🎄 Happy Holidays! 🥳

Most of Solace is closed December 24–January 1 so our employees can spend time with their families. We will re-open Thursday, January 2, 2024. Please expect slower response times during this period and open a support ticket for anything needing immediate assistance.

Happy Holidays!

Please note: most of Solace is closed December 25–January 2, and will re-open Tuesday, January 3, 2023.

DMR cluster started multiplying messages across the mesh instead of smart subscriptions handling

Aleksei
Aleksei Member Posts: 14

I've got a setup of one backend HA group and three nodes for MQTT clients. All connected into DMR mesh. One backend node (primary) receives about 500-1k msgs/sec from a single publisher. No subscribers on the mesh at any node. Messages are trafficked across the mesh bridge links unconditionally and it seems that they even return back to the published node. Message rate counters on the publisher node even hit 10k msgs/sec ingress limit of the standard edition. But only 500 or so per second actually enter the mesh. I saw this kind of behaviour when there are topology issues, but none are reported. Logs appear to be silent of any issues. Web-facing nodes do have an error that link is down and the channel is missing. I've checked the channels on those nodes, and it appears that message VPN channels are indeed not being created like they are on backend nodes of the mesh. All nodes have default(disabled, dynamic off) and my message VPN (on all nodes enabled, dynamic routing ON). I believe I hit some kind of a bug, cause configuration was diffed on all nodes and doesn't differ in any important way. Any thoughts or experience shared are appreciated!

Tagged:

Answers

  • Aaron
    Aaron Member, Administrator, Moderator, Employee Posts: 644 admin

    Bump this old thread. I don't remember ever seeing this? Did you eventually fix or did Solace fix or did this get resolved??

  • Aleksei
    Aleksei Member Posts: 14

    @Aaron To be honest I don't remember if I found the problem. I do remember it was working fine after rebuilding the mesh from scratch allowing time for nodes to sync properly before extending mesh. Maybe it was config issue of some sort cause setting up all those links can get error prone.