Mailing List Archive: mixed persistent and regular ipvsadm entries

mixed persistent and regular ipvsadm entries

Jan 31, 2001, 10:59 AM

Post #1 of 4 (717 views)

If I setup a 2 real-server VS-DR LVS with persistence,

ipvsadm -A -t $VIP -p -s -rr
ipvsadm -a -t $VIP -R $realserver1 $VS_DR -w 1
ipvsadm -a -t $VIP -R $realserver2 $VS_DR -w 1

as expected I can connect to any service on the
real-servers, always getting the same real-server.

If I now add an entry for telnet to both real-servers,

ipvsadm -A -t $VIP:telnet -s -rr
ipvsadm -a -t $VIP:telnet -R $realserver1 $VS_DR -w 1
ipvsadm -a -t $VIP:telnet -R $realserver2 $VS_DR -w 1

I can telnet to both real-servers in turn as would be
expected for an LVS serving only telnet, but all other
services go to the same first real-server (ie all
but telnet appear persistent).

I take it that the director is persisting all ports
except those that are explicitely set as non-persistent.

I expected that one of these sets of ipvsadm commands would
overwrite the other. It seems that persistent and non-persistent
connections can be made at the same time.

Is this by design and going to stay this way?

Joe

--
Joseph Mack PhD, Senior Systems Engineer, Lockheed Martin
contractor to the National Environmental Supercomputer Center,
mailto:mack.joseph@epa.gov ph# 919-541-0007, RTP, NC, USA

Re: mixed persistent and regular ipvsadm entries [ In reply to ]

mack at ncifcrf

Jan 31, 2001, 5:02 PM

Post #2 of 4 (712 views)

Permalink

On Thu, 1 Feb 2001, Julian Anastasov wrote:

> > I expected that one of these sets of ipvsadm commands would
> > overwrite the other. It seems that persistent and non-persistent
> > connections can be made at the same time.
> >
> > Is this by design and going to stay this way?
>
> Yes, the templates we use to keep the persistence are not
> considered when scheduling packets for non-persistent connections.

I'm impressed. I expected the two lots of ipvsadm commands to tread on
each other's toes.

> Not sure what happens when VIP:0 and VIP:21 are defined persistent
> together. Didn't checked such setup :)

I'll see if I can test it
Joe

--
Joseph Mack mack@ncifcrf.gov

Re: mixed persistent and regular ipvsadm entries [ In reply to ]

ja at ssi

Jan 31, 2001, 6:37 PM

Post #3 of 4 (716 views)

Permalink

Hello,

On Wed, 31 Jan 2001, Joseph Mack wrote:

> If I setup a 2 real-server VS-DR LVS with persistence,
>
> ipvsadm -A -t $VIP -p -s -rr
> ipvsadm -a -t $VIP -R $realserver1 $VS_DR -w 1
> ipvsadm -a -t $VIP -R $realserver2 $VS_DR -w 1
>
> as expected I can connect to any service on the
> real-servers, always getting the same real-server.
>
> If I now add an entry for telnet to both real-servers,
>
> ipvsadm -A -t $VIP:telnet -s -rr
> ipvsadm -a -t $VIP:telnet -R $realserver1 $VS_DR -w 1
> ipvsadm -a -t $VIP:telnet -R $realserver2 $VS_DR -w 1
>
> I can telnet to both real-servers in turn as would be
> expected for an LVS serving only telnet, but all other
> services go to the same first real-server (ie all
> but telnet appear persistent).
>
> I take it that the director is persisting all ports
> except those that are explicitely set as non-persistent.
>
> I expected that one of these sets of ipvsadm commands would
> overwrite the other. It seems that persistent and non-persistent
> connections can be made at the same time.
>
> Is this by design and going to stay this way?

Yes, the templates we use to keep the persistence are not
considered when scheduling packets for non-persistent connections.
Not sure what happens when VIP:0 and VIP:21 are defined persistent
together. Didn't checked such setup :)

> Joe

Regards

--
Julian Anastasov <ja@ssi.bg>

Re: mixed persistent and regular ipvsadm entries [ In reply to ]

mack.joseph at epa

Feb 1, 2001, 8:39 AM

Post #4 of 4 (713 views)

Permalink

Julian Anastasov wrote:

> Not sure what happens when VIP:0 and VIP:21 are defined persistent
> together. Didn't checked such setup :)

hmm. When you have an LVS with VIP:0, all ports are linked together.
I don't know what you mean by saying "VIP:21 persistent".

What ipvsadm commands would test what you are interested in here?

Joe

--
Joseph Mack PhD, Senior Systems Engineer, Lockheed Martin
contractor to the National Environmental Supercomputer Center,
mailto:mack.joseph@epa.gov ph# 919-541-0007, RTP, NC, USA