[Zope] brief storage server downtime

ethan mindlace fremen mindlace@digicool.com
Mon, 12 Feb 2001 12:11:18 -0500


well...

we just modified zdaemon so that it would not stop restarting zope if it 
died within 20 seconds, and modified z2.py so that if anything wierd 
happened there it would notify zdaemon to stop trying.  I tried to start a 
storage server (whose start.py hasn't been changed) with a zope that hadn't 
been compiled (oops).

So there was an error on Aquisition and the process died within a second. 
Zdaemon faithfully restarted it.

Zdaemon, however, has this little buglet that makes the PID change on each 
restart.

So I had 2 processes that were respawning every second with a new PID.  Not 
very easy to kill.

We tried moving start.py to start.save, but that just made the child die 
all the faster.

Eventually we created a start.py that just did time.sleep(3600), which kept 
the child awake long enough to let me kill zdaemon.

Now everything is back in order.

--
ethan mindlace fremen
zopatista community liason