Are there any tuning guidelines for improving the network performance of Nexus?
As part of our build process, we’re generating some large images (~9GiB) which are uploaded to a raw repository. We have 10Gbps connectivity between the machines, yet the upload into Nexus can take in excess of 5 minutes, which is a lot slower than expected.
Using iperf I can see we can get close to the wire speed of the NICs, so there doesn’t appear to be any network issue on the machines involved.
This is a single machine install of OSS 3.56.0-01 running on Windows Server 2019.
I’ve also tried with OSS 3.59.0-01 running on Debian 12, which is a bit quicker (3 minutes), but nowhere near as fast as I would expect.
Don’t know if my question is relevant but do you mean you have 9GB docker images you upload to raw repositories ? Why don’t you use upload them to a docker repository instead of raw ?
By the way if you mean network bandwidth isn’t an issue, could it be a disk issue on your host ?
Dom
It’s a live OS image, rather than docker, hence why we’re using a raw repository.
When I say network bandwidth isn’t an issue, what I mean is that the measured network performance between the hosts involved is really good - close to 10Gbps. However, Nexus doesn’t seem to be using it.
The disks aren’t the fastest, but I can create a 9GiB file in a little over a minute so I don’t think that’s the limiting factor.