It appears that when I try to make any network connection to a host that is 
either powered down or non existent it takes nearly 3.5 minutes for a timeout 
to occur. I'm transitioning from Linux where I'm used to a host sending 3 ARP 
broadcasts and then generating an ICMP host unreachable message and moving on.

We're seeing this in a high-availability cluster where a node is being powered 
off and it's taking down most of the cluster with it because it's fatally 
interrupting inter-cluster communication with this 3.5 minute timeout rather 
than acknowledging a host is down.

Is there a way to get comparable behavior on Solaris? Would you recommend a 
different approach altogether? So far the closest thing we can find is 
ndd -set /dev/tcp tcp_ip_abort_cinterval 1000
 
 
This message posted from opensolaris.org
_______________________________________________
networking-discuss mailing list
[email protected]

Reply via email to