If there is backend work going on but the portal tier is still up, I've removed the login form from login.html and replaced it with a "system maintenance" type message, but kept the portal tier up to serve it
If the portal tier services are down as well, I've brought up something like tinyhttpd on port 80 with a service outage page.
I like your idea to use an alternate http server to display an outage notice when the portal web services are down.
I just created a twitter account and used a jquery plugin to display twitter updates to our login page, which has come in very handy. Putting that onto an outage page would be even better.
Could probably modify the luminis start/stop scripts to automatically bring up/shut down tinyhttpd so it's got things covered while Luminis is not available.
We are using a load balancer that is capable of "detecting" if the Portal tiers are up and running in our parallel deployment. (I assume that it is issuing a telnet statement, or maybe curl).
If all of our Portal tiers are down, then the load balancer has a match rule to automatically redirect to an alternate server where a nice message informs the user to try again later, and with some relevant links.
I believe that Apache can use mod_proxy to create clusters, and intelligently route in the same way as our load balancer.
I see a couple of comments here that a load balancer is being used to redirect the Luminis login page to another web page during a maintenance outage. Is anyone willing to talk to our load balancer analyst for some direction on setting this up. Any help is greatly appreciated.
Outage page
If there is backend work going on but the portal tier is still up, I've removed the login form from login.html and replaced it with a "system maintenance" type message, but kept the portal tier up to serve it
If the portal tier services are down as well, I've brought up something like tinyhttpd on port 80 with a service outage page.
alternate web server for outages
I like your idea to use an alternate http server to display an outage notice when the portal web services are down.
I just created a twitter account and used a jquery plugin to display twitter updates to our login page, which has come in very handy. Putting that onto an outage page would be even better.
Could probably modify the luminis start/stop scripts to automatically bring up/shut down tinyhttpd so it's got things covered while Luminis is not available.
Load balancer
We are using a load balancer that is capable of "detecting" if the Portal tiers are up and running in our parallel deployment. (I assume that it is issuing a telnet statement, or maybe curl).
If all of our Portal tiers are down, then the load balancer has a match rule to automatically redirect to an alternate server where a nice message informs the user to try again later, and with some relevant links.
I believe that Apache can use mod_proxy to create clusters, and intelligently route in the same way as our load balancer.
Load balancer too, same as
Load balancer too, same as above!
Jack
Load Balancer Help
I see a couple of comments here that a load balancer is being used to redirect the Luminis login page to another web page during a maintenance outage. Is anyone willing to talk to our load balancer analyst for some direction on setting this up. Any help is greatly appreciated.
Thanks, Alicia
Idaho State University
tausalic@isu.edu