Score:-1

Slow Network Speed (11-12 MB/s) on Linux Home Server using Gigabit-capable Hardware

cr flag

I am currently trying to solve an issue with my tiny linux home server that's been causing slow network speeds when accessing SAMBA shares and transferring files to/from other computers on the network. Despite having a gigabit-capable network setup, the server is only achieving speeds of around 11-12 MB/s, which seems to me like something is limiting the network speed to 100 megabits. This is also the maximum speed I get when downloading files from the internet. Other devices connected to the same switch/router have no trouble achieving full gigabit speeds on the local network and around 800 Mbps downspeed from the internet, consistent with what I normally get out of my gigabit plan most of the time.

This is the setup:

  1. Fujitsu Esprimo Q920 running debian
  2. TP-Link Archer C6 with latest firmware
  3. All devices connected via CAT 7 cable

ethtool enp0s25:

Supported ports: [ TP ]
Supported link modes:   10baseT/Half 10baseT/Full
                        100baseT/Half 100baseT/Full
                        1000baseT/Full
Supported pause frame use: No
Supports auto-negotiation: Yes
Supported FEC modes: Not reported
Advertised link modes:  10baseT/Half 10baseT/Full
                        100baseT/Half 100baseT/Full
                        1000baseT/Full
Advertised pause frame use: No
Advertised auto-negotiation: Yes
Advertised FEC modes: Not reported
Speed: 1000Mb/s
Duplex: Full
Auto-negotiation: on
Port: Twisted Pair
PHYAD: 1
Transceiver: internal
MDI-X: on (auto)
Supports Wake-on: pumbg
Wake-on: g
Current message level: 0x00000007 (7) drv probe link

What I tryed so far:

  1. Attempted different CAT 7 cables to rule out a cable issue
  2. Experimented with forcing network mode to 1000 Mbps full duplex using ethtool
  3. Disabled autonegotiation on both the server and the network switch, manually setting the connection to 1000 Mbps full duplex

Your assistance is greatly appreciated. Thank you in advance for your expertise!

Jaromanda X avatar
ru flag
what speeds does your *tiny linux home server* achieve when downloading from the internet? comparing internet speed test with SAMBA (notoriously for being slow) is like comparing a race car with a Citroën 2CV ... NOTE: your ethtool output clearly shows link speed is `Speed: 1000Mb/s` so, the bottleneck is something else, stop futzing with the network adapter settings :p
Alex avatar
cr flag
Oh I did not know that SAMBA is not the best tool to benchmark this. I also get 100megabit speeds when downloading files from the internet. Just tested with the ubuntu image and got average 10.2mb/s. The ethtool output is actually what makes me wonder. Any idea what could be the bottleneck in this scenario? Something I could check?
Jaromanda X avatar
ru flag
OK, so your *tiny linux home server* is obviously connected to something with a gigabit port. But what is that connected to? no idea what your network layout is, but the speed of data between two devices is limited by the lowest link speed along the way. (ignoring any software shenanigans like traffic shaping of course)
Alex avatar
cr flag
The internet comes through a basement cable modem, connecting to a TP-Link Archer C6 in my ground floor office, which is then linked to another Archer C6 in the living room on the same floor. CAT 7 cables are used for all physical connections. I typically achieve around 800mbits downstream from the web in my office and living room, showing good performance. Even my Phone gets about 300mbits from the Archer's 5GHz WiFi right now, suggesting the network can handle more than the home server's 100mbits. So should be an issue with the server pc itself I guess
Jaromanda X avatar
ru flag
probably, since it can only do 10mb/s (that's bits or bytes?)
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.