And here I am thinking that my 15-17ms pings are good…
Like I said, I’m not saying that the pings are now crap, as they ain’t, but I am just wondering why they have doubled when for the last 7 months they were at 3-5. Just wondering if @Yayzi_Team have any ideas or done something that has changed this.
Single digit is great - totally! Especially when coming from VM…
Cheers.
Keeop
@Keeop you posted something about only getting between 1.6-1.7Gbps on synthetic speedtests. I’m seeing the same thing when using Opnsense on Proxmox, but if I switch to the Yayzi supplied I’m hitting over 2Gbps.
Opnsense is running on an i7-13700 so horsepower isn’t the issue here. Been through all the tuning options for Opnsense ( Receive-side scaling etc) on the web but makes no odds. Ntop shows the very little resource utilisation while running speedtests
Running Intel I226-V nics
Just wondered if you see the same thing as I vaguely remember you running Opensense also. Not a big issue but I did get full throughput on Opnsense at one point.
Hi @Lee
I have actually parked my Opnsense gateway for now as I am getting on OK with the UDM-SE. I still may well flip back though as the functionality on the UDM is nowhere near as good as Opnsense but I do like the ‘single pane of glass’ approach and the nice GUI!!
But, that was running on a VM with 6 cores of an E5-2687 allocated so again, no hardware limitations there.
I can’t remember what the specs of the UDM-SE are off the top of my head, but the CPU will be Noddy compared to your i7. But, just run a test directly from that and I got 1.9 down. It can be very variable though. For instance, I have it set to run a test very morning and the one it ran this morning only returned about 400Mb! In fact…
Those are the last few - bit of variance to say the least!!
Are you running your tests directly from the Opnsense box, or from a different client?
I’m unfamiliar with those NICs but I have played with both Intel X550 and X710 and while both should be 10Gb, I get lower throughput on X550 so I guess that may be a factor?
However, if you used to get full throughput, then I guess it must be Opnsense itself - possibly an update?
Sorry, can’t help really. If I switch back to Opnsense, I’ll be sure to run some tests and let you know.
Yeah could well be an Opnsense update thats done it. I’ll tinker round some more.
7ms?
Highest i’ve seen is 3 (Unifi saying 6).
Though i’m constantly getting errors about packet loss, something on the network is clearly not very stable.
Must depends on where you are. Some joining it has been 6-7ms, which beats the absolute snot out of VM HFC at 18-20ms, which in itself is good
Ironically VM were out today in my area making the FTTP live, so once i get that installed i’ll feedback on what its like.
Strange enough mine has always been 14-16ms but today its dropped to 9ms. So Yayzi DONT touch nothing as its the best I have had
2G down is crazy
Yeah, so my pings were like that but have suddenly doubled. On the plus side, I’m not experiencing the same packet loss as you.
I’ve noticed my average pings and spikes have started to increase again during peak hours. I’m wondering if we’re getting to that time again where new subscriber influx are starting to impact the port capacity again.
I’m not getting any packet loss yet, but from memory this is how it started last time.
I could be wrong, hopefully I am
@Yayzi_Team @Yayzi_Staff Any input / thoughts on the packet loss and high ping?
No issue our end guys. The next port upgrade has been done ahead of time. Absolutely loads of bandwidth available
Any reason for the packet loss and higher than normal pings then?
No, the part capacity was fixed
It sounds like it may not be a Yayzi problem at this time, But I’m sure they will investigate
Either one do not have packet loss and have had zero since joining, No, that was unplanned anyway
If it was something on our end I believe it would affect everyone like the port issue did.
But we are investigating
Could be the GPON stick
You’ve hit the nail on the head! I bet it’s the GPON stick interacting with the giggle sprocket back feeding RF rays down the hyper-line which is having an adverse affect on the Yayzi flux capacitors.