Hi Lindsay,
Another way to test a failure is, if you have access to the switch layer and know 100% which ports the node your wanting to simulate dying is, then just shut all ports down at the same time. This does a good job simulating a node dying.
Joe
On Mon, Dec 7, 2015 at 9:05 AM, Lindsay Mathieson <lindsay.mathieson@xxxxxxxxx> wrote:
I've been stress testing gluster (VM hosting), including testing how well it handles node outages.
Since I can't physically reboot or reset the nodes I've been doing the following:
killall glusterd glusterfsd
Is there a better way of emulating a node failure?
Is there a clean way to close glusterfsd without killing it?
Thanks - Lindsay
_______________________________________________
Gluster-users mailing list
Gluster-users@xxxxxxxxxxx
http://www.gluster.org/mailman/listinfo/gluster-users
_______________________________________________ Gluster-users mailing list Gluster-users@xxxxxxxxxxx http://www.gluster.org/mailman/listinfo/gluster-users
