All Questions
Tagged with performance networking
183
questions
0
votes
1
answer
116
views
Why MACsec can not reach more than 2Gbits/s?
i am configuring macsec and i have discovered that, when using this protocol to send/received data, the bitrate is so low.
I am using it over a 10Gb network adaptor, using iperf3 it almost reach that ...
0
votes
0
answers
296
views
Interpret Iperf3 results for bidirectional UDP Test
I am trying to make sense of the raw logs of an iperf3 UDP bidIrectional test.
Here are the iperf3 client logs:
[ ID][Role] Interval Transfer Bitrate Jitter Lost/Total ...
- 101
0
votes
0
answers
50
views
Lower throughput in a Bridge created as a Ethernet Switch in a Ubuntu Server -
The setup is given below. I have a server with NIC(eth0 and eth1). I have created a bridge br0. The idea is I want to send the packets received from eth0 via br0 to eth1.
(Laptop)__________Ubuntu ...
-2
votes
1
answer
114
views
Discovering the speed of data transfer between my server and another server that I don't have access to [closed]
What solution do you have so that I can get the speed of downloading from a server or uploading to a server on the TCP or UDP protocol, which I do not have access to that server? but I have root ...
- 97
1
vote
0
answers
75
views
Poor forwarding performance between two WireGuard tunnels
I've got a dedicated server (Ubuntu Server 20.04), running two WireGuard interfaces:
wg0 is used to connect end-users, like laptops, mobile devices, etc.
wg1 is a tunnel to a 3rd-party VPN service
...
- 296
0
votes
1
answer
723
views
Simple haproxy TCP passthrough results in very slow network transfer speed
I've setup a simple haproxy instance on a clean install of Debian 10 Buster. I've added some simple necessary config to enable the passthrough to the IP address in question (which has been redacted in ...
- 1,636
2
votes
1
answer
258
views
How to scale OpenVPN when client-to-client is a must?
I am trying to create an OpenVPN server cluster that can autoscale, I have found a lot of information online on how to create such a network.
As far as I understand, you basically use a round-robin ...
- 58
-2
votes
3
answers
309
views
I live in Australia, how is it that Stack Exchange is so fast without using any CDN?
I had a look at the Stack Exchange Servers:
https://meta.stackexchange.com/questions/10369/which-tools-and-technologies-are-used-to-build-the-stack-exchange-network
Seems they only have two data ...
- 1,345
1
vote
0
answers
36
views
Network performance with Kubernetes
I am running a small real time bidder application in Docker Swarm using host mode for networking. I found that other networking modes were giving much higher latency.
Soon I plan to switch from Docker ...
0
votes
3
answers
63
views
Accelerate loading time in others countries by installing some kind of VPN? (What is it called?) [closed]
My servers are hosted in France (OVH) and my users are mainly from France.
I am starting to have more and more users in Asia and North America, but some of them have huge speed problems : more than 3-...
- 171
0
votes
2
answers
577
views
Is there a speed and performance difference between VPNs (L2TP, etc.) vs. port forwarding?
Let's say you wanted to transfer a 1 GB file from a machine at site 1 to a machine at site 2. You can either connect the two machines via a VPN (L2TP) or basic port forwarding. Which connection method ...
- 559
1
vote
0
answers
586
views
Apache2 mod_dumpio logged output bytes does not match %O in access.log
I'm trying to understand why responses sent by my server are long despite the content (+header) being small.
For that I'm using mod_dumpio to log all data received and sent by apache. But the logs ...
- 71
6
votes
2
answers
2k
views
Why are concurrent downloads faster
Many download managers like this, this and this support downloading a file over multiple parallel connections, one per thread.
The concept is that each connection will download one part of the file ...
0
votes
2
answers
946
views
Slow initial connection to database (php / mssql)
Our webserver and database server are on a local network and the webserver uses an IP address to connect to the database.
$con = new \PDO('sqlsrv:Server=10.200.2.1\sql2017;Database=dbname, $user, $...
- 73
0
votes
1
answer
459
views
Poor network performance on host machine using libvirtd / Qemu / KVM
We've got two Fedora 27 servers that we use to host several virtual servers using libvirtd / Qemu / KVM that are experiencing the same issue. When freshly booted, network performance is great, pings ...
- 1
1
vote
1
answer
3k
views
Very high tcp retransmit rate to Internet on devices inside LAN
I work with Linux systems a lot as a developer and have a moderate level of networking knowledge, but the following situation is a real baffling mystery:
I'm running pfSense 2.4.3 (that's current) on ...
- 163
3
votes
1
answer
591
views
UDP. Sending speed is limited by the minimal Ethernet adapter configuration
I have 1Gbit Ethernet adapter on my computer. And I have the other computer with Ethernet adapter 100Mbit. The both in one LAN and connected through the 1Gbit switch.
1 Test: I try to send from my ...
1
vote
0
answers
69
views
Last 4 bytes of HTTP response come after >30s delay
I'm encountering a weird error with one of my websites. No matter which page I'm trying to display, about 3 out of 5 cases and up loading in no less than 30 seconds...
When loading the website with ...
- 671
-1
votes
2
answers
113
views
SSH server grinding to a halt when users grep large files - high NET_RX and high softirqs - How to limit?
I have a server running fedora 14 - Linux-2.6.35 with a 1Gbps NIC. It has NFSv3 mounts supplied by an Isilon server.
It freezes whenever any user works with a large file(grep or similar) that exists ...
- 167
5
votes
1
answer
6k
views
How to debug slow server response time?
Using the method described here to measure server response time, I can see that time_appconnect, time_pretransfer, and time_starttransfer are extremely high, but only when accessing the website ...
- 151
3
votes
1
answer
7k
views
How to enable ethernet 1 Gigabit advertising or troubleshoot the situation?
On an Debian Stretch box, I have a Gigabit ethernet network adapter, but this mode is not active (however displayed as supported).
Does this depend on the peripheral condition (i.e. cable is not ...
- 179
1
vote
1
answer
747
views
Setup a private network between my 7 dedicated servers in different data centers?
I have 7 dedicated servers running at a host provider.
Whenever I order a new server I cannot rely on having my servers being setup in the same rack, so I cannot have a physical private network ...
- 358
1
vote
2
answers
20k
views
What is the minimum gigabit Ethernet latency I can expect? [closed]
I have 2 servers sitting in a rack running Ubuntu 16.04 with a 1 meter Ethernet cable between them, both having standard Intel Ethernet adapters.
The ping between the two is about 300 us (...
- 858
1
vote
2
answers
384
views
How to interpret traceroute results
I am trying to figure out how to interpret the attached image.
First two hops are on our LAN, hops 3-5 belong to our WAN provider, while hop 6 is at our hosting provider and the hop 7 is a Windows ...
7
votes
2
answers
2k
views
Why multicast UDP can be slower than unicast UDP?
I have two machines (with Centos 7.3 on both) connected via 1G switch.
When I measure UDP thoughput between macines with iperf it gives about 850 Mbit/s.
However when I measure multicast UDP thoughput ...
- 173
1
vote
2
answers
273
views
Outgoing network bandwidth is high even for empty response [closed]
I have a simple nginx server that monitors the requests and store them in the access logs. The requests contain data in query parameters and are about 500 bytes. My output is just a HTTP 204 response. ...
- 11
-1
votes
2
answers
55
views
VPS vs Real Hardware performance correlation
Assuming I have application running in a VPS and consuming 50% of assigned CPU resource. Will it correct to assume that on real hardware with same properties CPU usage will not be more than 50% ...
- 210
0
votes
1
answer
3k
views
Windows file server got slow response and high latency even ping localhost by running about 7 days after rebooting
We have a file server of Windows Server 2012 R2 on DELL PowerEdge R720, and get a very weird issue in recent days about network/disk performance. The box have a RAID-1 disk group for OS (disk0) under ...
2
votes
2
answers
4k
views
Speeding up vCenter Template Cloning
My VM templates are <4GB in size, and when I finish cloning them into a Virtual Machine, they are still <4GB in size:
The template:
The VM created from that template:
My ESXi machine is ...
- 315
1
vote
2
answers
314
views
Methodology to find a "stable" IP addres within some ISP's AS network
Some definition(s)/meaning(s):
Stable IP address - a host within Internet Service Provider (ISP) Autonomous system (AS) network, replaying to ICMP type 8 echo request.
Genesis:
My routers monitor ...
- 1,471
1
vote
0
answers
228
views
slow single connection between two otherwise fast machines
I'm having trouble with larger file transfers from one of our virtual servers to our office network. The virtual server acts as a VPN server, therefore we call it "VPN", but the usage of VPN is not of ...
- 141
0
votes
1
answer
111
views
Internet routing through linux speeds down
I'm having the most odd problem:
I'm providing internet routing for a network of pcs through a server based on linux. I used a broadband modem and everything worked fine.
Now, to the question, I got ...
- 1
1
vote
1
answer
76
views
Understanding network performance impact of 32 bytes per request
My company is using redis on a fairly performance critical path. An nginx server calls out to it once per request. The call itself has several arguments, which are currently like 80 bytes. This call ...
- 113
10
votes
4
answers
13k
views
Small writes to SMB network share are slow on Windows, fast over CIFS Linux mount
I have been struggling to fix a performance problem with a SMB/CIFS share when performing small writes.
First, let me describe my current network setup:
Server
Synology DS215j (with SMB3 support ...
- 103
3
votes
1
answer
2k
views
Download speed low over http but Iperf is showing a high bandwidth
I was trying to download some files off of a server that I recently started renting. It is a dedicated server located in Paris-FR. I believe the box is in online.net's datacenter. I am renting it from ...
- 31
9
votes
2
answers
515
views
Higher TCP latencies in the latest versions of Linux
In my research group, we recently upgraded the OS on our machines from Red Hat 6.2 to Debian 8.3 and observed that the TCP round trip time through the integrated Intel 1G NICs between our machines had ...
- 91
13
votes
3
answers
2k
views
Effect of distance from server
I have a web server in Ireland (Amazon AWS). This server appears fast from Germany (orange line) but slow from the USA (black line). The HTTP request used for the test is the same.
I think this is ...
- 485
2
votes
1
answer
3k
views
Huge performance penalty in gbit link with ip forwarding / masquerade through clear VPN tunnel
I have a problem with IP forwarding between my home computer and a remote server (from online.net). My home connection is a 1 Gbit/s FTTH and the remote server has a 2.5 Gbit/s bandwidth.
The context
...
- 129
3
votes
2
answers
426
views
Measuring 40Gbe performance
I've connected two hosts via a 40Gbe fibre link and would like to determine a baseline for data transfer between them. I'm seeing relatively low speeds (~620MB/s). It seems that the signalling rate is ...
- 343
8
votes
3
answers
9k
views
How to load balance with respect to memory, disk usage, and other attributes
I've found load balancers such as NGINX but those seem to work only by keeping in mind CPU usage and network traffic. How would I load balance with respect to other variables such as the amount of ...
- 199
1
vote
0
answers
461
views
iperf actual send rate always lower than the expected target bandwidth until some equilibrium reached. Why?
I did an UDP iperf 2.0.5-Ubuntu test between two VMs on the same hypervisor. As can be seen in the iperf output below I started my first test with "-b 686M" flag. However, iperf reported that it was ...
- 113
2
votes
1
answer
1k
views
What's your experience with the AWS NAT Gateway performance?
In December/2015 Amazon launched their VPC NAT Gateway service which is basically a managed alternative to creating your own (and it looks very good in comparison to NAT instances).
I was just about ...
- 830
13
votes
1
answer
2k
views
Virtualization - Ten 1Gbps links or one 10Gbps link? (Performance)
I have a machine that has several VM's (5) and 3 physical networking cards (with each 2 ports), with a total of six 1Gbps ethernet ports.
I have a SPF-capable switch, having a total of 48Gbps ...
- 289
0
votes
0
answers
620
views
How to find the bottleneck for slow rsync
I want to find the bottleneck for a slow rsync.
I see these parts (on byte flow order):
IO on source
CPU on source
Network
CPU on destination
IO on destination
How can I debug this?
- 3,603
2
votes
3
answers
147
views
Check the cause of slow loading time among different server
A e-shop has developed using perstashop and put to the three server:
the first 2 are amazon EC2, should be same setting
Server 1: http://be-pure.com/en/women/3-slim-y-tank.html
Server 2: http://52....
- 121
0
votes
1
answer
2k
views
Router - huge difference in Tx vs. Rx
I have two computers:
(B) Banana Pi - 1 Gigabit Ethernet, Wi-Fi (supports 802.11n)
(D) Desktop PC - 1 Gigabit Ethernet, Wi-Fi (2x2 MIMO, supports 802.11n)
And (R) Mikrotik RouterBOARD RB951G-2HnD - ...
- 1,872
7
votes
2
answers
18k
views
Page load long stall times. What is this? [closed]
I have been studying the performance of a website I am building, and notice a curious thing that I can't find the answer to: two requests (always the same two) mysteriously have prolonged stalled ...
- 131
1
vote
1
answer
84
views
Different network speeds from servers on same subnet
I have a pretty strange issue, I'm looking for advice on how I can debug it.
I have two servers (A) and (B) on the same subnet and switch. Server A->B transfers at a full 12MB/s and vice-versa.
...
- 800
-2
votes
1
answer
69
views
Servers in different DC's - performance impact
Today Linode eat my nerves completely by shutting down main DB server for 25 minutes due to their internal network problems.
Right today I was going to set up another one server at linode to set up ...
- 97
6
votes
1
answer
1k
views
4X10GB Bonded NIC - only producing 4-5GB throughput
I have Two HPBL685c G6 blade servers running Ubuntu 15.04
Each server has 4X10GB NIC
2x10GB NIC are connected to a single VirtualConnect 10/10G Ethernet module
The other 2x10GB NIC are connected to a ...
- 149