LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

Re: LVS vs Piranha

To: "David D.W. Downey" <david.downey@xxxxxxxxxxxxxxxxxxxxxxxx>
Subject: Re: LVS vs Piranha
Cc: Michael Loftis <zop12@xxxxxxxxxxxx>, Q@xxxxxxxxxx, kel@xxxxxxxx, lvs-users@xxxxxxxxxxxxxxxxxxxxxx
From: Keith Barrett <kbarrett@xxxxxxxxxx>
Date: Thu, 14 Sep 2000 17:03:31 -0400
"David D.W. Downey" wrote:
> 
> 
> Kieth, that is bull crap.

As I have said, I am not going to participate in this type of
discussion.


> How does one pass the -v switch to
> nanny to get more verbosity so we can see where things are failing. 

Two ways:

1. Run nanny directly with the -v switch, specifying the same command line
   you saw when it was started by pulse.

2. Run pulse with "-n -v" commands. The latest RPMs will pass this along
   to nanny.


> Is nanny being called hardcoded from inside pulse or some other spot or a
> script? 

Hardcoded fork. You have the source code. The routine name would be
"activateSOMETING", where "SOMETHING" depends on whether you are using
FOS or LVS.


> How do we pass the verbose switch to lvs itself so we can seee
> what IT'S doing?

1. Run lvs manually, specifying a command line that include "-v"

2. Use the current RPMs and start pulse with a "-n -v".


> We can see that the lvs daeemon is parsing the file to
> some degree since the NAT IP and the VIP are coming online. It's dropping
> the ball on the virtual declaration(s). How do we follow along with it to
> find out exactly WHERE it's dropping the ball?

The best way is what you appear to be thinking, enable debug mode and
examine the displayed and logged output. It will display all the times
ifconfig is invoked to define or remove a VIP address.


Hope that helps. Have a nice day.


-- 

Keith Barrett
Red Hat Inc. HA Team
kbarrett@xxxxxxxxxx


<Prev in Thread] Current Thread [Next in Thread>