- **Milestone**: 4.6.2 --> 5.0.RC1
- **Comment**:

This five second delay is configurable using the variable 
CLMNA_ELECTION_DELAY_TIME in /etc/opensaf/clmna.conf Could you re-test after 
changing this setting to something very low, like 100 ms:

export CLMNA_ELECTION_DELAY_TIME=100

The default value is 5s, which I think is reasonable in a cluster with many 
system controllers. The delay only happens during initial cluster start (and 
cluster re-start), and it is there to reduce the probability for split-brain. 
However, in a cluster with only two system controllers, it is probably safe to 
use a shorter delay (or even zero?), since it will be just as good - or just as 
bad - as before. Mabe for backwards compatibility the default ought to be 
100ms, but then the documentation should be really clear that the default value 
is not optimal for clusters with many system controllers.



---

** [tickets:#1724] Opensaf is taking more time to start on active controller **

**Status:** unassigned
**Milestone:** 5.0.RC1
**Created:** Wed Apr 06, 2016 07:03 AM UTC by Srikanth R
**Last Updated:** Wed Apr 06, 2016 07:03 AM UTC
**Owner:** nobody


Setup :  single controller with opensaf changeset 7436 5.0 FC

Opensaf with 5.0 FC is taking more time  to start , when compared to opensaf 
4.7GA.

CLMNA is taking 5 seconds to promote the first node of the cluster to system 
controller and another two seconds to declare as active in 5.0. In 4.7,RDE 
takes 2 to three seconds to declare the first node as active.

Below is syslog on 5.0.


Apr  4 20:34:03 CONTROLLER-1 opensafd: Starting OpenSAF Services(5.0.FC - ) 
(Using TIPC)
Starting OpenSAF Services (Using TIPC):Apr  4 20:34:03 CONTROLLER-1 kernel: 
[22205.292335] 
...
Apr  4 20:34:03 CONTROLLER-1 kernel: [22205.297844] TIPC: Enabled bearer 
<eth:eth0>, discovery domain <1.1.0>, priority 10
Apr  4 20:34:03 CONTROLLER-1 osafclmna[10490]: Started
Apr  4 20:34:08 CONTROLLER-1 osafclmna[10490]: NO Starting to promote this node 
to a system controller
Apr  4 20:34:08 CONTROLLER-1 osafrded[10499]: Started
Apr  4 20:34:08 CONTROLLER-1 osaffmd[10508]: Started
Apr  4 20:34:08 CONTROLLER-1 osafimmd[10518]: logtrace: trace enabled to file 
/var/log/opensaf/osafimmd, mask=0xffffffff
....
Apr  4 20:34:08 CONTROLLER-1 osafrded[10499]: NO Requesting ACTIVE role
Apr  4 20:34:08 CONTROLLER-1 osafrded[10499]: NO RDE role set to Undefined
Apr  4 20:34:10 CONTROLLER-1 osafrded[10499]: NO Running 
'/usr/lib64/opensaf/opensaf_sc_active' with 0 argument(s)
Apr  4 20:34:10 CONTROLLER-1 osafrded[10499]: NO Switched to ACTIVE from 
Undefined
Apr  4 20:34:10 CONTROLLER-1 osaffmd[10508]: NO Starting activation 
supervision: 300000ms
Apr  4 20:34:10 CONTROLLER-1 osafimmnd[10529]: NO IMMD service is UP ... 
ScAbsenseAllowed?:0 introduced?:0
Apr  4 20:34:10 CONTROLLER-1 osafimmd[10518]: IN node with dest ADDED 
564117257723920
Apr  4 20:34:10 CONTROLLER-1 osafimmnd[10529]: NO SERVER STATE: 
IMM_SERVER_ANONYMOUS --> IMM_SERVER_CLUSTER_WAITING
...
Apr  4 20:34:10 CONTROLLER-1 osafimmd[10518]: NO Attached Nodes:1 Accepted 
nodes:0 KnownVeteran:0 doReply:1
Apr  4 20:34:10 CONTROLLER-1 osafimmd[10518]: NO First IMMND on SC found at 
2010f this IMMD at 2010f. Cluster is loading, *not* 2PBE => designating that 
IMMND as coordinator
Apr  4 20:34:10 CONTROLLER-1 osafimmnd[10529]: NO This IMMND is now the NEW 
Coord
Apr  4 20:34:10 CONTROLLER-1 osafimmnd[10529]: NO SETTING COORD TO 1 CLOUD PROTO
Apr  4 20:34:13 CONTROLLER-1 osafimmnd[10529]: NO SERVER STATE: 
IMM_SERVER_CLUSTER_WAITING --> IMM_SERVER_LOADING_PENDING
...
Apr  4 20:34:14 CONTROLLER-1 osafamfnd[10590]: NO Assigned 
'safSi=SC-2N,safApp=OpenSAF' ACTIVE to 'safSu=SC-1,safSg=2N,safApp=OpenSAF'
Apr  4 20:34:14 CONTROLLER-1 opensafd: OpenSAF(5.0.FC - ) services successfully 
started


Below is the syslog output for startup on opensaf 4.7 FC.

Apr  5 15:39:50 CONTROLLER-2 opensafd: Starting OpenSAF Services(4.7.0 - ) 
(Using TIPC)
Starting OpenSAF Services (Using TIPC):Apr  5 15:39:50 CONTROLLER-2 kernel: [ 
5024.923512] TIPC: Activated (version 2.0.0)
Apr  5 15:39:50 CONTROLLER-2 kernel: [ 5024.923603] NET: Registered protocol 
family 30
...
Apr  5 15:39:50 CONTROLLER-2 kernel: [ 5024.930263] TIPC: Enabled bearer 
<eth:eth3>, discovery domain <1.1.0>, priority 10
Apr  5 15:39:50 CONTROLLER-2 osafrded[4071]: Started
Apr  5 15:39:52 CONTROLLER-2 osafrded[4071]: NO No peer available => Setting 
Active role for this node
Apr  5 15:39:52 CONTROLLER-2 osaffmd[4080]: Started
Apr  5 15:39:52 CONTROLLER-2 osafimmd[4090]: Started
Apr  5 15:39:52 CONTROLLER-2 osafimmnd[4101]: Started
Apr  5 15:39:53 CONTROLLER-2 osafimmd[4090]: NO New IMMND process is on ACTIVE 
Controller at 2020f
Apr  5 15:39:53 CONTROLLER-2 osafimmd[4090]: NO First SC IMMND (OpenSAF 4.4 or 
later) attached 2020f
...
Apr  5 15:39:53 CONTROLLER-2 osafimmnd[4101]: NO SERVER STATE: 
IMM_SERVER_ANONYMOUS --> IMM_SERVER_CLUSTER_WAITING
....
Apr  5 15:39:56 CONTROLLER-2 osafimmnd[4101]: NO SERVER STATE: 
IMM_SERVER_CLUSTER_WAITING --> IMM_SERVER_LOADING_PENDING
...
Apr  5 15:39:57 CONTROLLER-2 osafamfnd[4168]: NO Assigned 
'safSi=SC-2N,safApp=OpenSAF' ACTIVE to 'safSu=SC-2,safSg=2N,safApp=OpenSAF'
Apr  5 15:39:57 CONTROLLER-2 opensafd: OpenSAF(4.7.0 - ) services successfully 
started



---

Sent from sourceforge.net because [email protected] is 
subscribed to https://sourceforge.net/p/opensaf/tickets/

To unsubscribe from further messages, a project admin can change settings at 
https://sourceforge.net/p/opensaf/admin/tickets/options.  Or, if this is a 
mailing list, you can unsubscribe from the mailing list.
------------------------------------------------------------------------------
_______________________________________________
Opensaf-tickets mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensaf-tickets

Reply via email to