|
On Fri, Sep 24, 1999 at 07:34:46PM +0800, Wensong Zhang wrote:
> Virtual Server patch for Linux 2.2 - Version 0.9.0 - September 24, 1999
[...]
> - Added new persistent service handling
[...]
> For all the
> persistent services except FTP, we create a masq template like
> <daddr, 0, maddr, mport, saddr, sport>. So, the persistent services
> won't disturb each other, and it fixes the wrong accounting bug for
> different persistent services.
[...]
> For passive
> FTP, FTP server tells the clients the port that it passively listens
> to, and the client issues the data connection. In the tunneling or
> direct routing mode, the load balancer is on the client-to-server half
> of connection, the port number is unknown to the load balancer. So, a
> template masq like <daddr, 0, maddr, 0, saddr, 0> is created for
> persistent FTP service.
Now, if I read this right, it says that I could persistently redirect
different services on one virtual IP to different servers -- as long as
no passive FTP is involved.
When passive FTP is involved on a virtual IP, however, ALL services on
that IP are persistently linked to one server (based, of course, on
source addr.) For example, I wouldn't be able to reroute :80 to a
set of web-only servers, and reroute :20, :21, :<passive-ports> to FTP-
only servers. Right?
If this is indeed the case, would it be possible to circumvent this
nature by providing some kind of "fallback" LVS entry? Something like:
# reroute HTTP
ipvsadm -A -t 10.0.0.1:80 -s rr
ipvsadm -a -t 10.0.0.1:80 -r 192.168.0.66 -g
ipvsadm -a -t 10.0.0.1:80 -r 192.168.0.130 -g
# reroute HTTPS
ipvsadm -A -t 10.0.0.1:443 -s rr -p
ipvsadm -a -t 10.0.0.1:443 -r 192.168.0.66 -g
ipvsadm -a -t 10.0.0.1:443 -r 192.168.0.130 -g
# fallback (FTP, passive FTP, and everything else)
ipvsadm -A -t 10.0.0.1:0 -s rr -p
ipvsadm -a -t 10.0.0.1:0 -r 192.168.0.67 -g
ipvsadm -a -t 10.0.0.1:0 -r 192.168.0.131 -g
So that if a connection comes around that's specificially defined (like
an HTTP or HTTPS connect), it works as it should. However, if a specific
entry for that IP's port is NOT found, the fallback :0 entry would apply.
Is this how things currently work? And if not, would it be possible or
even desirable to do it this way?
--
Jake Buchholz, ExecPC Sr. Systems Administrator : /~\
jake@xxxxxxxxxx -- http://www.execpc.com/~jake : ASCII \ / Against
Fngrpt PGP262: 29f3322af2da07994dc03fdf1f50aed2 : Ribbon X HTML
GnuPG: 53d97ab3d279f731e7d2fc1197ca38544423882b : Campaign / \ Mail
----------------------------------------------------------------------
LinuxVirtualServer.org mailing list - lvs-users@xxxxxxxxxxxxxxxxxxxxxx
To unsubscribe, e-mail: lvs-users-unsubscribe@xxxxxxxxxxxxxxxxxxxxxx
For additional commands, e-mail: lvs-users-help@xxxxxxxxxxxxxxxxxxxxxx
|
