'Re: [Spread-users] Local connection heartbeats'

[prev in list] [next in list] [prev in thread] [next in thread] 

List:       spread-users
Subject:    Re: [Spread-users] Local connection heartbeats
From:       "Alec H. Peterson" <alec.peterson () messagesystems ! com>
Date:       2006-09-07 19:54:51
Message-ID: 2CA9A3A7-2E70-4463-B429-C3ACB12D90D1 () messagesystems ! com
[Download RAW message or body]

[Attachment #2 (multipart/alternative)]

Hi,

On Sep 8, 2006, at 3:26, Camp, TracyX E wrote:

> I’ve worked on a forked variant of spread in the recent past.  I’m  
> planning on posting a few of the more useful bits of the variant  
> fairly soon.  Anyways we did have a watchdog facility in our  
> variant.  It worked by having an API that scheduled a timed event  
> to the main client event loop (i.e. an E_queue() call).  The  
> callback the API specified would send a heartbeat message to the  
> local spread daemon and schedule another upcall.  The spread-daemon  
> had an aliveness counter for each socket that if it wasn’t updated  
> in a certain number of seconds would cause various things to  
> happen.  In our case we took the entire spread daemon down, but in  
> your case it could simply do a sess_kill() on the offended socket,  
> which would in turn cause membership to change.
Interesting, that sounds like it would work perfectly.  Thanks for  
the info!

Alec

[Attachment #5 (unknown)]

<HTML><BODY style="word-wrap: break-word; -khtml-nbsp-mode: space; -khtml-line-break: \
after-white-space; ">Hi,<DIV><BR><DIV><DIV>On Sep 8, 2006, at 3:26, Camp, TracyX E \
wrote:</DIV><BR class="Apple-interchange-newline"><BLOCKQUOTE type="cite"> <DIV \
class="Section1"><P class="MsoNormal"><FONT size="2" face="Arial"><SPAN \
style="font-size:10.0pt; font-family:Arial">I’ve worked on a forked variant of spread \
in the recent past.  I’m planning on posting a few of the more useful bits of the \
variant fairly soon.  Anyways we did have a watchdog facility in our variant.  It \
worked by having an API that scheduled a timed event to the main client event loop \
(i.e. an E_queue() call).  The callback the API specified would send a heartbeat \
message to the local spread daemon and schedule another upcall.  The spread-daemon \
had an aliveness counter for each socket that if it wasn’t updated in a certain \
number of seconds would cause various things to happen.  In our case we took the \
entire spread daemon down, but in your case it could simply do a sess_kill() on the \
offended socket, which would in turn cause membership to \
change.<O:P></O:P></SPAN></FONT></P></DIV></BLOCKQUOTE>Interesting, that sounds like \
it would work perfectly.  Thanks for the info!<BR></DIV><DIV><BR \
class="khtml-block-placeholder"></DIV><DIV>Alec</DIV><DIV><BR \
class="khtml-block-placeholder"></DIV><BR></DIV></BODY></HTML>

_______________________________________________
Spread-users mailing list
Spread-users@lists.spread.org
http://lists.spread.org/mailman/listinfo/spread-users

[prev in list] [next in list] [prev in thread] [next in thread] 