It makes
component failover. In avnd_su_si_oper_done() the initial
buffer assignments invoked and it caused the loop of
10f
Author: thang.d.nguyen
Date: Thu, 22 Oct 2020 14:05:08 +0700
amf: fix coredump in start up [#3186]
During node start up, it loses connection with Active SC.
At that time CLM join cluster and AMFND tries to convert
CLM node to AMF node. But IMMND at that time is down
(e.i, unregister the
During node start up, it loses connection with Active SC.
At that time CLM join cluster and AMFND tries to convert
CLM node to AMF node. But IMMND at that time is down
(e.i, unregister then register with MDS). The IMM OM API
return asap and cause the coredump.
Need to retry in this case for more
f8e
Author: thang.d.nguyen
Date: Mon, 12 Oct 2020 09:45:22 +0700
amf: fix coredump in start up [#3186]
During node start up, it loses connection with Active SC.
At that time CLM join cluster and AMFND tries to convert
CLM node to AMF node. But IMMND at that time is down
(e.i, unregister the
During node start up, it loses connection with Active SC.
At that time CLM join cluster and AMFND tries to convert
CLM node to AMF node. But IMMND at that time is down
(e.i, unregister then register with MDS). The IMM OM API
return asap and cause the coredump.
Solution is to add some retry to
revision fa78173f280133ceb47224bfbaf9e83b96873fc5
Author: thang.d.nguyen
Date: Sat, 3 Oct 2020 09:58:47 +0700
amf: ignore sufailover when shutdown initiated [#3223]
When active assignment is on going, node shutdown and
sufailover happened. Amfnd tries to sufailover but
not successful. Stop node stucks due to amfnd
When active assignment is on going, node shutdown and
sufailover happened. Amfnd tries to sufailover but
not successful. Stop node stucks due to amfnd wait until
csiSetCallbackTimeout.
Amfnd needs to stop this SU-SI assigment and move it
to unassigned state.
---
src/amf/amfnd/susm.cc | 11
During SURestart, another compoenent is failovered. The
SU-SI modify event message need buffering to process later.
The new active assignment on SU need processing after it is
instantiated.
---
src/amf/amfnd/comp.cc | 3 ++-
src/amf/amfnd/susm.cc | 62 ++-
revision 9f2218fe38fc1f48db08936b30df1e05ac5b3b95
Author: thang.d.nguyen
Date: Fri, 7 Aug 2020 08:31:14 +0700
amfnd: handle component failover during SURestart [#3207]
During SURestart, another compoenent is failovered. The
SU-SI modify event message need buffering to process later.
The new active assignment on SU need
revision b0086e3c5da87fad844e76c8c648f6dc6e7ae73a
Author: thang.d.nguyen
Date: Wed, 8 Jul 2020 09:37:52 +0700
lgs: not invoke saAmfComponentErrorReport() [#3197]
Should not invoke saAmfComponentErrorReport() to AMF
before exit with failure.
In case invoking, AMF don't know how to handle it.
And logd does not s
Should not invoke saAmfComponentErrorReport() to AMF
before exit with failure.
In case invoking, AMF don't know how to handle it.
And logd does not start again.
---
src/log/README | 8
src/log/logd/lgs_mbcsv.cc| 16
src/log/logd/lgs_oi_admin.cc | 7
Should not invoke saAmfComponentErrorReport() to AMF
before exit with failure.
In case invoking, AMF don't know how to handle it.
And logd does not start again.
---
src/log/logd/lgs_mbcsv.cc| 16
src/log/logd/lgs_oi_admin.cc | 7 +++
src/log/logd/lgs_util.cc | 4
revision 053cd41fb7cfc3d94197df31cc44be3395ae7e4b
Author: thang.d.nguyen
Date: Thu, 2 Jul 2020 13:58:23 +0700
lgs: correct inform failure to AMF [#3197]
Should not invoke saAmfComponentErrorReport() to AMF
before exit with failure.
In case invoking, AMF don't know how to handle it.
And logd does not start again.
Complet
Fix definitely lost reported by valgrind.
---
src/base/daemon.c | 5 +++--
src/log/logd/lgs_imm.cc | 7 +++
src/log/logd/lgs_mbcsv.cc | 2 ++
3 files changed, 12 insertions(+), 2 deletions(-)
diff --git a/src/base/daemon.c b/src/base/daemon.c
index 48a0665f2..71eba7bff 100644
---
uot; above):
-
revision d28ee50720d5e57edba6ee5c27e8b2bebb0638fa
Author: thang.d.nguyen
Date: Thu, 11 Jun 2020 11:59:10 +0700
lgs: fix memory leak reeported by valgrind [#3195]
Fix definitely lost reported by valgrind.
Complete diffstat:
--
src/bas
Fix definitely lost reported by valgrind.
---
src/base/daemon.c | 6 --
src/log/logd/lgs_imm.cc | 8
src/log/logd/lgs_mbcsv.cc | 2 ++
3 files changed, 14 insertions(+), 2 deletions(-)
diff --git a/src/base/daemon.c b/src/base/daemon.c
index 48a0665f2..2b23b43c3 100644
---
revision 23f7edb74a8459f6c051e9eb3daddbf4d33f6e9d
Author: thang.d.nguyen
Date: Tue, 2 Jun 2020 10:49:55 +0700
lgs: fix memory leak reeported by valgrind [#3195]
Fix definitely lost reported by valgrind.
Complete diffstat:
--
src/base/daemon.c | 2 --
src/log/logd/lgs_imm.cc | 8
sr
Fix definitely lost reported by valgrind.
---
src/base/daemon.c | 2 --
src/log/logd/lgs_imm.cc | 8
src/log/logd/lgs_mbcsv.cc | 1 +
3 files changed, 9 insertions(+), 2 deletions(-)
diff --git a/src/base/daemon.c b/src/base/daemon.c
index 48a0665f2..56f5aa8ff 100644
---
revision 1b2b6f0406ac092696ac160dae8e9eaac12b0107
Author: thang.d.nguyen
Date: Wed, 20 May 2020 14:32:07 +0700
amfd: process IMM evt higher than MBCSV evt in polling [#3190]
Amfd on standby somehow busy/stuck in a while. The si object was
removed and added again. Then this SI was assigned. Then amfd on standb
Amfd on standby somehow busy/stuck in a while. The si object was
removed and added again. Then this SI was assigned. Then amfd on standby
continue working again. But it will process event from MBCSV first
(e.g, saAmfSINumCurrActiveAssignments) and it will due to the SI not
created in data based
1 - 100 of 195 matches
Mail list logo
