01-09-2013 02:56 PM - edited 03-18-2019 12:24 AM
We have two TMS servers configured to a load balancer, and for some reason the TMSPE connection between TMS and VCS isn't working.
According to the VCS and TMS it says "IO Error" and "() timed out".
We've setup the LB the way the TMS redundancy guide stated, and I can ping the TMS server from the VCS. Any idea of what could be the cause?
01-09-2013 09:33 PM
Patrick
The IO error might be related to that you are using TLS or that the vcs isnt able to reach the tms. Are you using the nlb address for tmspe? What happens if you bypass the nlb by typing one of the tms ip directly?
/Magnus
Sent from Cisco Technical Support iPhone App
01-09-2013 09:45 PM
Hello Mangus - A little bit more background to the issue, as well as the results of what you had asked.
We first has TMS set to it's own default gateway and had the load balancer manage the routing in and out of the servers (best way I can describe it), and the TMSPE connecection worked just fine. However over time we found our codecs were being set to "Behind Firewall" automatically by TMS, we assumed the HTTP headers were being altered by the NLB as we've see that happen with other website when pulling reports, so we changed the default gateway of the TMS servers to be the NLB as the redundancy documentation recommended. In turn, we fixed the behind firewall issue, but got this new issue as a result.
TLS was set on the VCS provisioning page, I even tried it without, and it still gives the IOError when using either the NLB IP or URL of the TMS frontend.
When I changed the address on the provisioning page to go directly to one of the TMS servers, and disabled TLS because it was throwing an error, I'm getting all green checks and everything is connecting between the TMS/VCS.
-------------
On a side note, our MCU is also throwing an error that it can't connect to TMS too, after we made the TMS/NLB change today. I found that out tonight when I was looking at something else in it's logs. It says it can't connect to TMS via port 80?
01-10-2013 03:53 PM
Opened up a case with TAC, apparently it's related to the load balancer and needing TMS to be on its desperate vlan. Something you wouldn't think to be an issue if all the managed servers were on the same vlan in the same data center. So we're going back and forth. Something that was missed when setting up the redundancy.
Sent from Cisco Technical Support iPhone App
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide