We use Netscaler where I work and we didn't require anything special in our config other than what bmc says.
Thanks for the reply. Got a few ?? if you don't mind:
Load Balancing Methods: Have you found that Remedy or the user's experience is better using least connections, round robin, or something else?
Sessions persistence: Again do you see performance improvement when using cookie, server IP, source IP, or something else for this?
Monitoring/Health Checking: Any good suggestions in this area?
One of the things I read in the docs is that post 7.6.04 the sticky bit is no longer needed on your load balancer as long as all your Remedy apps are above 7.6.04. It also in mentions the need to configure your MidTier to use the Enable Lifespan option in the MT Configuration Tool. Any thoughts or comments about sticky bit usage/Lifespan option?
Thanks for taking the time LJ,
I'm not the net scaler administrator....but...
I have always asked for 'least connections, because the other options don't seem to make us much sense in the long run .
I believe we use Cookie for our Session persistence.
Monitoring/Health Checking....now that's an entirely different discussion. I believe most people would go with a port check...80/443 for their Mid-Tier and whatever port you choose for your application server...I have had problems with this approach however and have written custom applications that do what I consider 'smart checking'....for example, on the app server, I have a program that logs on, performs a query, issues a few key commands to ensure plugin servers are functional, etc...when all of that passes, I open up a port that the LB checks...if everything's not kosher, I don't open the port...which stops traffic from flowing to the server...similar efforts on the web side, ensuring that you actually REACH the home page when doing a health check, and only if you do, then consider it 'up'...takes a fair amount of work, but I think it's worth it if your app is considered as critical as some Remedy systems are in their respective infrastructures.
Yes...Sticky (Session Persistence) is not needed on the App LB anymore...and I don't recall seeing any guidelines regarding setting the lifespan (but it's been a few years, so I'm not entirely sure)...I know we configured it, but don't recall where we got the values we used...
I realise this post is very old now, but was wondering (hoping) you may still have this info to hand.
I am working with a client who is using a Netscaler LB on Remedy 8.1.02 and the users are regularly facing the HTTP Communication Error 504 (ARERR 9351).
The clients LB team don't know where to look to confirm their settings, and I can't find a straight forward doc / guide to point them to.
Do you have the details of the settings they need to implement?
3 of 3 people found this helpful
Keith, I have a tool that myself and a friend put together (Load Balancer Monitor – A Programming Legacy ) that allows you to do smart monitoring of either an arserver or a mid-tier and report it's availability via a port being opened, thus simplifying the LB config so all you need to do is do a TCP Monitor to determine if the service is available.