|
We have 2 tomcat servers behind an lvs+ldirectord load balancer. The
pertinent section of ldirectord.cf looks like this:
# Virtual Server for tomcat(site008), Outside to Inside
virtual=192.168.5.100:3008
real=192.168.10.61:3008 masq
real=192.168.10.62:3008 masq
service=http
request="/checkup.html"
receive="site008_tomcat_is_up"
scheduler=lblc
protocol=tcp
checktype=3
persistent=360
Normally, ipvsadm looks something like this:
[root@lb1 ]# ipvsadm|grep 3008
TCP extrovert.mydomain.com:3008 lblc persistent 360
-> appftp2.mydomain.com:3008 Masq 1 1 4
-> appftp1.mydomain.com:3008 Masq 1 2 6
We turned off one of the servers and now it looks like this, which is
expected:
[root@lb1 ]# ipvsadm|grep 3008
TCP extrovert.mydomain.com:3008 lblc persistent 360
-> appftp2.mydomain.com:3008 Masq 0 0 3
-> appftp1.mydomain.com:3008 Masq 1 2 6
However, users who were connected to appftp2 got "page could not be
displayed" errors, which tells me that LVS is was still directing
traffic to the down server.
After a while, the situation cleared up. I assume this is because
ipvsadm started directing traffic to the up server.
Does anyone know why this would be happening? Is my persistent=360
statement causing ipvsadm to ignore the down state of the realserver?
--
Eric Robinson
Director of Information Technology
Physician Select Management, LLC
775.720.2082
Disclaimer - May 15, 2007
This email and any files transmitted with it are confidential and intended
solely for LinuxVirtualServer.org users mailing list.. If you are not the named
addressee you should not disseminate, distribute, copy or alter this email. Any
views or opinions presented in this email are solely those of the author and
might not represent those of Physician Select Management (PSM) or Physician's
Managed Care (PMC). Warning: Although the message sender has taken reasonable
precautions to ensure no viruses are present in this email, neither PSM nor PMC
can accept responsibility for any loss or damage arising from the use of this
email or attachments.
|
