Score:1

NETGEAR NAS capped at 1.3MB/s transfer rate across VPN

cl flag

We have many NAS devices at our company and one is remote. It is connected to a Meraki MX64 which connects to our local network at the company. While transferring data over to it, the speed seems to cap out at 1.3MB/s very specifically. I'm wondering what sorts of settings to look for as to why this rate would be capped, or just any reasoning at all. From what I can see, neither of the connecting Meraki devices have their rates capped that low, nor is there anything on our Netgear devices that should limit it to that rate specifically.

This is especially weird since I saw the rate at several times this speed (I believe it was around 30MB/s, possibly more or less) just a few weeks ago. I haven't changed any settings anywhere to prompt this change in rate, but I did notice that this current backup job started fast but very slowly and steadily decreased to this rate. This is not the first time I have seen the rate cap out at 1.3MB/s, but usually the job would start out at an extremely tiny pace and climb up to the 1.3MB/s point as opposed to slowing down to it over time.

One thing to note is that 1.3MB/s converts to 10.4Mb/s. If we assume that the actual transfer rate is 10Mb/s, which is 1.25MB/s, then it's possible that I am seeing 1.3MB/s because the web interface is rounding up. I'd imagine that this might have something to do with what could be limiting it, since 10Mb/s is such a round, clean number. Any ideas would be appreciated.

Edit: After some testing, the issue does not seem to be with any settings on the NAS devices. Both links are set to 1Gb/s, 4500MTU, and the measured bandwidth between my computer and the remote NAS is 11Mb/s aka 1.3MB/s. I cannot test the speed on other devices on the remote network because there currently are no other devices. The issue likely lies somewhere in the middle, at any rate.

vidarlo avatar
ar flag
What kind of VPN? What performance does iperf3 show across the same link? What protocol is used to access the file share? What is the latency of the link? What is the raw performance of the underlying links of the VPN?
JasonB576 avatar
cl flag
@vidarlo The VPN is the built-in Meraki Site-to-Site VPN, no special settings or details there afaik. I can't do iperf3 at the moment but this does look like a useful tool. Protocol is RSync. Average lag is 34ms according to a ping. Not sure what you mean by "raw performance of the underlying links."
vidarlo avatar
ar flag
What performance can the links at either end support? And do a measurement with iperf3, to have an idea of where the problem is. Make it a priority in fault finding...
JasonB576 avatar
cl flag
@vidarlo Sorry for the long silence, had issues with the remote NAS. I got iperf running and it measures the bandwidth as the speed listed by the backup jobs: 11Mb/s aka 1.3MB/s. The links at either end support and are set at 1Gb/s. Discrepancy must be somewhere in the middle, then? Either way, thank you for showing me iperf. This tool rocks.
vidarlo avatar
ar flag
Certainly. I'd start with your meriaki devices. I would suggest to [edit] your question with the new information, and focus it on increasing VPN performance. Update the tags as well as the question :) The storage performance I think is sufficiently ruled out.
vidarlo avatar
ar flag
You write that you have 4500B MTU. What size packages are sent over the VPN? I would try to reduce MTU to the normal 1500B, and see how performance is affected. If possible, try with 1400 bytes as well, as that should allow for some overhead...
mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.