Presenting an unavailable webpage when pool is down (Thick Client)

If you had a second monitor watching tcp/808, then the loss of that port would also take the pool member down.

If there are no pool members available, then the virtual server will go down. However, if there's an irule attached to that server, which references another pool, the virtual server will remain up, even with no pool members. The thinking here is that the irule might send the traffic to another pool member that is still up, so we need to keep the virtual up, so that the irule gets a chance to run.

So with that in mind, you could create an irule that looks at the number of members active, and redirects to another page or pool member if that is case. Probably something similar to this

Of course this all depends on whether your application is smart enough to handle such a response in an intelligent way when you send it unexpected data.

Hi, You can add another node to your pool. This node may have this requirement : - This node refer to maintenance area (this specific check) - Have less priority than then other one (applicative nodes)

Then all applicative nodes are down (by monitor test), maintenance node take the lead, and you application can show maintenance page. You can have multiple maitenance node (with the same priority value).

I hope it can help you

I think wnat you may be after is the LB_FAILED event

when LB_FAILED {
         Either all pool members are down OR the pool member selected has not responded
        HTTP::respond 503 content "Sorry your page is down" noserver Retry-After 300
        return
}

Alternatively if you happen to have the sorry page locally stored in an iFile;-

when LB_FAILED {
         Either all pool members are down OR the pool member selected has not responded
        HTTP::respond 503 -version 1.1 content [ifile get "sorry_page_html"] noserver Content-Type "text/html;charset=utf-8" Server $server Retry-After 300 noserver
            return

}