HAST configuration?

[Pawel Jakub Dawidek]

On Sat, Apr 24, 2010 at 10:49:10AM +0900, hiroshi at soupacific.com wrote:
> Hi !
> 
> I checked ucarp_up.sh and modified like this.
> 
> # Wait few seconds for provider to appear.
> ##for i in `jot 50`; do
> for i in `jot 2000`; do
> 	[ -c "${device}" ] && break
> 	sleep 0.1
> done
> 
> Then 30 seconds to a minute after login the box, /dev/hast and /hapool 
> appears.
> 
> So hastctl role primary test takes a while when nodeb is downed.
> 
> I hope this info help you some.

That makes sense. Connection attempts are made in background, but with
one exception - role change. When the role is changed to primary I try to
connect to secondary _before_ creating /dev/hast/ provider, so if
secondary is fine, we will connect and everything will work. I wanted
to avoid situation where we create /dev/hast/ provider, system will
start to use it and we will connect to secondary a bit later. This will
trigger synchronization, which isn't desiriable when both nodes are in
good shape. I think the way to fix it is would be to decrease initial
connection timeout. I'll look into this.

-- 
Pawel Jakub Dawidek                       http://www.wheelsystems.com
pjd at FreeBSD.org                           http://www.FreeBSD.org
FreeBSD committer                         Am I Evil? Yes, I Am!
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 196 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-fs/attachments/20100424/831869b1/attachment.pgp