It implies that I spend a significant amount of time travelling on the London Underground.
Incidentally, that may be true for some routers, but not all. I have an RT-AC86U, which has a fairly powerful (1800MHz) dual-core CPU. One of the changes Merlin made to the standard Asuswrt software is that the OpenVPN client runs exclusively on the second CPU core, while all of the actual routing is handled by the first core; consequently, even if the VPN is active, unless there is significant data throughout from another device, it makes almost no difference.
sry, kernel 120 may not be in the repo. I was trying to avoid possible incompatibility between kernel and other packages. 117 is the last release kernel, so try
sudo apt-get install --reinstall vero364-image-3.14.29-117-osmc
Hi,
i have tested the new 121 kernel and the results are worse than before. Now even through the router the speed drops to <20 Mbits with >500 retrys. Same when i connect the vero4k+ to the switch.
After reverting the kernel to 120 and powercycling the vero4k+ i get the old bad numbers through the switch but through the router it works with ~930 Mbits in both directions again.
BTW How do i address the powersuply problem (5V 1A). My new (“faulty”) switch is just two weeks old. What powersupply do you recommend?
Hi Sam, I ran the updates like you described, updated fine.
Sadly it didn’t fix anything for me, i got exactly the same results after this update.
Last login: Mon Oct 1 10:40:24 2018 from 192.168.1.100
osmc@Vero4K:~$ iperf3 -c 192.168.1.100
Connecting to host 192.168.1.100, port 5201
[ 4] local 192.168.1.128 port 51492 connected to 192.168.1.100 port 5201
[ ID] Interval Transfer Bandwidth Retr Cwnd
[ 4] 0.00-1.00 sec 174 KBytes 1.42 Mbits/sec 33 4.28 KBytes
[ 4] 1.00-2.00 sec 34.2 KBytes 280 Kbits/sec 14 4.28 KBytes
[ 4] 2.00-3.00 sec 42.8 KBytes 350 Kbits/sec 11 4.28 KBytes
[ 4] 3.00-4.00 sec 2.85 KBytes 23.3 Kbits/sec 4 1.43 KBytes
[ 4] 4.00-5.00 sec 0.00 Bytes 0.00 bits/sec 1 1.43 KBytes
[ 4] 5.00-6.00 sec 0.00 Bytes 0.00 bits/sec 1 1.43 KBytes
[ 4] 6.00-7.00 sec 1.43 KBytes 11.7 Kbits/sec 8 2.85 KBytes
[ 4] 7.00-8.00 sec 22.8 KBytes 187 Kbits/sec 6 1.43 KBytes
[ 4] 8.00-9.00 sec 5.70 KBytes 46.7 Kbits/sec 3 2.85 KBytes
[ 4] 9.00-10.00 sec 8.55 KBytes 70.1 Kbits/sec 4 1.43 KBytes
[ ID] Interval Transfer Bandwidth Retr
[ 4] 0.00-10.00 sec 292 KBytes 239 Kbits/sec 85 sender
[ 4] 0.00-10.00 sec 221 KBytes 181 Kbits/sec receiver
I also tried this direct Vero>PC with no router or switch involved at all and results the same.
As I suspected looks like my Vero is faulty. I have dropped you a PM as I would like an exchange for a working unit asap.
thanks for your help in this.
Thanks @grahamh , that did the trick. Now getting just over 700 Mbits/sec with the 4K+ as client (and just under 900 Mbits/sec as server. I’m happy with that. I’ll stop fiddling!!
So, 11 days ago I raised the issue here, 5 days ago Sam says he’ll be following up with those of us with problems to get this resolved and then what I actually need to do is… send an email to the generic sales address?
OK.
@sam_nazarko would it be possible to describe the ‘bad’ device symptoms please and perhaps how you are testing them to filter them out of the supply chain?
I have posted my experience of 940Mb in one direction but only 45Mb in the other but have not had confirmation if this is normal or fixable?
I have also tried the hotfix you posted earlier in this thread but it made no real difference.
Thanks
Sam reads that e-mail, it’s not a generic address. But if you read this thread you can clearly see that the problem is being very actively worked on. It would be nice if you could wait for the final solution, but if you really want a refund I know Sam will cheerfully give you one.
And if you read the half dozen threads on the subject that I’ve been reading and posting in for the past 2 weeks, you’ll see I’ve been actively involved in testing the device. And it’s been confirmed there is a hardware issue. So, what solution do you suggest I wait for exactly?
What would be nice is if new (tested) devices had been sent out to the half dozen of us with problematic devices to test. But instead, whilst orders continue to be fulfilled, those of us with broken, faulty Veros, sit here on a thread for two weeks with entirely unhelpful people suggesting nonsensical network solutions to a hardware fault, whilst the fabled RMA solution doesn’t appear and the promised contact from Sam (from 5 days ago) doesn’t happen.
But thanks for the input.
If Sam hasn’t contacted you, it could be because he doesn’t think your particular device has a hardware problem. There were symptoms like yours in pre-production models which were eliminated with software adjustments so that all the testers got good performance with their switches. Now these new veros are getting exposed to a wider variety of network devices some new issues have turned up. I know Sam is working night and day to get those sorted. It will be solved.
Hi grahamh,
I’m totally a newbie on Linux embedded systems so I really do not where to start in checking whether my unit is healthy or not.
Is it possible to get a trusted procedure will all steps to be executed in order also to provide reliable feedbacks to Vero Team not to flood you with crappy info?
Thanks in advance
If you are using wired ethernet and haven’t noticed any problems then you are probably OK. But if you want to do some testing, have a look here: [How To] Check Network Performance with iperf3
Note that real-world throughput depends on what other traffic is on your network, cables, connections etc.
Many thanks for the link.
According to you, what are realistic data that are going to confirm the device is healthy if I’m going to execute the tests with only one server and one client?
Sam says above anything over 200Mbit/s and you don’t have one of the faulty units. If it’s less than that it could be hardware but it still could be all sorts of other things.
Sorry to hear about your frustration. Keep in mind that I didn’t receive a faulty switch + unit combination until Friday.
I then made some driver improvements available on Saturday which improved the stability of the Ethernet driver.
On Sunday, I then worked on improving the test suite so that no faulty units can be dispatched to a customer.
We haven’t shipped any units today as we’ve been preparing RMA units for affected customers and they will be resolved first. But it didn’t make sense to send more potentially faulty units out to customers.
I’ve collated the list of users in need of replacement and they’ll receive RMA instructions this evening.
Cheers
Sam
@matt.capes: Personally, I feel that Vero (and especially Sam) are doing a great job in finding out what is going on. I’m in hardware&software engineering myself and can tell you from first-hand experience that diagnosing a problem can be a painstaking job.
I rather have to wait for a bit, then to get a rushed ‘solution’ that isn’t a solution at all. It even might be two problems manifesting here: My Vero has really, really low values, I suspect it’s a hardware issue with my unit. Others report that the fix of Saturday did resolve their problems…
On the meantime, I’m on WiFi and can stream non-4k material perfectly.
So please, have some patience!
Based on a private-message conversation Sam and I were having earlier today, I think that statement needs qualifying a little.
There is a specific hardware issue that is known to affect the Ethernet on a small number of 4K+ units; Sam has been working to diagnose it. Based on his testing of affected boxes, we can say that if a Vero 4K+ can exceed 200Mb/s then it is not suffering from that issue.
However, the symptoms my 4K+ is exhibiting are rather different. Sam thinks it sounds like a hardware issue; I agree. My opinion (and I don’t know if Sam agrees with this, he may not!) is that it’s not the same hardware issue as the one he’s been researching up till now, but a different one.
If your box has the same problem mine has then it will sometimes go well above 200Mb/s, but not consistently - iperf numbers swing around wildly, even though there’s virtually no other traffic on the network, and can occasionally go as low as 4Mb/s or as high as 500. On mine, the traffic is also affected only in one direction, not both ways.