On Mon, 2006-11-13 at 13:55 -0800, Al Chu wrote: > Hi Jean, > > > Here is what I propose. Instead of using a getDeviceId for the > > keepalive mechanism, we could use a simple empty SOL transmit. (Valid > > sequence number, no acknowledge, length of 0). This will trig no work > > on the BMC side and will no generate also an answer. > > This solution would bother me a bit. I don't think the IPMI spec is > specific enough on the behavior that should occur if an empty SOL packet > is sent. Some other implementations may return an error instead
I note that CVS accepted that patch, and can confirm that the strategy causes problems on at least one BMC implementation, and in the code itself. The problem is that the BMC never acknowledges the keepalive packet, and so every 30 seconds you have a huge stall because ipmitool exhausts retries, but doesn't close connection. BTW, this also means that ipmitool didn't close the connection when it never received a keepalive response, so on this count also there seems to be a problem that it fails to detect what could be a dead connection. ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys - and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _______________________________________________ Ipmitool-devel mailing list Ipmitool-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ipmitool-devel
