You can use my former post (quoted below) publically. We're also going to
use wackamole on a pair of clones that will be running our public webserver
(apache) and external smtp relays (sendmail w/procmail filters). The
"master" server will by rsync'd with the other in the event that it fails or
we need to take it down for maintenance. We'll only run one public IP and
the other server will sit there idle until the first server goes down.
Nothing elaborate, just cheap redudancy for two highly visible services.
What I'd love to see is the ability for Wackamole to monitor service ports.
The other server might still be online, but sendmail or apache might have
crashed, so I'd love to be able to take over due to a service failure. We
monitor everything, so essentually we'll be doing this manually if a service
fails by downing wackamole, but it'd be nice to automate it.
>From: "Jason Roysdon" <jroysdon@hotmail.com>
>Reply-To: wackamole-users@lists.backhand.org
>To: wackamole-users@lists.backhand.org
>Subject: [Wackamole-users] Solaris 8 test environment / What are you using
>Wackamole for?
>Date: Tue, 28 May 2002 15:47:23 -0700
>
>Our proposed environment is using round-robin DNS pointing at the VIPs of
>each Apache server running the Weblogic Load-balancing/clustering Plug-in,
>which in turn keep track of all BEA servers, which in turn pass all state
>information and communicate to both Oracle DB servers which in turn are
>constantly talking to each other (I'm not a DBA or even pretend to know
>what is going on with Weblogic, but I understand networking). My ASCII-art
>only shows a pair of each server, but you could scale any portion of this
>to include more servers depending on where the load and bottlenecks are.
>
>-------------- --------------
>| Apache w/ | | Apache w/ |
>| Weblogic |---| Weblogic |
>| Plug-in #1 | | Plug-in #2 |
>-------------\ /-------------
> | X |
>--------------/ \--------------
>| Weblogic | | Weblogic |
>| BEA #1 | | BEA #2 |
>--------------\ /--------------
> | X |
>--------------/ \--------------
>| Oracle DB |---| Oracle DB |
>-------------- --------------
>
>Without Wackamole, we'd need something like a pair of Cisco Local Directors
>out front, but those are $10K each plus a $200 failover cable (you'd think
>they'd just give the failover cable to you for spending $20K).
Jason Roysdon, CCDP/CCNP:Security, MCSE, CNA, Network+, A+
Public email : jroysdon@hotmail.com
Homepage : http://jason.roysdon.net/
PS: Don't use a subject like that looks like spam, as your message will end
up in my junk filter and never be read.
_________________________________________________________________
MSN Photos is the easiest way to share and print your photos:
http://photos.msn.com/support/worldwide.aspx
use wackamole on a pair of clones that will be running our public webserver
(apache) and external smtp relays (sendmail w/procmail filters). The
"master" server will by rsync'd with the other in the event that it fails or
we need to take it down for maintenance. We'll only run one public IP and
the other server will sit there idle until the first server goes down.
Nothing elaborate, just cheap redudancy for two highly visible services.
What I'd love to see is the ability for Wackamole to monitor service ports.
The other server might still be online, but sendmail or apache might have
crashed, so I'd love to be able to take over due to a service failure. We
monitor everything, so essentually we'll be doing this manually if a service
fails by downing wackamole, but it'd be nice to automate it.
>From: "Jason Roysdon" <jroysdon@hotmail.com>
>Reply-To: wackamole-users@lists.backhand.org
>To: wackamole-users@lists.backhand.org
>Subject: [Wackamole-users] Solaris 8 test environment / What are you using
>Wackamole for?
>Date: Tue, 28 May 2002 15:47:23 -0700
>
>Our proposed environment is using round-robin DNS pointing at the VIPs of
>each Apache server running the Weblogic Load-balancing/clustering Plug-in,
>which in turn keep track of all BEA servers, which in turn pass all state
>information and communicate to both Oracle DB servers which in turn are
>constantly talking to each other (I'm not a DBA or even pretend to know
>what is going on with Weblogic, but I understand networking). My ASCII-art
>only shows a pair of each server, but you could scale any portion of this
>to include more servers depending on where the load and bottlenecks are.
>
>-------------- --------------
>| Apache w/ | | Apache w/ |
>| Weblogic |---| Weblogic |
>| Plug-in #1 | | Plug-in #2 |
>-------------\ /-------------
> | X |
>--------------/ \--------------
>| Weblogic | | Weblogic |
>| BEA #1 | | BEA #2 |
>--------------\ /--------------
> | X |
>--------------/ \--------------
>| Oracle DB |---| Oracle DB |
>-------------- --------------
>
>Without Wackamole, we'd need something like a pair of Cisco Local Directors
>out front, but those are $10K each plus a $200 failover cable (you'd think
>they'd just give the failover cable to you for spending $20K).
Jason Roysdon, CCDP/CCNP:Security, MCSE, CNA, Network+, A+
Public email : jroysdon@hotmail.com
Homepage : http://jason.roysdon.net/
PS: Don't use a subject like that looks like spam, as your message will end
up in my junk filter and never be read.
_________________________________________________________________
MSN Photos is the easiest way to share and print your photos:
http://photos.msn.com/support/worldwide.aspx