well, this is what I got:
bypass the firewall (just 2 PCs connected via the switch):
iperf -c server-ip -t 60 -M 500
380Mb/s
iperf -c server-ip -t 60 -M 500 -d
477Mb/s
422 Mb/s
comparing to the tests with bce driver:
iperf -c server-ip -t 60 -M 500
300Mb/s
52 -85kpps
iperf -c server-ip -t 60
From: Lenny [mailto:five2one.le...@gmail.com]
Sent: Tuesday, August 04, 2009 4:04 AM
To: support@pfsense.com
Subject: Re: [pfSense Support] Re: Can't get
more than 15kpps
On Sun, Aug 2, 2009 at 12:21 PM, Tim Dressel tjdres...@gmail.com wrote:
Install on both sides, not on pfsense.
i.e. install on a machine on the WAN side, and on the LAN site. Or if
you are testing between LAN and an OPT interface, put a machine on
both subnets and test that way.
iPerf on
On Mon, Aug 3, 2009 at 3:52 AM, Lennyfive2one.le...@gmail.com wrote:
On Sun, Aug 2, 2009 at 12:21 PM, Tim Dressel tjdres...@gmail.com wrote:
Install on both sides, not on pfsense.
i.e. install on a machine on the WAN side, and on the LAN site. Or if
you are testing between LAN and an OPT
On Sat, Aug 1, 2009 at 2:15 AM, David Rees dree...@gmail.com wrote:
Most of the time, the real issue is that scp has to encrypt the data
on one end and decrypt it on the other - that takes a lot of CPU power
that could otherwise be used for tossing packets around.
-Dave
Install on both sides, not on pfsense.
i.e. install on a machine on the WAN side, and on the LAN site. Or if
you are testing between LAN and an OPT interface, put a machine on
both subnets and test that way.
iPerf on pfsense will not give you a throughput of the firewall (at
least nothing that
On Sun, Aug 2, 2009 at 12:21 PM, Tim Dressel tjdres...@gmail.com wrote:
Install on both sides, not on pfsense.
i.e. install on a machine on the WAN side, and on the LAN site. Or if
you are testing between LAN and an OPT interface, put a machine on
both subnets and test that way.
iPerf on
Lenny wrote:
So do you have any other ideas? I NEED this to work.
just for a sanity check, could you boot a live linux CD and make various
tests with that (iptraf, timed netcat etc)?
-
To unsubscribe, e-mail:
Paul Mansfield wrote:
Lenny wrote:
So do you have any other ideas? I NEED this to work.
just for a sanity check, could you boot a live linux CD and make various
tests with that (iptraf, timed netcat etc)?
-
To
Lenny wrote:
But to my biggest shame I'd never made those tests. Is there a chance
you'd give me some pointers?
Thanks.
boot a live linux disk like ubuntu
try a speed test website.
for network testing...
set up the interfaces
create a 1G test file, e.g. dd if=/dev/urandom of=/tmp/random
Paul Mansfield wrote:
boot a live linux disk like ubuntu
try a speed test website.
for network testing...
set up the interfaces
create a 1G test file, e.g. dd if=/dev/urandom of=/tmp/random bs=1024
count=1048576
then use time scp /tmp/random otherhost:/tmp/blah or use netcat -l -p
1234 on
Paul Mansfield schrieb:
boot a live linux disk like ubuntu
try a speed test website.
for network testing...
set up the interfaces
create a 1G test file, e.g. dd if=/dev/urandom of=/tmp/random bs=1024
count=1048576
then use time scp /tmp/random otherhost:/tmp/blah or use netcat -l -p
Rainer Duffner wrote:
Paul Mansfield schrieb:
boot a live linux disk like ubuntu
try a speed test website.
for network testing...
set up the interfaces
create a 1G test file, e.g. dd if=/dev/urandom of=/tmp/random bs=1024
count=1048576
then use time scp /tmp/random otherhost:/tmp/blah
I may be wrong, but his problem is pps (packets per second).
That's not the same as being able to download a large file.
Unfortunately.
How does one generate a large a mount of (small) packets with useful
an genuine traffic?
actually you're right. But I think there was a tool for that.
Rainer Duffner wrote:
I may be wrong, but his problem is pps (packets per second).
That's not the same as being able to download a large file.
Unfortunately.
How does one generate a large a mount of (small) packets with useful
and genuine traffic?
set the MTU to a low value (200?) so that
On Fri, Jul 31, 2009 at 10:30 AM, Paul
Mansfieldit-admin-pfse...@taptu.com wrote:
Rainer Duffner wrote:
I may be wrong, but his problem is pps (packets per second).
That's not the same as being able to download a large file.
Unfortunately.
How does one generate a large a mount of (small)
On Fri, Jul 31, 2009 at 7:16 AM, Rainer Duffnerrai...@ultra-secure.de wrote:
Paul Mansfield schrieb:
boot a live linux disk like ubuntu
try a speed test website.
for network testing...
set up the interfaces
create a 1G test file, e.g. dd if=/dev/urandom of=/tmp/random bs=1024
then use time scp /tmp/random otherhost:/tmp/blah or use netcat -l -p
1234 on one to create a listen and on other time cat /tmp/random |
netcat -p 1234 otherhost to see how long it takes
scp doesn't perform well over fast links, it's not really a good tool
for testing. I can barely get
On Fri, Jul 31, 2009 at 12:09 PM, Keenan Timskt...@gotroot.ca wrote:
then use time scp /tmp/random otherhost:/tmp/blah or use netcat -l -p
1234 on one to create a listen and on other time cat /tmp/random |
netcat -p 1234 otherhost to see how long it takes
scp doesn't perform well over fast
On Wed, Jul 29, 2009 at 11:38 PM, Chris Buechler c...@pfsense.org wrote:
On Wed, Jul 29, 2009 at 3:38 PM, Lennyfive2one.le...@gmail.com wrote:
That's all understandable when speaking of errors and packet loss, but
would
it really cause the CPU hit 100% at 50kpps?
both em0 and em1?
On Wed, Jul 29, 2009 at 11:27 PM, Evgeny Yurchenko
evgeny.yurche...@frontline.ca wrote:
*From:* Lenny [mailto:five2one.le...@gmail.com]
*Sent:* July 29, 2009 3:38 PM
Evgeny Yurchenko wrote:
I would try to swap cables and interfaces in config and see errors. Do they
go to em0?
On Wed, Jul 29, 2009 at 11:27 PM, Evgeny Yurchenko
evgeny.yurche...@frontline.ca wrote:
10kpps - 20% CPU
50kpps - 100% CPU
looks like we have some logic here. I've looked at my graphs - there is no
relation between cpu load and pps. Do you have this relation? Not talking
about your
On Thu, Jul 30, 2009 at 3:25 AM, Lennyfive2one.le...@gmail.com wrote:
I have in the past, not with the recent setup. As far as I understood it
doesn't help much in the latest releases of FreeBSD.
It can if you're getting killed by interrupts but that doesn't seem to
be the case.
On Thu, Jul 30, 2009 at 4:00 PM, Chris Buechler cbuech...@gmail.com wrote:
On Thu, Jul 30, 2009 at 3:25 AM, Lennyfive2one.le...@gmail.com wrote:
I have in the past, not with the recent setup. As far as I understood it
doesn't help much in the latest releases of FreeBSD.
It can if
From: Lenny [mailto:five2one.le...@gmail.com]
Sent: July 30, 2009 3:28 AM
On Wed, Jul 29, 2009 at 11:27 PM, Evgeny Yurchenko
evgeny.yurche...@frontline.ca wrote:
From:
On Thu, Jul 30, 2009 at 4:07 PM, Evgeny Yurchenko
evgeny.yurche...@frontline.ca wrote:
Weird, I do not have any relation between cpu and bandwidth/packets:
http://img43.imageshack.us/img43/4127/bandwidth.png
http://img78.imageshack.us/img78/8375/cpu.png
From: Lenny [mailto:five2one.le...@gmail.com]
Sent: July 30, 2009 9:16 AM
On Thu, Jul 30, 2009 at 4:07 PM, Evgeny Yurchenko
evgeny.yurche...@frontline.ca wrote:
Weird, I do not have
On Thu, Jul 30, 2009 at 4:25 PM, Evgeny Yurchenko
evgeny.yurche...@frontline.ca wrote:
My traffic spike is between em and bge interfaces... I have another box
with two bge interfaces with load peaking at 250Mb/s and packets 24kpps and
there I have cpu-bandwidth relation.
If you do not mind
On Thu, Jul 30, 2009 at 4:25 PM, Evgeny Yurchenko
evgeny.yurche...@frontline.ca wrote:
My traffic spike is between em and bge interfaces... I have another box
with two bge interfaces with load peaking at 250Mb/s and packets 24kpps and
there I have cpu-bandwidth relation.
If you do not mind
On Thu, Jul 30, 2009 at 9:32 AM, Lennyfive2one.le...@gmail.com wrote:
bce0: Broadcom NetXtreme II BCM5708 1000Base-T (B2) mem
0xc800-0xc9ff irq 18 at device 0.0 on pci4
Are things any better/different if you use the onboard Broadcom NICs instead?
On Thu, Jul 30, 2009 at 8:06 PM, Chris Buechler c...@pfsense.org wrote:
On Thu, Jul 30, 2009 at 9:32 AM, Lennyfive2one.le...@gmail.com wrote:
bce0: Broadcom NetXtreme II BCM5708 1000Base-T (B2) mem
0xc800-0xc9ff irq 18 at device 0.0 on pci4
Are things any better/different if you
On Thu, Jul 30, 2009 at 1:17 PM, Lennyfive2one.le...@gmail.com wrote:
Actually I've never tried with this server. Only with the old one and then
the interrupt was pretty high.
But I remember you and the other guys advised against using Broadcom in
favor of Intel.
Are you suspecting the NIC
On Thu, Jul 30, 2009 at 8:21 PM, Chris Buechler c...@pfsense.org wrote:
On Thu, Jul 30, 2009 at 1:17 PM, Lennyfive2one.le...@gmail.com wrote:
Actually I've never tried with this server. Only with the old one and
then
the interrupt was pretty high.
But I remember you and the other guys
Date: Thu, 30 Jul 2009 20:24:27 +0300
From: five2one.le...@gmail.com
To: support@pfsense.com
Subject: Re: [pfSense Support] Re: Can't get more than 15kpps.
On Thu, Jul 30, 2009 at 8:21 PM, Chris Buechler c...@pfsense.org wrote:
On Thu, Jul 30, 2009 at 1:17 PM, Lennyfive2one.le...@gmail.com
Hi guys,
I know how sick of me you are by now, but I've had some developments here
and now I'm stuck again.
So, FINALLY I convinced the management to buy a new server. We bought an IBM
x3550 with 2 Quad Core CPUs E5420 and 2GB RAM PC2-5300 667MHz. Not just
that, we bought 2 of them ( we need the
I have forgotten and am too lazy to go through all my emails again to
read, but have you tried standard intel server NIC's for this?
Curtis LaMasters
http://www.curtis-lamasters.com
http://www.builtnetworks.com
On Wed, Jul 29, 2009 at 10:30 AM, Lennyfive2one.le...@gmail.com wrote:
Hi guys,
Oh yeah, sorry, forgot to mention.
The NIC is the same one: Dual Intel PCI-X.
Lenny.
Curtis LaMasters wrote:
I have forgotten and am too lazy to go through all my emails again to
read, but have you tried standard intel server NIC's for this?
Curtis LaMasters
Not sure what your talking about with top posting. I just replied to the list.
Curtis LaMasters
http://www.curtis-lamasters.com
http://www.builtnetworks.com
On Wed, Jul 29, 2009 at 11:07 AM, Eugen Leitleu...@leitl.org wrote:
On Wed, Jul 29, 2009 at 10:33:19AM -0500, Curtis LaMasters wrote:
On Wed, Jul 29, 2009 at 10:22 AM, Lennyfive2one.le...@gmail.com wrote:
Oh yeah, sorry, forgot to mention.
The NIC is the same one: Dual Intel PCI-X.
Have you ruled out your switches as bottleneck?
db
-
To unsubscribe,
On Wed, Jul 29, 2009 at 12:30 PM, Lennyfive2one.le...@gmail.com wrote:
By the way, I checked sysctl net.inet.ip.intr_queue_drops and it's 0.
On the interfaces I see that em0(outside) has 0 errors, but on the
em1(inside) there are 3666587/0.
6 of the CPUs(cores) are usually 100% idle, while the
I did it half a year ago, but yes, without the pfSense, Alteon was able
to deal with all the load.
David Burgess wrote:
On Wed, Jul 29, 2009 at 10:22 AM, Lennyfive2one.le...@gmail.com wrote:
Oh yeah, sorry, forgot to mention.
The NIC is the same one: Dual Intel PCI-X.
Have you
Just like I answered previously, without the pfSense Alteon was able to
handle the load without problems.
Cisco switch also didn't have any errors on the interface. Plus, I only
started to see errors, when the high load began and at that same time I
also saw some packet loss on the firewall.
Lenny,
Do you have commercial support on these box's?
Curtis LaMasters
http://www.curtis-lamasters.com
http://www.builtnetworks.com
On Wed, Jul 29, 2009 at 11:54 AM, Lennyfive2one.le...@gmail.com wrote:
Just like I answered previously, without the pfSense Alteon was able to
handle the load
On Wed, Jul 29, 2009 at 12:54 PM, Lennyfive2one.le...@gmail.com wrote:
Just like I answered previously, without the pfSense Alteon was able to
handle the load without problems.
Cisco switch also didn't have any errors on the interface. Plus, I only
started to see errors, when the high load
No, it's actually automatic on both sides.
But would it be the reason for the CPU to raise this high?
Should I input a higher value in kthreads of em driver?
Chris Buechler wrote:
On Wed, Jul 29, 2009 at 12:54 PM, Lennyfive2one.le...@gmail.com wrote:
Just like I answered previously,
Well, I believe, a standard support of NBD. Except for the NIC, which
was bought on ebay.
Curtis LaMasters wrote:
Lenny,
Do you have commercial support on these box's?
Curtis LaMasters
http://www.curtis-lamasters.com
http://www.builtnetworks.com
On Wed, Jul 29, 2009 at 11:54 AM,
Who is the guy that runs some of the Fox websites. I believe he is in
this area as far as PPS. Maybe he could shed some light though he may
be only available via the Forums.
Curtis LaMasters
http://www.curtis-lamasters.com
http://www.builtnetworks.com
On Wed, Jul 29, 2009 at 12:00 PM,
On Wed, Jul 29, 2009 at 1:59 PM, Lennyfive2one.le...@gmail.com wrote:
No, it's actually automatic on both sides.
But would it be the reason for the CPU to raise this high?
Should I input a higher value in kthreads of em driver?
Try to fix speed and duplex both on the switch and the pfsense
From: Lenny [mailto:five2one.le...@gmail.com]
Sent: July 29, 2009 11:31 AM
By the way, I checked sysctl net.inet.ip.intr_queue_drops and
it's 0.
On the interfaces I see that em0(outside) has 0 errors, but on
the em1(inside) there are 3666587/0.
6 of the
Evgeny Yurchenko wrote:
**
I would try to swap cables and interfaces in config and see
errors. Do they go to em0? stay on em1? It's pointless trying to
fix tcp/ip without eliminating problem on media.
Eugene.
That's all understandable when speaking of errors and
From: Lenny [mailto:five2one.le...@gmail.com]
Sent: July 29, 2009 3:38 PM
Evgeny Yurchenko wrote:
I would try to swap cables and interfaces in
config and see errors. Do they go to
On Wed, Jul 29, 2009 at 3:38 PM, Lennyfive2one.le...@gmail.com wrote:
That's all understandable when speaking of errors and packet loss, but would
it really cause the CPU hit 100% at 50kpps?
both em0 and em1?
By the way, it worked for 3 weeks with regular load (about 10kpps) and the
CPU was
Thanks for all the suggestions, guys.
Anyway, I found it very interesting that the new snapshots have yandex
driver in them, so I decided to try it.
Of course, as I don't have the new server yet, I had to try on my old IBM
x335.
But here are a couple of things that wouldn't let me try it:
with 2
On Wed, May 13, 2009 at 7:47 PM, Scott Ullrich sullr...@gmail.com wrote:
On Wed, May 13, 2009 at 8:36 PM, Dimitri Rodis
dimit...@integritasystems.com wrote:
My understanding is that Giant lock is gone from the FreeBSD network stack
in 8:
Hi again,
sorry to wake an old thread, but this is still an issue for me.
I was offered a Dell R200 server today, which comes with a single Xeon x3220
2.4GHz Quad Core CPU.
(I understand it's a repacked Q6600 or something).
I was wondering if this would be sufficient for my needs.
The better
Lenny wrote:
I was offered a Dell R200 server today, which comes with a single Xeon
x3220 2.4GHz Quad Core CPU.
(I understand it's a repacked Q6600 or something).
I was wondering if this would be sufficient for my needs.
I use a R200 at work for our pfSense, and we've had no issues with it.
Hi,
thanks for answering, but I guess you didn't read the whole thread. I don't
blame you, since it's a zillion posts:)
Anyway, I need it for a website, where I have about 300Mb traffic, which is
around 150kpps, and I need some CPU power to spare.
IBM x336 with dual Xeon 3.6GHz could only handle
On Wed, May 13, 2009 at 6:54 AM, Lenny five2one.le...@gmail.com wrote:
Hi again,
sorry to wake an old thread, but this is still an issue for me.
I was offered a Dell R200 server today, which comes with a single Xeon x3220
2.4GHz Quad Core CPU.
(I understand it's a repacked Q6600 or
Bill Marquette schrieb:
Ask the vendors for eval gear and make sure it supports the load before you
buy.
AFAIK, SUN still provides eval-systems for free.
I would evaluate one of the new X2270 with the Nehalem Xeons.
This should provide a 50% boost even on 5400-series Xeons.
Also, they
On Wed, May 13, 2009 at 10:21 AM, Rainer Duffner rai...@ultra-secure.de wrote:
AFAIK, SUN still provides eval-systems for free.
I would evaluate one of the new X2270 with the Nehalem Xeons.
This should provide a 50% boost even on 5400-series Xeons.
Also, they use Intel NICs, IIRC.
The
Rainer Duffner wrote:
AFAIK, SUN still provides eval-systems for free.
I would evaluate one of the new X2270 with the Nehalem Xeons.
This should provide a 50% boost even on 5400-series Xeons.
Also, they use Intel NICs, IIRC.
we've got a shiny Dell R710 with twin L5520 Xeon-EPs, very nice;
Sounds like over kill for pfsense! :D
Message sent from my iPhone
Bill McIlhargey Jr
COMPUTERONIX, LLC
978.500.5936
supp...@compute-ronix.com
www.compute-ronix.com
On May 13, 2009, at 11:24 AM, Paul Mansfield it-admin-pfse...@taptu.com
wrote:
Rainer Duffner wrote:
AFAIK, SUN still
On Wed, May 13, 2009 at 10:25 AM, Bill McIlhargey Jr b...@mcilhargey.com
wrote:
Sounds like over kill for pfsense! :D
Message sent from my iPhone
Bill McIlhargey Jr
COMPUTERONIX, LLC
978.500.5936
supp...@compute-ronix.com
www.compute-ronix.com
It's only overkill if you don't need the
...@gmail.com]
Sent: Wednesday, May 13, 2009 4:13 PM
To: support@pfsense.com
Subject: Re: [pfSense Support] Re: Can't get more than 15kpps.
On Wed, May 13, 2009 at 10:25 AM, Bill McIlhargey Jr b...@mcilhargey.com
wrote:
Sounds like over kill for pfsense! :D
Message sent from my iPhone
Bill
On Wed, May 13, 2009 at 8:36 PM, Dimitri Rodis
dimit...@integritasystems.com wrote:
My understanding is that Giant lock is gone from the FreeBSD network stack
in 8:
http://unix.derkeiler.com/Mailing-Lists/FreeBSD/arch/2009-04/msg00075.html
PF is still protected by one giant lock and does not
Chris Buechler wrote:
On Mon, Apr 13, 2009 at 6:13 AM, Lenny five2one.le...@gmail.com wrote:
Hi guys,
first of all, thanks for all the support!
Anyway, unfortunately, after all the hell I've been through with this, our
CEO is not interested in buying a new server:(
heh.. How about sorry,
*Subject:* Re: [pfSense Support] Re: Can't get more than 15kpps.
Well, actually, it's not the NICs that pushed me away from this server,
but the expansion slots.
I intend to insert my dual port Intel, and it's PCI-X, but the Sun only has
PCI-e, so it was no good.
Also, today, looking on ebay
Am 13.04.2009 um 12:13 schrieb Lenny:
Hi guys,
first of all, thanks for all the support!
Anyway, unfortunately, after all the hell I've been through with
this, our CEO is not interested in buying a new server:(
But let's put all the smart decisions aside as I have to figure
out what can I
Well,
I succeeded in installing m0n0wall before I saw the limitations of it.
Although I did have to use the IDE drive, and not the SCSI.
But would you say it would take care of the traffic I have?
OpenBSD scares me a bit:)
Regarding the iptables stuff, weird as it may sound - the CEO said that
On Mon, Apr 13, 2009 at 06:53, Lenny five2one.le...@gmail.com wrote:
OpenBSD scares me a bit:)
It shouldn't, really. The initial installer dialog is awful, but once
you get past it and get stuff running, it's about as smooth and
seamless as any good BSD setup. For that matter, neither pf nor
On Mon, Apr 13, 2009 at 6:13 AM, Lenny five2one.le...@gmail.com wrote:
Hi guys,
first of all, thanks for all the support!
Anyway, unfortunately, after all the hell I've been through with this, our
CEO is not interested in buying a new server:(
heh.. How about sorry, but there is no other
Lenny wrote:
Also, today, looking on ebay, I realized that it's not such an easy task
- to find a modern server with a dual core AMD (second generation) and
at least 1 PCI-X slot. The same is with Intel. And I already have 4
PCI-X cards, so I'd rather use them.
we've found the Tyan 5391
On Mon, Mar 23, 2009 at 9:26 AM, Vick Khera vi...@khera.org wrote:
On Mon, Mar 23, 2009 at 8:30 AM, Lenny five2one.le...@gmail.com wrote:
I got offered a Sun Fire X2200 with Opteron Dual Core 2210(that's 1.8GHz).
Will that do it? (for ~150kpps)
That's a little slower than what I use in prod
Well, actually, it's not the NICs that pushed me away from this server,
but the expansion slots.
I intend to insert my dual port Intel, and it's PCI-X, but the Sun only
has PCI-e, so it was no good.
Also, today, looking on ebay, I realized that it's not such an easy task
- to find a modern
On Mon, Mar 23, 2009 at 12:33 AM, Lenny five2one.le...@gmail.com wrote:
It's 530 (bytes?)
(and yet for 50kpps I had around 150Mb of traffic. Is this possible?)
http://www.ccievault.net/index.php/tools says it's possible
--Bill
Right, sorry, forgot to multiply by 8:)
Anyway, thank you for all the support,
I guess I should search for the new server then.
Lenny.
Bill Marquette wrote:
On Mon, Mar 23, 2009 at 12:33 AM, Lenny five2one.le...@gmail.com wrote:
It's 530 (bytes?)
(and yet for 50kpps I had around
I got offered a Sun Fire X2200 with Opteron Dual Core 2210(that's 1.8GHz).
Will that do it? (for ~150kpps)
Lenny.
On Mon, Mar 23, 2009 at 8:59 AM, Lenny five2one.le...@gmail.com wrote:
Right, sorry, forgot to multiply by 8:)
Anyway, thank you for all the support,
I guess I should
On Mon, Mar 23, 2009 at 8:30 AM, Lenny five2one.le...@gmail.com wrote:
I got offered a Sun Fire X2200 with Opteron Dual Core 2210(that's 1.8GHz).
Will that do it? (for ~150kpps)
Double check the NICs in that box. I believe they're broadcom and
nvidia (yes, Sun does a mix and match on the same
On Mon, Mar 23, 2009 at 8:30 AM, Lenny five2one.le...@gmail.com wrote:
I got offered a Sun Fire X2200 with Opteron Dual Core 2210(that's 1.8GHz).
Will that do it? (for ~150kpps)
Stick with boxes that feature EM (Intel) NICS.
Scott
On Sat, Mar 21, 2009 at 6:00 PM, Lenny five2one.le...@gmail.com wrote:
Hi Bill,
snip
Now, for the bad part. I got to a total of almost 50kpps, and that was via
70% CPU. Which probably means that at about 70kpps or so I'd hit 100%. Which
actually was a lot like what you said about Xeons (you
Hi Bill,
ok, thanks.
So as I understand it, in my production environment I will not be able
to get more than say 150-200kpps even if I had the best CPU available on
the market today? Which, by the way, equals around 450-600Mb in my case.
And that is for dual port NIC, of course.
Also, I
On Sun, Mar 22, 2009 at 3:32 PM, Lenny five2one.le...@gmail.com wrote:
Hi Bill,
ok, thanks.
So as I understand it, in my production environment I will not be able to
get more than say 150-200kpps even if I had the best CPU available on the
You should be able to hit much more than that.
BTW, whats your average packet size?
--Bill
-
To unsubscribe, e-mail: support-unsubscr...@pfsense.com
For additional commands, e-mail: support-h...@pfsense.com
Commercial support available - https://portal.pfsense.org
sorry, you got me there:) how do I check that?
Bill Marquette wrote:
BTW, whats your average packet size?
--Bill
-
To unsubscribe, e-mail: support-unsubscr...@pfsense.com
For additional commands, e-mail:
here you go.
Are those CPUs close to you old Xeons?
Copyright (c) 1992-2008 The FreeBSD Project.
Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994
The Regents of the University of California. All rights reserved.
FreeBSD is a registered trademark of The FreeBSD
I believe so. The newer Core designs have lower Ghz ratings. Any
chance you know the models? I'm not seeing the VTX feature in your
dmesg, which makes me think it's not a 5xxx series CPU (which would
get you more throughput).
--Bill
On Sun, Mar 22, 2009 at 4:18 PM, Lenny
On Sun, Mar 22, 2009 at 4:39 PM, Bill Marquette
bill.marque...@gmail.com wrote:
On Sun, Mar 22, 2009 at 4:13 PM, Lenny five2one.le...@gmail.com wrote:
sorry, you got me there:) how do I check that?
Bill Marquette wrote:
BTW, whats your average packet size?
Easiest way to get in the
On Sun, Mar 22, 2009 at 4:13 PM, Lenny five2one.le...@gmail.com wrote:
sorry, you got me there:) how do I check that?
Bill Marquette wrote:
BTW, whats your average packet size?
Easiest way to get in the ballpark should be to:
tcpdump -w /tmp/pps.pcap -i WAN -c 1
substitute WAN for
On Sun, Mar 22, 2009 at 5:33 PM, Bill Marquette
bill.marque...@gmail.com wrote:
I believe so. The newer Core designs have lower Ghz ratings. Any
chance you know the models? I'm not seeing the VTX feature in your
dmesg, which makes me think it's not a 5xxx series CPU (which would
get you
On Sun, Mar 22, 2009 at 5:20 PM, Chris Buechler c...@pfsense.org wrote:
On Sun, Mar 22, 2009 at 5:33 PM, Bill Marquette
bill.marque...@gmail.com wrote:
I believe so. The newer Core designs have lower Ghz ratings. Any
chance you know the models? I'm not seeing the VTX feature in your
dmesg,
It's 530 (bytes?)
(and yet for 50kpps I had around 150Mb of traffic. Is this possible?)
Bill Marquette wrote:
On Sun, Mar 22, 2009 at 4:39 PM, Bill Marquette
bill.marque...@gmail.com wrote:
On Sun, Mar 22, 2009 at 4:13 PM, Lenny five2one.le...@gmail.com wrote:
sorry, you got me
Hi Bill,
so I followed your advise and disabled the hyperthreading in the BIOS
(No APIC settings there).
So here's what's happening. First of all, I realized how dumb I am,
since I always looked only
on the outside of the wan interface when watching the throughput, so
all this time it
Hi,
ok, I'm back with some tests and results.
I read a lot about the em driver settings, and this is what I did:
in /etc/sysctl.conf I added:
dev.em.0.rx_processing_limit=1600
dev.em.1.rx_processing_limit=1600
although I also tried -1 and some smaller values.
in /boot/loader.conf I added:
Also, while searching the net for the emX taskq solution, I read that few
people are successfully running the modified em driver from Yandex.
This is their README:
RX queue is being processed w/more than one thread. Use sysctl
dev.em.X.rx_kthreads to alter number of threads.
TX interrupts has
You would have to build a kernel yourself without the em/ixgbe modules
to be able to use yandex driver.
Ever checked if you have MSI enabled on your motherboard and what
happens if you disable it?
On Wed, Mar 18, 2009 at 4:27 PM, five2one.le...@gmail.com wrote:
Also, while searching the net for
On Wed, Mar 18, 2009 at 7:32 AM, five2one.le...@gmail.com wrote:
Hi,
ok, I'm back with some tests and results.
I read a lot about the em driver settings, and this is what I did:
in /etc/sysctl.conf I added:
dev.em.0.rx_processing_limit=1600
dev.em.1.rx_processing_limit=1600
although I
On Wed, Mar 18, 2009 at 10:27 AM, five2one.le...@gmail.com wrote:
So the question is, should I go for it? Will it help me in any way? I mean,
if I have 2 Xeon CPUs and Hyper Threading enabled, I can actually divide it
into 4 threads, right?
Don't use hyperthreading. It's likely to cause you
Hi Bill,
thanks for answering.
a couple of questions:
I'm gonna disable hyperthreading tomorrow, but tell me, should I do it
in BIOS and just boot it up,
without any change(reinstall)? or should I use this method:
FreeBSD supports hyperthreading on Intel CPU's on the i386 platform.
Hi,
ok, thanks.
Regarding MSI - I never checked, but as far as I remember the BIOS
settings - I never saw it there.
I'll check tomorrow.
thanks,
Lenny.
Ermal Luçi wrote:
You would have to build a kernel yourself without the em/ixgbe modules
to be able to use yandex driver.
Ever
On Wed, Mar 18, 2009 at 3:12 PM, Lenny five2one.le...@gmail.com wrote:
Hi Bill,
thanks for answering.
a couple of questions:
I'm gonna disable hyperthreading tomorrow, but tell me, should I do it in
BIOS and just boot it up,
All our older Intel machines had it disabled in BIOS. The
1 - 100 of 120 matches
Mail list logo