Hi,
it may be that the MDS thread is a real time thread SCHED_RR and the
main thread is SCHED_OTHER, and as long as the
MDS thread do not yield, the main thread will not be scheduled. This may
be more common in an container
environment, I think we need to look into, if this starts to be a more
common problem, how to correct this and not be dependent on
scheduling policies. I'm looking at two cases that looks similar.
/BR HansN
On 07/06/2017 06:42 AM, praveen malviya wrote:
Hi Gary,
When SC joins back after SC Absence state, there will not be many
messages in the queue to be processed even in large cluster also.
Also there is no error like ipc send failure.
Is AMFD stuck somewhere and it could not take any message from queue?
I think raising priority will not solve problem in such a situation.
Thanks
Praveen
On 23-Jun-17 7:18 AM, Gary Lee wrote:
---
src/amf/amfd/ndmsg.cc | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/src/amf/amfd/ndmsg.cc b/src/amf/amfd/ndmsg.cc
index 11bc8ac35..9bfab6fd8 100644
--- a/src/amf/amfd/ndmsg.cc
+++ b/src/amf/amfd/ndmsg.cc
@@ -371,7 +371,12 @@ uint32_t avd_n2d_msg_rcv(AVD_DND_MSG *rcv_msg,
NODE_ID node_id,
evt->info.avnd_msg = rcv_msg;
- if (m_NCS_IPC_SEND(&cb->avd_mbx, evt, NCS_IPC_PRIORITY_HIGH) !=
+ NCS_IPC_PRIORITY priority = NCS_IPC_PRIORITY_HIGH;
+ if (evt->rcv_evt == AVD_EVT_NODE_UP_MSG) {
+ priority = NCS_IPC_PRIORITY_VERY_HIGH;
+ }
+
+ if (m_NCS_IPC_SEND(&cb->avd_mbx, evt, priority) !=
NCSCC_RC_SUCCESS) {
LOG_ER("%s: ncs_ipc_send failed", __FUNCTION__);
avsv_dnd_msg_free(rcv_msg);
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Opensaf-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensaf-devel
