> writes:
> >If your daemons are "going down" that often then I really think you
> >have some other issue you should be looking at.
> My daemons are not failing, however it is a possibility and some daemons
> over time have locked up or a rare occasion.
I run close to 80 virtual domains/hosts with a fair load (approx 3 to 10
connections at any time). My web server has not gone down once in
almost 60 days (Jan 25 will be 60 days). Every morning at 9:30am, I
send a "killall -HUP httpd" to the server. It pages me with "gonzo:
httpd 9:30 restart OK -- accepting connections on port 80". Meaning
gonzo (the server) restarted successfully. Every 10 minutes it checks
the server to make sure it's up. If it does not get a proper responce
back from the server, it pages me "gozno: httpd <time> DOWN --
attempting restart". If it restarts the server successfully, it pages
me "gonzo: httpd <time> up -- accepting connections on port 80". If it
fails to restart, it pages me "gonzo: httpd <time> ALERT: httpd restart
FAILED! server needs remote attention!" (in cases where the whole
server is down and telnet/ssh access isn't available, it pages me
"gonzo: SERVER <time> ALERT: server needs personal attention! (message
from bosco)" This message is coming from my remote logger on bosco
(another server) telling me that gonzo needs personal attention.
Anyway, the point is I have a nice logging program that tells me stats
about the server. If you don't have an alpha-numeric pager, you can
still make nice little scripts that can watch/monitor/maintain the
server.
Quote:> >Unless you have specific reasons (eg. running them as different users,
> >restricting traffic to one server, etc.) it is a huge waste to run
> >a seperate copy for each virtualhost, especially if some of your vhosts
> >are like many with little traffic.
> Many virtual hosts do have very little traffic. It is my impression that the
> main apache daemon will only start enough child process that are needed to
> handle the load. Thus saving considerable on memory.
Very seldom does my server ever crash. As I said, it's almost been 60
days, and last time I was doing a kernel upgrade.
Thanks,
T.J. Weber
--
T.J. Weber | huked 0n f0niks
President, Interplanetary Media | werked four mee!
phone: 847.205.5200 |----------------------------
fax: 847.205.5201 | part time president, student,
web: http://www.ipmedia.net |
Side note to that "special someone" in my life:
-----------------------------------------------------------
| "You are the epitome of everything I've ever looked for |
| in another human being." --Chasing Amy, 1997 |
-----------------------------------------------------------