Sorry, for the swedish, translation:
I just noticed that the poll loop in the MDS thread is first checking the DSock
fd and then the BSRSock fd.
This I assume would mean that an ava_down can override a response message. I
don’t know if this order
can be changed in the poll loop without causing other problems.
/Bertil
From: Bertil Engelholm [mailto:[email protected]]
Sent: den 18 september 2013 13:03
To: [opensaf:tickets]
Subject: [opensaf:tickets] Re: #570 AMF: ava down event ignored in TERMINATING
state
Tja, jag såg just I MDS-koden att man I poll-loopen först tittar på DSock och
sedan BSRSock.
Det borde kunna leda till att meddelandet om att porten försvunnit (ava_down)
kan köra om
responsmeddelandet. Vet ej om det kan ställa till problem om man ändrar denna
ordningen i poll-loopen.
/Bertil
From: Hans Feldt [mailto:[email protected]]
Sent: den 18 september 2013 08:46
To: Ticket 570
Subject: [tickets]<http://sourceforge.net/p/opensaf/tickets/_discuss/>
[opensaf:tickets] #570 AMF: ava down event ignored in TERMINATING state
________________________________
[tickets:#570]<http://sourceforge.net/p/opensaf/tickets/570/>http://sourceforge.net/p/opensaf/tickets/570/
AMF: ava down event ignored in TERMINATING state
Status: unassigned
Created: Wed Sep 18, 2013 06:45 AM UTC by Hans Feldt
Last Updated: Wed Sep 18, 2013 06:45 AM UTC
Owner: nobody
If a component crash or exit in context of the terminate callback, AMF will not
use the "ava down" event to trigger cleanup and finish component termination.
Instead the CallbackTimeout is awaited which can be very long.
This is a problem if this happens during an upgrade, it will cause the upgrade
to fail potentially leading to system restore.
Ignoring the event (in avnd_err.c) was added in:
changeset: 1646:92e6e65eefc0
user: Nagendra Kumar
[email protected][email protected]<mailto:[email protected]>
date: Thu Aug 26 18:43:31 2010 +0530
summary: Ticket 1433: Allowing dynamic configuration changes for AMF logical
entities
reasons unclear. This has to be revisited. As far as I can tell there is no
"race" between the ava down event and the response message. Normally a process
calls saAmfResponse(OK) and then exit(0). saAmfResponse(OK) under the hood
sends a message and at least TIPC will do "run to completion" meaning it will
post it to the receivers (amfnd) socket receive buffer. After that the process
exit and a topology event is created and written to another socket receive
buffer (MDS lib for amfnd). Eventually the MDS thread in amfnd context will
receiver both messages and write them to amfnd mailbox with the same prio.
________________________________
Sent from sourceforge.net because
[email protected][email protected]<mailto:[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.
________________________________
[tickets:#570]<http://sourceforge.net/p/opensaf/tickets/570/> AMF: ava down
event ignored in TERMINATING state
Status: unassigned
Created: Wed Sep 18, 2013 06:45 AM UTC by Hans Feldt
Last Updated: Wed Sep 18, 2013 06:45 AM UTC
Owner: nobody
If a component crash or exit in context of the terminate callback, AMF will not
use the "ava down" event to trigger cleanup and finish component termination.
Instead the CallbackTimeout is awaited which can be very long.
This is a problem if this happens during an upgrade, it will cause the upgrade
to fail potentially leading to system restore.
Ignoring the event (in avnd_err.c) was added in:
changeset: 1646:92e6e65eefc0
user: Nagendra Kumar [email protected]<mailto:[email protected]>
date: Thu Aug 26 18:43:31 2010 +0530
summary: Ticket 1433: Allowing dynamic configuration changes for AMF logical
entities
reasons unclear. This has to be revisited. As far as I can tell there is no
"race" between the ava down event and the response message. Normally a process
calls saAmfResponse(OK) and then exit(0). saAmfResponse(OK) under the hood
sends a message and at least TIPC will do "run to completion" meaning it will
post it to the receivers (amfnd) socket receive buffer. After that the process
exit and a topology event is created and written to another socket receive
buffer (MDS lib for amfnd). Eventually the MDS thread in amfnd context will
receiver both messages and write them to amfnd mailbox with the same prio.
________________________________
Sent from sourceforge.net because you indicated interest in
https://sourceforge.net/p/opensaf/tickets/570/
To unsubscribe from further messages, please visit
https://sourceforge.net/auth/subscriptions/
---
** [tickets:#570] AMF: ava down event ignored in TERMINATING state**
**Status:** unassigned
**Created:** Wed Sep 18, 2013 06:45 AM UTC by Hans Feldt
**Last Updated:** Wed Sep 18, 2013 06:45 AM UTC
**Owner:** nobody
If a component crash or exit in context of the terminate callback, AMF will not
use the "ava down" event to trigger cleanup and finish component termination.
Instead the CallbackTimeout is awaited which can be very long.
This is a problem if this happens during an upgrade, it will cause the upgrade
to fail potentially leading to system restore.
Ignoring the event (in avnd_err.c) was added in:
changeset: 1646:92e6e65eefc0
user: Nagendra Kumar <[email protected]>
date: Thu Aug 26 18:43:31 2010 +0530
summary: Ticket 1433: Allowing dynamic configuration changes for AMF
logical entities
reasons unclear. This has to be revisited. As far as I can tell there is no
"race" between the ava down event and the response message. Normally a process
calls saAmfResponse(OK) and then exit(0). saAmfResponse(OK) under the hood
sends a message and at least TIPC will do "run to completion" meaning it will
post it to the receivers (amfnd) socket receive buffer. After that the process
exit and a topology event is created and written to another socket receive
buffer (MDS lib for amfnd). Eventually the MDS thread in amfnd context will
receiver both messages and write them to amfnd mailbox with the same prio.
---
Sent from sourceforge.net because [email protected] is
subscribed to http://sourceforge.net/p/opensaf/tickets/
To unsubscribe from further messages, a project admin can change settings at
http://sourceforge.net/p/opensaf/admin/tickets/options. Or, if this is a
mailing list, you can unsubscribe from the mailing list.
------------------------------------------------------------------------------
LIMITED TIME SALE - Full Year of Microsoft Training For Just $49.99!
1,500+ hours of tutorials including VisualStudio 2012, Windows 8, SharePoint
2013, SQL 2012, MVC 4, more. BEST VALUE: New Multi-Library Power Pack includes
Mobile, Cloud, Java, and UX Design. Lowest price ever! Ends 9/20/13.
http://pubads.g.doubleclick.net/gampad/clk?id=58041151&iu=/4140/ostg.clktrk
_______________________________________________
Opensaf-tickets mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensaf-tickets
