Something irritating I’ve noticed awhile ago, but only looked at now. ‘amportal stop’ works by sending ‘core stop gracefully’, and apparently that is causing asterisk to die with segfault. Dunno why yet (and I will look and post back results). Only mentioning that here because the net effect is that safe_asterisk thinks it should restart asterisk for you. Because the stop command is piped into ‘grep -v’ for ‘no such command’ (or somesuch), no obvious error results to indicate an issue, then when you do ‘amportal restart’ you are confounded when you’re told it’s already running. I worked around this by changing the ‘gracefully’ to ‘now’.
This is (was) digium issue 13476. I was running 1.6.0.1. Checking the latest release candidate (asterisk-1.6.0.3-rc1), it looks like this was fixed. I upgraded and retried the repro, and it is in fact fixed now 
I have Asterisk 1.6.1.0 and this issue is still there.
I haven’t noticed that thread earlier and posted my question here http://www.freepbx.org/forum/freepbx/users/error-stopping-freepbx
I am running 1.6.0.9 (ran into an issue in 1.6.1.0 - don’t remember what offhand), and it is fine here.