When the PBE hung, amfd can process the events with
below order when a node was started then stop then started
- clm_track_cb for node down event
- clm_track_cb for second node up event
- avd_mds_avnd_down_evh was called to process amfnd down event
And it cause the node can not join the cluster.
-
When the PBE hung, amfd can process the events with
below order when a node was started then stop then started
- clm_track_cb for node down event
- clm_track_cb for second node up event
- avd_mds_avnd_down_evh was called to process amfnd down event
And it cause the node can not join the cluster.
-
Hi Thang,
Ack from me.
Thanks
Minh
On 11/6/19 4:11 pm, Minh Hon Chau wrote:
Hi Thang,
I can see it's a race in main thread that how amfd processes the mds
down and clm callback.
Node is going down
<143>1 2019-06-11T15:16:42.157517+10:00 SC-1 osafamfd 272 osafamfd
[meta sequenceId="38507
Hi Thang,
I can see it's a race in main thread that how amfd processes the mds
down and clm callback.
Node is going down
<143>1 2019-06-11T15:16:42.157517+10:00 SC-1 osafamfd 272 osafamfd [meta
sequenceId="38507"] 275:amf/amfd/mds.cc:398 >> avd_mds_svc_evt
<143>1 2019-06-11T15:16:42.157526+1
When the PBE hung, amfd can process the events with
below order when a node was started then stop then started
- clm_track_cb for node down event
- clm_track_cb for second node up event
- avd_mds_avnd_down_evh was called to process amfnd down event
And it cause the node can not join the cluster.
-
