Hello,
no problem, thanks for letting me know.

Best regards
 

Thank you for your follow-up. After further investigation we have determined that the issue was caused by a race condition in our own restart sequence, not by QuickOPC or TOP Server behavior.

When restarting a group task, our code was closing the shared channel writer before cancelling the CancellationToken. This left the OPC session in a partially torn-down state — subscriptions were still technically active while the channel was already closed. In that window, QuickOPC was sending the empty Publish request (subscriptionId=0, sequenceNumber=0) as a keep-alive on the broken session, which TOP Server correctly rejected with BadTimeout.

After correcting the shutdown sequence to cancel the CancellationToken first and then close the channel, the Warning 4051 no longer appears under any restart scenario — individual group restart, restart all, or service shutdown.

No changes to QuickOPC configuration were needed. Apologies for the noise — the issue was entirely on our side. Thank you for your assistance.

Hello.

Are you seeing a PublishRequest where the SubscriptionAcknowledgements array contains a single element where subscriptionId = 0 and sequenceNumber = 0? My understanding is that such request would be invalid - but I have not observed it so far. If so, do you perhaps have a WireShark capture?

Or, are you seeing PublishRequest-s where the subscriptionAcknowledgements array is simply empty?

Regards
 

TOP Server
Toolbox OPC Power Server 6.17
Version V6.17.240.0
Thanks for your help!

Hello,
thank you for reporting this. Please indicate the version of TOP Server you are using.

We will try to reproduce and investigate. although the behavior likely comes from the OPC Foundation UA SDK.

Best regards
 

Every ~2 minutes after session creation, the client sends a PublishRequest with subscription ID 0 and sequence number 0 
— an empty publish used as a session keep-alive. TOP Server rejects this with BadRequestTimeout (0x80850000) / BadTimeout (0x800A0000). 
Warning 4051 fires once per session then self-suppresses. Active subscriptions exist on these sessions — the keep-alive should not be necessary. 
Is there a parameter to disable the empty Publish keep-alive or configure a different keep-alive mechanism?