[LON-CAPA-admin] loncnew

[Guy Albertelli II]

Hi Guy,

> I installed 1.1.99.5, and after restarting loncontrol I got 77 loncnew
> processes and 23 lonsql processes running and clogging my cpu.

Err, clogging?  They are for the most part completely idle.
(Looking at s10, the idle loncnew are using about 12 seconds of
processor time a day, which means it is using about 6% of a single
processor a day. Idle lonsql use as close to zero seconds a day as
possible.)


Anyway the lonsql is a bit high should be 3 or 4, not sure why so many
are firing off but they will easily be swaped out and shouldn't effect
anything. 


As for loncnew the current architecture calls for 1 loncnew
process per host in the network.

Which is actually better than before, since you used to have 1 lonc
and 1 lond process per host in the network, now the lond is fired up
only on an as needed basis and trimmed back if it goes idle for 10
minutes.

We'll be trying to make things more dynamic in upcoming releases
(hopefully by next summer) such that loncnew will fork off the child
for each host only when it needs to bring up the connection.

-- 
guy at albertelli.com  LON-CAPA Developer  0-7-3-9-