Commit 93d78f49 authored by Mike Hibler's avatar Mike Hibler
Browse files

When using the "-e pid,eid" on a firewalled experiment, don't reboot

the firewall itself.  See the comment for more.
parent c6d1231f
#!/usr/bin/perl -wT
# Copyright (c) 2000-2005 University of Utah and the Flux Group.
# Copyright (c) 2000-2006 University of Utah and the Flux Group.
# All rights reserved.
use English;
......@@ -178,8 +178,30 @@ elsif (defined($options{"e"})) {
" There are no nodes reserved in pid/eid $pid/$eid\n");
# If this is a firewalled experiment, don't reboot the firewall.
# If you want to reboot the firewall, you have to do it individually.
# Two reasons. One is that the firewall is not conceptually part of
# the experiment and the user should not "be aware" of it. This
# is a pretty lame reason because pretty much everywhere else, the
# firewall IS part of the experiment. That leads to reason number
# two: rebooting the firewall causes all other nodes in the experiment
# to become disconnected until the firewall reboots. For some machines,
# that are also rebooting as you recall, not getting PXE info for a
# significant amount of time causes them to fail to the next boot.
# For some machines this might mean halting ("Strike any key to continue"),
# as there is no other boot possibility. This means ya gotta come back
# later and reboot all those nodes again.
my $firewall = "";
if (TBExptFirewall($pid, $eid, \$firewall)) {
warn("WARNING: NOT rebooting firewall node $firewall\n");
while (my ($nodeid) = $query_result->fetchrow_array()) {
push(@nodes, $nodeid);
push(@nodes, $nodeid)
if ($nodeid ne $firewall);
else {
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment