Redo shutdown to match what it says in the paper; Contact the CMs in
parallel uding ParRun() so that its a lot faster. Also deal with transient errors and retry. Note though, that I do not deal with timeout at the RPC level cause I do not know how to get that error back yet. Also cleanup credential usage, although not active yet until the new CM makes it out; we must use a clearinghouse self signed credential not the bogus slice credential that the clearinghouse was using.
tmcd/freebsd/rc.bridges
0 → 100755