LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

Re: Will failover lose requests?

To: "ratz" <ratz@xxxxxx>, "huatao" <thua@xxxxxxxxxxxxxxx>
Subject: Re: Will failover lose requests?
Cc: <lvs-users@xxxxxxxxxxxxxxxxxxxxxx>
From: "Putchong Uthayopas" <pu@xxxxxxxx>
Date: Fri, 17 Nov 2000 17:00:25 +0700
HI,

Are there any way to backup and restore the state of persistance connection
from LVS kernel.  (via. device interface, api etc) We can send it to backup
node and restore when primary fail. This at least will help maintain some
live connection.

Putchong.

----- Original Message -----
From: "ratz" <ratz@xxxxxx>
To: "huatao" <thua@xxxxxxxxxxxxxxx>
Cc: <lvs-users@xxxxxxxxxxxxxxxxxxxxxx>
Sent: Thursday, November 16, 2000 10:43 PM
Subject: Re: Will failover lose requests?


> Hi Huatao,
>
> huatao wrote:
> >
> > Hello lvs-users,
> >   It's my first time to write to this list, nice to meet you all!
> >   I've got some questions:
> >   One is when failover takes place, will the requests during the
failover
> >   course lose?
>
> Yes, unfortunately. And this is very nasty for persistent setups
> in an e-commerce environment. Take for example a simple e-com site
> providing some subjects to buy. You can browse and view all their
> goodies. At a certain point you want to buy something. Ok, it is
> common nowadays that people can buy over the Internet with CC. Obviously
> this is done f.e. with SSL. SSL needs persistency enabled in the
> lvs-configuration. Imaging having 1000 users (conn ESTABLISHED)
> that are entering their VISA information when the database server
> crashes and the healthcheck takes out the server; or even more simple
> when the server/service itself crashes. Ok, all already established
> connections (they have a persistent template in the kernel space)
> are lost and these 1000 users have to reauthenticate. How does this
> look from a clients point of view which has no idea about the technology
> behind a certain site.
> To summarize: Yes, you loose the information in the kernel space about
> the templates. Good thing: The design for a possible synchronisation
> of the masqerading table from one director to another one is done the
> implementation not.
>
> >   The other is how to benchmark failover, that is how long the services
> >   provided by LVS will be unavailable during the failover course.
>
> This is up to different things: Your failover software used, the switch
> setup and learning delay (extremely stupid feature of cisco switches) and
> your setup.
>
> HTH and best regards,
> Roberto Nibali, ratz
>
> --
> mailto: `echo NrOatSz@xxxxxxxxx | sed 's/[NOSPAM]//g'`
>



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