Description of problem: When a node is fenced there is nothing reported in the syslog by the fencer. The only way I know who's fencing who is because I added debugging info to fence_apc. In 6.0 and 5.2.1 there were messages stating who was about to be fenced and then weither or not it was successful. The only way to know what's happening now is to run fenced with the debugging flag and in the foreground. How reproducible: Always Steps to Reproduce: 1. attempt recovery senario
How does this look? RCS file: /cvs/cluster/cluster/fence/fenced/recover.c,v retrieving revision 1.2 diff -r1.2 recover.c 181a182 > syslog(LOG_INFO, "fencing node \"%s\"", node->name); 187a189,191 > > syslog(LOG_INFO, "fence \"%s\" %s", node->name, > error ? "failed" : "success");
looks good. fix verified.
This is good, but it is also necessary for troubleshooting to have the output of the fence agent in here, as it was for 6.0.0. I'm currently in a state where a log is reporting: Sep 22 14:23:32 tng3-1 fenced[2267]: fencing node "tng3-3" Sep 22 14:23:32 tng3-1 fenced[2267]: fence "tng3-3" failed over and over again with no additional info on what's gone wrong.
With today's checkin it worked with my test.
the output of the fence agent does now appear. fix verified.
Updating version to the right level in the defects. Sorry for the storm.