Internet is slow behind pfsense
-
i love Pfsense more than my wife :)
:o
I'd recommend to see if there's an upgrade pack for wife available ;D ;D ;D ;D
-
A more recent model of wife could be good to get the speed you want…
-
Try running a download test instead. You can download a file from a known good source directly on the pfSense box. That way you know if the restriction you're seeing is at the WAN or LAN. For example:
[2.1.5-RELEASE][root@pfsense.fire.box]/root(1): fetch -o /dev/null http://download.thinkbroadband.com/50MB.zip /dev/null 100% of 50 MB 8550 kBpsThat work well in the UK but you may want to choose some file closer to you. Though at 1Mbps you're probably fine.
Compare that with downloading the file behind pfSense or directly.
Steve
Thank you Steve,
this behind the Pfsense :
/root: fetch -o /dev/null http://download.thinkbroadband.com/50MB.zip /dev/null 8% of 50 MB 169 kBps 05m47sthis directly to the internet :
etch -o /dev/null http://download.thinkbroadband.com/50MB.zip /dev/null 53% of 50 MB 1965 kBps 00m14sPS : i've updated to the 2.2 release
-
So, are either of those on the pfSense VM itself?
-
And what vm do you have it setup on? You stated this was a VM. Workstation, Player, Esxi? How is it setup? How much cpu, how much ram - how the nics connected to your physical network/host? Are they bridged or natted?
So my pfsense is also vm..
[2.2-RELEASE][root@pfSense.local.lan]/root: fetch -o /dev/null http://www.wswd.net/testdownloadfiles/512MB.zip
/dev/null 30% of 512 MB 7284 kBps 00m52sNeeded a bit bigger file 50MB done so fast could get a good speed indication. And the UK bit far from me, wasn't seeing good speeds there.
this is from another vm behind pfsense, didn't have fetch so used wget
135,725,046 7.11MB/s eta 60sGoing to need some details of how your VM is setup, etc.
-
And what vm do you have it setup on? You stated this was a VM. Workstation, Player, Esxi? How is it setup? How much cpu, how much ram - how the nics connected to your physical network/host? Are they bridged or natted?
So my pfsense is also vm..
[2.2-RELEASE][root@pfSense.local.lan]/root: fetch -o /dev/null http://www.wswd.net/testdownloadfiles/512MB.zip
/dev/null 30% of 512 MB 7284 kBps 00m52sNeeded a bit bigger file 50MB done so fast could get a good speed indication. And the UK bit far from me, wasn't seeing good speeds there.
this is from another vm behind pfsense, didn't have fetch so used wget
135,725,046 7.11MB/s eta 60sGoing to need some details of how your VM is setup, etc.
Hi John thank you for trying to help me fix this.
the Guest OS :FreeBSD (64-bit)
Compatibility ESXI 5.5 ( VM Version 10 )
VMware Tools : Running version 2147483647
Pfsense has two Virtuals NICS, VMX3
NIC 1 is LAN
NIC 0 is WAN
2 CPUs
Memory 2048 MB, ( 112 USED )
Two Disks : 8 GB SSD for the OS, and 40 GB for cashing " Not configured yet "
Firewall rules are standart nothing special really, no VLANS.
NAT Port 443,Between Pfsense and Internet there is ISP Modem,
Thank you
So, are either of those on the pfSense VM itself?
yes they are the same VM, only first test is using Pfsense as it Gateway ( the slow one ) and second test is not ( the fastest one )
I am not saying pfsense is slowing the traffic, i think there is somewhere miss configuration.
i am using pfsense on a production and very happy about it . -
So run that command on the pfSense VM itself and you will know if the speed restriction is at the WAN or LAN interface.
Steve
-
So run that command on the pfSense VM itself and you will know if the speed restriction is at the WAN or LAN interface.
Steve
i am sorry not sure quiet i understand you,
what commands do i need to run ? -
Run the fetch command at the pfSense console.
If you see the full download speed there you know the throttling exists between pfSense and the LAN side clients.Steve
-
"VMware Tools : Running version 2147483647"
What tools? The native tools, there has been nothing but issues with the native tools.. I think that build number reflects 3rd party tools, where did you get that build number? That is not a valid number. The current version number of the openvm tools package is 1280544_10 that I see
I would suggest you do a clean install of pfsense, not install the tools - test speed. Then install the openvm tools not the native ones. What instructions did you follow, there are some things that need to be fixed up with the openvm tools to get the shutdown to work correctly, etc.
When I was playing with the native tools couldn't get anything to work other than ping, etc.
And what is your esxi host, and how is it connected to the real world? What vswitch setup do you have? Where exactly did you do those fetch tests on - you didn't do them on pfsense itself?
-
What tools? The native tools, there has been nothing but issues with the native tools..
Yeah, that package is totally horrible, best removed altogether from 2.2
-
Not sure where he is getting that build number for his tools, that is not a valid number.
on pfsense with the openvpn tools install I show this
[2.2-RELEASE][root@pfSense.local.lan]/root: vmware-toolbox-cmd -v
9.4.0.25793 (build-1280544)On a linux box with the current native tools I show this
ubuntu:~$ vmware-toolbox-cmd -v
9.4.11.42879 (build-2400950)If I google that buld number for vmware tools shows as unmanaged.. not a standard build number and is just the decimal value of 0x7fffffff, so not sure where or what he has installed for tools if any to be honest.
Other question I have is what build number of esxi 5.5 - the version of freebsd 10.1, in 2.2 is not fully supported until update 2 of esxi 5.5, build 2068190, the current build number is 2456374 which 4 patches past update 2.
I am running 2.2 pfsense on vm on esxi with the openvm tools and it is running fantastic.. Other than the apinger showing me that my gateway is like 1.2 ms, I WISH ;) I have really not seen any issues at all.
My connections in esxi host to the real world is like this - see attached. So would be curious to see how he has it setup.
My host has 4 physical nic ports connected to vswitches. Pfsense has a nic in each vswitch, the wan physical nic is directly connected to my cable modem. This puts a public IP on pfsense wan. Then there is wlan and lan and dmz. The vmkern is on its own nic not connected to pfsense just tied to the lan. When you share port groups with vmkern and another network (lan say) seems to slow down performance moving files to and from the datastore - since I have enough phsyical nics why not break it out, etc.
I run a vlan on the wlan nic in pfsense for guest wireless, etc. The lan and wlan physical nics connect to a managed switch with vlans for the lan and wlan/guestwlan segments because there a few physical devices on the wlan not just AP and controller, printer makes airprint easier when on the wireless network. And I put my dvrs which are wired on it as well since makes it easier or ipad app to find them, etc.
So curious how the OP setup is.. I would assume it should be something very sim to this.
-
I am experiencing a slow down on the wan port when running pfsense 2.2 on esxi 5.5 u2 with open-vm-tools install. Everything seems normal on the lan port. I have tried replacing nic and cables but the results are the same. Cannot find the problem.
Here are some test results running iperf client on pfsense
–----------------------------------------------------------
LAN side iperf server listening on TCP port 5001
TCP window size: 85.3 KByte (default)[ 4] local 192.168.5.22 port 5001 connected with 192.168.5.3 port 41971 (pfsense on esxi)
[ ID] Interval Transfer Bandwidth
[ 4] 0.0-10.0 sec 712 MBytes 596 Mbits/sec
[ 5] local 192.168.5.22 port 5001 connected with 192.168.5.1 port 37565 (pfsense on hyperv)
[ 5] 0.0-10.3 sec 684 MBytes 559 Mbits/sec
[ 4] local 192.168.5.22 port 5001 connected with 192.168.5.2 port 51702 (pfsense on atom pc)
[ 4] 0.0-10.0 sec 442 MBytes 370 Mbits/sec–----------------------------------------------------------
WAN side iperf server listening on TCP port 5001
TCP window size: 85.3 KByte (default)[ 4] local 192.168.1.22 port 5001 connected with 192.168.1.3 port 39355
[ ID] Interval Transfer Bandwidth
[ 4] 0.0-11.2 sec 7.12 MBytes 5.32 Mbits/sec <-?????? esxi pfsense running iperf client to wan side server
[ 5] local 192.168.1.22 port 5001 connected with 192.168.1.1 port 54856
[ 5] 0.0-10.9 sec 700 MBytes 536 Mbits/sec
[ 4] local 192.168.1.22 port 5001 connected with 192.168.1.2 port 58495
[ 4] 0.0-10.9 sec 537 MBytes 413 Mbits/sec -
Hi John,
i've run the command on the box to detect the install vmware tools[2.2-RELEASE][root@firewall.pfsense.lan]/root: vmware-toolbox-cmd -v 9.4.0.25793 (build-1280544) [2.2-RELEASE][root@firewall.pfsense.lan]/root:my switch is 1Gigabit switch and all ports are 1GB
on my ESXI 5.5 i've got 4 1GB NICS, and have all of them on one group.
all the traffic is travelling on those 4 NICS at once over 1Gigabit switch ( HP 1810-24G v2 ).
i have the same switch and configuration on production all seems to works fine even faster .
i have removed the vmware tools and reinstalled them but the issue still exist .my ESXI 5.5 is having the latest patches.
are you saying the pfsesne 2.2 isn't supporting the latest U2 Patches ? because my production is patched Build Number: 2456374 -
"on my ESXI 5.5 i've got 4 1GB NICS, and have all of them on one group."
What?? That makes NO SENSE – how do you isolate between your wan and your lan on pfsense..
So you have 1 vswitch? Please post your esxi setup like I did for vswitches and lets see this setup of your 4 nics - did you setup a lagg on your switch?
-
Long time no bridge… ::)
-
"on my ESXI 5.5 i've got 4 1GB NICS, and have all of them on one group."
What?? That makes NO SENSE – how do you isolate between your wan and your lan on pfsense..
So you have 1 vswitch? Please post your esxi setup like I did for vswitches and lets see this setup of your 4 nics - did you setup a lagg on your switch?
between pfsense and internet there is my ISP Modem.
this is a lab and the VM's are having a static iP, and the ISP Modem DHCP is off.
so all the traffic is going through 4 NICS ( WAN and LAN ).yes using one vSwitch but all the VMS are statics.
-
So your trying to run 2 networks over the same wire (wan and lan)?? Just by changing the ips.. Yeah sorry it doesn't work that way..
Break out a nic for you WAN, then another one for your LAN put them on 2 different vswitches. I would also break out your vmkern.. IF you have a managed switch you can use the same switch with vlans. But your wan connection should really just be direct to your esxi host interface..
-
So your trying to run 2 networks over the same wire (wan and lan)?? Just by changing the ips.. Yeah sorry it doesn't work that way..
Break out a nic for you WAN, then another one for your LAN put them on 2 different vswitches. I would also break out your vmkern.. IF you have a managed switch you can use the same switch with vlans. But your wan connection should really just be direct to your esxi host interface..
thank John,
you mean 1 NIC for the WAN and Managenement of the ESXI and 3 Groups NICS for the LAN and traffic.Like WAN = gonna use Vlan 2
and LAN= Vlan 3
ect.. ?According to VMware NIC Teaming is the best practise for ESXI provide reduncency, speed, and increase the netwrok capacity
isn't John ?
-
dude this is a LAB not some mission critical production setup.. And even if it was you sure wouldn't run 4 nics all to the same vswitch for both wan and lan networks. Was it a trunk and you were setting up vlans on pfsense? If so you made no mention of vlans in your setup.
Break out your NICS What is your internet speed again, how and the F do you think you need 3 nics to lan?? If you want to do failover then put 2 on each and sure you can share your vmkern port with your lan. But it hurts performance to and from the datastore for moving files. This is normally rare - but you have 4 nics to play with.
Doesn't really matter what vlan numbers you put on them.. They will be physically isolated I would hope - but sure if you have to run your wan through your switch it can be on a different vlan. Pfsense doesn't really care in that sort of setup.
I can tell you what vmware doesn't say is try and run 2 different networks over the same freaking wire that is for damn sure.. So what mode did you have these nics in when groups? teamed/lagged/loadbalanced/failover? Did you lagg them on the switch?
