It's good practice for providing a comfort page so the user would know the site is down versus hitting dry air. The site being down could be from a host of issues
- planned maintenance
- un-planned maintenance
- critical errors
- code-application errors
Concept #1 is a hosted HTML PAGE via the F5 LTM
Concept#2 is to host a pool of ever-green servers that house the SITE_DOWN pages. Bigger application hosting outfits typical this method
Concept#3 is to host a server node that houses the site down page
A hosted a data iFile and that's call out as required
Out of the above the pool and or node is the better method , since the administrators of the web server that hosts the site down can modify the content as required, If you set a NULL_POOL, you can always call up the "dummy" site and check the page. With a off-F5-LTM hosted site_down page, the F5 engineer would not be tied up with managing Ifile or HTML respond content.
With the NULL_POOL defined, you can point your browser at the http://ww11.example.com will call up the hosted our site_down for testing
And lastly, you could a simple site http redirect or fallback statement, but be advise of mistakes with bad 302 redirects or even worst redirect loops.
Finally the title of this post was specific with examples for using Irules. We have one more method that's even simpler.
By using a priority-group and with a node(s) set as the SITE_DOWN, you can apply that node to your pools to host the site DOWN when all other nodes are down in that pool.
See this visio diagram where we have two priority groups and the last group with the server node "SSD" while be hit for the SiteDown page when !!!!! ALL OTHER NODES HAVE FAILED !!!!
The node SSD could be across all pools for simple execution and could be a micro-image low linux tinyhttpd server.
NSE ( network security expert) and Route/Switching Engineer
kfelix -----a----t---- socpuppets ---dot---com
=( @ @ )=