Commit bde090d9 authored by Mike Hibler's avatar Mike Hibler
Browse files

Make the event system startup failure after a swap modify be non-fatal.

The logic here is:
 * this is the last step and that a failure at this point causes a
   complete swapout of the experiment, possibly losing important state.
 * we can probably recover from this by just manually starting the
   event system.
parent 3dac3cb8
......@@ -2,7 +2,7 @@
#
# EMULAB-COPYRIGHT
# Copyright (c) 2000-2011 University of Utah and the Flux Group.
# Copyright (c) 2000-2012 University of Utah and the Flux Group.
# All rights reserved.
#
use English;
......@@ -1545,10 +1545,23 @@ sub doSwapin($) {
print "Starting the event system.\n";
TBDebugTimeStamp("eventsys_control started");
if (system("eventsys_control start $pid,$eid")) {
tberror({type => 'secondary', severity => SEV_SECONDARY,
error => ['eventsys_start_failed']},
"Failed to start the event system.");
return 1;
#
# If we are doing a modify, a failure here will result in
# a swapout and associated angry user. At this point I deem
# it better to let the swapin succeed without the event
# system running and just let testbed-ops know so that we
# can run eventsys_control manually.
#
if ($type == MODIFY) {
tbwarn "Failed to start the event system, ".
"must restart manually.";
return 0;
} else {
tberror({type => 'secondary', severity => SEV_SECONDARY,
error => ['eventsys_start_failed']},
"Failed to start the event system.");
return 1;
}
}
TBDebugTimeStamp("eventsys_control finished");
}
......
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