Forum Discussion
Achisu
Nimbostratus
NimbostratusWhen user goes through LB the server page has stripped information
I have created a pretty simple round robin load balancing for a user with three servers. As a part of this I also have DNS LB in place that sends the traffic to two VIPs that are connected to the three nodes in a pool I have created on my LTM F5. User accesses the LB DNS URL I provide via Https://<>.com > VIP > Pool > Nodes. There is a certificate applied to the clientssl and serverssl profiles attached to the VIPs.
The user is able to get to their backend servers/nodes when going through the load balancer, but we are coming across an interesting issue. When the user goes through the F5 the server dashboard page they usually see is stripped of information on that dashboard. Typically, there would be tiles shown on the server dashboard, but it is just the basic UI and none of the tiles.
When the user goes directly to their server, all the information/tiles are shown as normal. I have never experienced this problem before and am not sure how to prove out the F5 is causing the issue or how it is happening.
Any insight would be greatly appreciated!
*Attached file shows what I'm explaining.
- zamroni777
Nacreous
to see the failed http object requests, you can open network tab in browser's developer tools (press f12)
probably the webpage includes objects referenced using fixed paths, which doesnt go through the f5 vserver, instead of relative paths.
i suggest involve the web developers in the troubleshooting. - Injeyan_Kostas
Cirrus
Is there any kind of persistence in the pool?
Have you tried with a single backend server only in the pool?
Are any other profiles applied in the VS?
AchisuThere is no persistence applied to the pool. There is just the client and server ssl profiles applied with the default configuration other than the certificates applied to them.
I will try making the pool be only available with one member and see if we can narrow down the issue there. Thanks for the tips.
- Mayur_Sutare
MVP
I had faced a similar issue in the past. At that time, developer tool helped to fix it. In my case, there were multiple requests going at backend to different destinations/urls to load all the required tabs. Few of the requests going from F5 to those destinations were not through which was causing the issue. You can review it as suggested in earlier posts.
The 2nd suggestion also looks good to test. At times, with single server everything works but creates a problem when multiple servers are available. In some cases, persistence is required to handle request appropriately when multiple pool members are active.
Looking forward to your response to understand root cause in your case.
