We had a PVS farm with 2 PVS servers (Server 2022)...
# citrix-provisionings
Srinivasan Shanmugam
11/13/2025, 9:57 PMWe had a PVS farm with 2 PVS servers (Server 2022) and 610 targets. We built new PVS servers running Server 2025, added them to the farm to replace existing servers. One of the new servers has rebalancing issues.
NewServer1 - 305 connections
NewServer2 - 305 connections
Stop stream service on NewServer1, connections doesnt fail over to NewServer2 and connection count stays 305.
Stop stream service on NewServer2, connections fail over to NewServer1 with new connection count 610.
Compared NewServer1 and NewServer2, there are no differences. Any pointers on what could cause this issue?
r
Rob Zylowski
11/13/2025, 10:15 PMIt depends on how you do the bootstrapping. This should help
https://support.citrix.com/support-home/kbsearch/article?articleNumber=CTX208519
I assume tis all efi so you should o over the efi section.
Rob Zylowski
11/13/2025, 10:27 PMThinking about it if the targets are already booted that's not the issue. There are licensing issues that stop has but its working in the one direction. Maybe someone from the pvs team will chime in soon
l
Lennart Hermansson
11/13/2025, 10:55 PMHave all the targets been rebooted since you added the new PVS servers?
If not reboot them and see if rebalancing works
s
Srinivasan Shanmugam
11/13/2025, 10:57 PM@Lennart Hermansson - Yes, we have weekly reboots and all the targets have been booted at least once since new servers were added.
s
Steven Brown
11/13/2025, 11:27 PMUpdate your BDM partitions or bootstrap config.
s
Srinivasan Shanmugam
11/13/2025, 11:29 PMBootstrap is updated. Like I said, no issues with targets booting up. It’s just rebalancing on one of the servers.
Srinivasan Shanmugam
11/14/2025, 4:23 AMAdded the PVS servers to Citrix Monitor (Infra Monitoring). There is a column "Total Rejected Logins" and the server having rebalance issues has 0 without a denominator. Not sure if it tried to login, I don't know how to interpret this information.
s
Steven Brown
11/14/2025, 5:29 AMYou could attempt to do some Wireshark traces on a vda to see if it attempted to reach out to the fail over server. Maybe a firewall issue?
Steven Brown
11/14/2025, 5:31 AMDo you have any vdisk properties that might account for this? Subnet affinity?
j
jvitech
11/14/2025, 12:06 PMNetwork on your PVS is correct including the mask? also the PVS not taking the failover works fine with other VMs. Weird. All VMs are in the same subnet?
Can you boot the VM not failing over from this PVS? Create an ISO having just this PVS IP to try. I would do that.
s
Srinivasan Shanmugam
11/19/2025, 2:26 AMIssue resolved - we had a typo in Option 11 for one of the PVS servers. That’s why rebalance was working one way.
We had to reboot all target devices to clear DHCP cache after updating scope option 11. Where does target cache this information? Just want to understand if it is possible to update without a reboot.
j
jvitech
11/19/2025, 9:51 AMThanks for sharing the solution! I use BDM iso so never came across this option 11!
DHCP option 11, also known as the Resource Location Protocol (RLP) Server option, is a DHCP option for specifying the IP addresses of one or more servers. It is primarily used with UEFI-based network booting, most notably with Citrix Provisioning Services (PVS), to provide a list of server addresses to the client during the boot process.
3 Views