The MDS send call(AVA_SEND) returns after sending the message! Subsequently the
process exit would generate an AVA down. But, i think theoretically the chances
that an AVA_RESP and AVA_DOWN could reach out of order is extremely less(the
only situation i can think of is when the main thread of application has
starved the MDS thread) *than* when comparing to the chances of an out of order
of AVA_FINALIZE and AVA_DOWN.
The AVA_FINALIZE results in a MDS shutdown.
I just checked, in the numerous iterations i tried on a 64-bit setup, the AMF
response always reached first and the ava down later!
Having said that, if the component timesout on the callback response or returns
error, AMF proceeds to cleanup the component.
The following are the possible outcomes in this scenario of admin restart
operation:
(a) IF all goes well, the component gets restarted(re INSTANTIATED)
(b) IF component timesout or returns error then AMF would invoke the cleanup
script of the component and then component gets restarted(re INSTANTIATED).
(c) IF it can happen that the AVA_DOWN(process exit) message had preceeded the
AVA_send, AMF would invoke the cleanup script and then component gets restarted.
While the end result is the same for all the cases, the case (c) could still be
considered safe because the component would have responded to AMF only after
preparing itself for the exit/kill! Hmm..
---
** [tickets:#501] amf: No node directors register to AMF within time after "#7
cleanup instead of terminate used at component restart"**
**Status:** unassigned
**Created:** Thu Jul 11, 2013 07:47 AM UTC by Ingvar Bergström
**Last Updated:** Thu Jul 18, 2013 01:31 PM UTC
**Owner:** nobody
After introduction of patches solving "#7 cleanup instead of terminate used at
component restart”, no node directors registers to AMF within time according to
messages log.
I have tried SMFND, CPND, GLND and MQND.
It seems however that the main routines of the node director daemons are not
started until 10 seconds after the terminate callback (after the registration
timeout).
It is very easy to see the fault by entering command "amf-adm restart
safComp=xxxND,safSu=SC-1,safSg=NoRed,safApp=OpenSAF"
---
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.
------------------------------------------------------------------------------
See everything from the browser to the database with AppDynamics
Get end-to-end visibility with application monitoring from AppDynamics
Isolate bottlenecks and diagnose root cause in seconds.
Start your free trial of AppDynamics Pro today!
http://pubads.g.doubleclick.net/gampad/clk?id=48808831&iu=/4140/ostg.clktrk
_______________________________________________
Opensaf-tickets mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensaf-tickets
