|
Hey guys!
Do you guys know how to configure nanny so that it doesn't output so much to
the error log? Right now, it yaps about bad load average strings from
ruptime (when the average load is 0.00). Plus, I am seeing it complain
about a timeout on port 2401 of a virtual server (cvs.mydomain.com). That's
a new occurence, and it seems as though the lvs is responding to the port
2401 tcp connection (because WinCVS complains about an 'object is remote'
error when I do the cvs login command).
I just need to reduce the output from nanny cuz the message log is about 18
megs per day!!! (uncompressed)
Also, any of you guys ever use lvs with the cvs pserver on port 2401? Right
now I only have a single real server behind the virtual cvs.mydomain.com ...
but it doesn't seem to work.
ipvsadm=0.9.14
Piranha= 0.4.14
kernel=2.2.16-4
when I run ipvsadm -L, I get:
IP Virtual Server version 0.9.14 (size=4096)
Prot LocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
TCP kryptonite.javatopia.com:www wlc
-> web1:www Masq 100 0 0
-> web2:www Masq 100 0 0
TCP kryptonite.javatopia.com:cvspserver wrr
TCP kryptonite.javatopia.com:https rr
so ... why don't my real servers show up for the cvspserver and https? Is
it because the nanny 'ping' on the servers:port is timing out?? Why would
that test connection fail? I can telnet to the IP of those servers
(10.75.1.30:2401 and 10.75.1.11:443) from the LVS machine.
Thanks.
-- Jake
javadude@xxxxxxxxx
|
