]> git.entuzijast.net Git - imunes.git/commit
When cleaning up a running experiment, always attempt to kill
authormarko <marko>
Mon, 8 Oct 2007 23:28:13 +0000 (23:28 +0000)
committermarko <marko>
Mon, 8 Oct 2007 23:28:13 +0000 (23:28 +0000)
commitab30d7214fd4396fb5f51abe7e5341960fec87d0
tree0b1fc82a6937b2c7909a99ce02f5e5522b6c7454
parentbd2d15abc86d02990f77a82af25a201a8c18b1d3
When cleaning up a running experiment, always attempt to kill
all processes in all vimages first.

It seems that quagga has a bug which sets all zebra processes in an
infinite loop if a network interface disappears while the suite is
running.  Before this change, we were "unplugging" the ng_eifaces
while zebra & co. were still running, resulting in each vimage
attempting to hog the CPU.  This would leave very little CPU time
for IMUNES scripts, and for any user interaction with the system
for that matter, so cleaning up a more complex topology could
last for quite a while.

In short, the cleanup should be now roughly as fast as it used to be
in 4.11 days.

Bug found by:
Submitted by:
Reviewed by:
Approved by:
Obtained from:
exec.tcl