I upgraded my test instances to 6.3.2 this week and encountered something odd while upgrading my indexer cluster. I had first successfully upgraded a standalone instance and a heavy forwarder, but my peer indexers failed when I tried to upgrade them. Every time I upgrade like this I first stop the Splunk service on both peers and the master before I do anything, but I eventually figured out that I had to have the master running during this peer upgrade. I'm not sure why this upgrade was different, but I thought I would share this information for anyone else who was in a similar situation.
