On 29.06.2013 13:50, Sami Halabi wrote:
I think I was misunderstood...
Here is the situation i want to handle:
My box is a router that handles several /24 behind.
One of my links (em0) is connected to a private network 192.168.0.1 is me,
my neighbour is 192.168.0.2.
I want to make that any
. You need both of source and
destination addresses get translated. Reverse NAT translates does,
well, reverse thing: it translates destination IP keeping source IP intact.
So, you just need setup two ipfw nat instances, one general and one reverse
and pass your packets through both instances.
Eugene
optimal
performance, but at least works out of the box?
Or, better, make nmbclusters auto-tuning smarter, if any.
I mean, use more nmbclusters for machines with large amounts of memory.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http
timeouts, please :-)
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd-net-unsubscr...@freebsd.org
On 03.06.2013 18:43, Andreas Nilsson wrote:
Hello,
Still trying out the tablearg functionality of ipfw and found the following:
1)
# ipfw table 100 add 192.168.0.0/24 10.0.0.1
# ipfw table 100 list
192.168.0.0/24 167772161
I guess it is correct, but not user friendly. Can't the
On 14.05.2013 08:08, Barney Cordoba wrote:
The defaults are broken.
Finally we managed to get to the point. Hallelujah!
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to
. I, as system administrator, expect from a
driver
to spit packets as fast as possible. I can employ some packet buffering system
if I need it
or use ng_car or dummynet or something to deal with bottlenecks. Or not to deal
with them for purpose.
NIC driver cannot know my needs.
Eugene Grosbein
On 10.05.2013 20:07, M. V. wrote:
hi,
I have a FreeBSD 8.2 server with some em igb interfaces. (em driver version
is 7.3.2 igb driver version is 2.3.1)
It doesn't have a full-load traffic on it (maximum 50-100Mbps per-interface
on 1Gbps interface). but occasionally one of its interfaces
On 10.05.2013 23:02, jbiskofski wrote:
Hey Everyone,
I just cant get this to work. I would appreciate any suggestions.
It seems GRE protocol is blocked in the direction from mpd5 server to client.
You should blame your provider. Or change VPN type from PPtP to L2TP that uses
UDP only,
not
the load on several CPU cores.
Therefore, I don't care of CS theory for this particular case.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd-net-unsubscr
running in igb's context.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd-net-unsubscr...@freebsd.org
On 09.05.2013 23:25, Barney Cordoba wrote:
Network device driver is not guilty here, that's just pf's
contention
running in igb's context.
Eugene Grosbein
They're both at play. Single threadedness aggravates subsystems that
have too many lock points.
It can also be solved with using
.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd-net-unsubscr...@freebsd.org
On 29.04.2013 16:24, Eugene Grosbein wrote:
Hi!
Today I discovered really subtle bug in connection of lagg(4), vlan(4) and
igb(4) drivers.
Here is what I have in /etc/rc.conf:
cloned_interfaces=lagg0 lagg1
ifconfig_lagg1=laggproto lacp laggport igb0 laggport igb1
12.04.2013 05:31, Karl Denninger пишет:
Is there a cookbook for setting this up? There are examples for
setting up a tunnel between two fixed-address networks (e.g. a remote
LAN that needs to be integrated with a central LAN over IPSec but I
can't find anything addressing the other situation
The following reply was made to PR kern/162558; it has been noted by GNATS.
From: Eugene Grosbein egrosb...@rdtc.ru
To: bug-follo...@freebsd.org
Cc: Luigi Rizzo ri...@iet.unipi.it
Subject: Re: kern/162558: [dummynet] [panic] seldom dummynet panics
Date: Tue, 22 Jan 2013 12:33:51 +0700
Hi
Hi!
Recently libradius(3) got long-awaited 'dead_time' option
that allows to skip 'dead' radius servers for 'dead_time' timeout
while dealing with multiple servers.
I'd like to ask for small improvement of the code.
Presently it will fail without a try if all servers marked 'dead'.
Instead, in
read ip(4) manual page on IP_PORTRANGE. You can choose one of three
ranges
for your socket (default range, low and high).
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send
is based on.
So, the warning applies to mpd's internal NAT too. I guess this should be a
culprit.
Can you exclude NAT to perform clean experiment? Does the issue disappear
when you do?
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http
22.10.2012 11:09, Nikolay Nevzorov пишет:
Any traffic throuhg NAT does not cause problems. And any routed traffic so on.
Problem only with traffic that generated on host with alc0, because host
generate packets much more bigger than MTU (about 2300 bytes per packet with
MTU 1500), a see it
offloading (TSO). Thus, to reliably nat your net-
work traffic, please disable TSO on your NICs using ifconfig(8).
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any
05.09.2012 14:08, Adrian Chadd пишет:
I'm so sorry for dropping off the radar here.
Can you please compile your kernel with KTR enabled so we can capture
some schedgraph traces?
Yes, I can and will. What should I do besides of rebuilding
the kernel with KTR and PREEMPTION?
one.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd-net-unsubscr...@freebsd.org
and done that before, during and after noted failure -
all counters are zero except of good frames conters (in/out).
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail
.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd-net-unsubscr...@freebsd.org
31.08.2012 22:54, Eugene Grosbein пишет:
I've rebuilt by kernel with SCHED_ULE and excluded PREEMPTION.
Stock driver works without changes in behaviour and driver from HEAD
now works very similar to old one: LA is not higher than 2 and
userland is pretty responsive. Also, transfer speed
, up to 4.5 for old one and up to 9+ for new).
I ran tests with same custom kernel, loading/unloading old/new drivers as
modules
without reboot. Schedules is default SCHED_ULE.
Another note: I run mpd/PPPoE/ng0 over vr1 and http transfer were through ng0.
Eugene Grosbein
In previous letter I've described my attempts to try vr(4) from HEAD.
Now I'd like to explain why I've tried it.
The problem is that stock vr(4) from 8.3-STABLE/i386 has serious issues for my
system.
I have home router with two vr interfaces, vr0 is for LAN (IPoE) and vr1 is for
WAN
31.08.2012 12:19, Eugene Grosbein пишет:
With HEAD driver, for same test LA pikes to 8 and higher and it takes up to
10 seconds
for userland applications like shell or screen(1) to respond to physical
console events:
last pid: 1335; load averages: 8.27, 4.05, 2.04up 0+00:14
07.08.2012 23:32, Marko Zec пишет:
On Tuesday 07 August 2012 18:20:28 Eugene Grosbein wrote:
On Tue, Aug 07, 2012 at 04:33:11PM +0200, Marko Zec wrote:
There's no need to use epair at all here. Just create a vnet jail, and
assign one of the igb interfaces into it, done:
# jail -c name=x
09.08.2012 11:54, Bjoern A. Zeeb пишет:
Modern jails are cool but their manual page is not ;-)
Depends, / is not a fairly complete file system install of FreeBSD in
your world view? ;-) [ which is what no path argument implies ]
First I thought no path argument emplies some kind of empty
Hi!
This is still the problem for 9.1-PRERELEASE: ng_source is unusable for 64bit
systems.
The patch really helps and fixes the problem for 64bit (and breaks for 32bit :-)
Perhaps, it may be improved with some #ifdef's
Eugene Grosbein
___
freebsd-net
The following reply was made to PR kern/120304; it has been noted by GNATS.
From: Eugene Grosbein egrosb...@rdtc.ru
To: bug-follo...@freebsd.org, n...@freebsd.org n...@freebsd.org
Cc:
Subject: Re: kern/120304: [netgraph] [patch] netgraph source assumes 32-bit
timeval on AMD64
Date: Tue, 07 Aug
netstat -I to see how many packets got through).
I would really like using normal TCP/UDP/ICMP streams in addition of ng_source
tests.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
of NETGRAPH mesh.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd-net-unsubscr...@freebsd.org
to state that most of us do NOT
need scientifically exact ipfw counters values when pushing hardware to its
maximum.
Personaly, I'd like to have tunable that gives me another 15% of speed
at cost of bad ipfw counters I don't use anyway.
Eugene Grosbein
16.07.2012 07:13, Bjoern A. Zeeb пишет:
On 15. Jul 2012, at 20:54 , Mike Tancsa wrote:
On 7/10/2012 2:24 AM, Przemyslaw Frasunek wrote:
It seems, Przemyslaw Frasunek uses proxyarp?
I have no such problems but I do not use proxyarp.
Could you get rid of it, Przemyslaw?
No, I don't use
ipencap packets
without ipfw nat processing and the problem has disappeared.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd-net-unsubscr...@freebsd.org
at receiving gif
interface,
and only them.
Note, there is no such problem for ICMP fragmented packets in this setup,
IPIP fragments are affected only.
Please help.
Eugene Grosbein.
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org
09.07.2012 16:49, Eugene Grosbein пишет:
Hi!
For long time I suffer from RELENG_8's inability to reassemble several types
of incoming fragmented packets when they arrive of-of-order.
For example: my gif(4) interface has MTU=1500, so it gets fragmented
when it goes out to the Internet
, the in6 case was even worse, as the in6 equivalent of
arptimer is significantly more complicated and likes to do crazy
things like dropping locks.
It seems, Przemyslaw Frasunek uses proxyarp?
I have no such problems but I do not use proxyarp.
Could you get rid of it, Przemyslaw?
Eugene Grosbein
0x80502ab2 at amd64_syscall+0x302
#12 0x804ea8cc at Xfast_syscall+0xfc
Uptime: 17d7h18m38s
Did you set net.isr.direct=0 (and/or direct_force)?
If so, don't do that. Get back to default 1 for these two sysctls.
Eugene Grosbein
___
freebsd
is required.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd-net-unsubscr...@freebsd.org
- 1.3.6.1.2.1.25.3.3.1.2
And without -n flag, please :-)
I'd like it to resolve OIDs to their names.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd-net-unsubscr
17.06.2012 01:13, Arnaud Lacombe пишет:
Hi,
On Fri, Jun 15, 2012 at 7:50 AM, Eugene Grosbein egrosb...@rdtc.ru wrote:
15.06.2012 18:33, Przemyslaw Frasunek пишет:
Dear All,
unfortunately, one of my mpd5 PPPoE access servers started panicing every
few
hours.
I'm running recent 8.3
? It matters too.
For systems having USB keyboard there is another patch needed to obtain
crashdumps (by Andriy Gapon):
http://www.kuzbass.ru/freebsd/patches/stop_scheduler_on_panic.usb.diff
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http
15.06.2012 18:50, Eugene Grosbein пишет:
unfortunately, one of my mpd5 PPPoE access servers started panicing every few
hours.
I'm running recent 8.3-STABLE (as of 23th May) with WITNESS, INVARIANTS and
DEBUG_MEMGUARD compiled. Unfortunately, I'm unable to catch crashdump. For
some
reason
Hi!
How do I make FreeBSD 8-based router/NAT/security gateway
first perform NAT for outgoing packets then apply IPSEC transport mode
for plain TCP traffic?
Presently, locally originated packets are encrypted just fine
but routed and NAT-ed packet go out unencrypted.
I use ipfw nat.
Eugene
15.06.2012 03:21, Michael Sierchio пишет:
On Thu, Jun 14, 2012 at 9:42 AM, Eugene Grosbein egrosb...@rdtc.ru wrote:
How do I make FreeBSD 8-based router/NAT/security gateway
first perform NAT for outgoing packets then apply IPSEC transport mode
for plain TCP traffic?
Forgive me, but I
the OS, IIRC.
Basically, he wants tunnel mode to create full-blown network interface
without overhead for extra gre/gif packet header.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
this with FreeBSD 8.2-STABLE/amd64,
em(4) 7.3.2, Intel PRO/1000 EB.
There is a workaround, though: ifconfig em0 -rxcsum -txcsum mtu 9000
work from the beginning.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman
Hi!
It seems, I350 adapters are supported with igb(4) driver in 9.0-RELEASE
out-of-the-box.
However, not Hardware Release Notes nor igb(4) manual page have a word about
that.
Why? Isn't the driver ready for I350?
Eugene Grosbein
___
freebsd-net
Hi!
What is current status of IPSec offload support
for Intel 82576EB Gigabit Ethernet controller in igb(4) driver?
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any
fixed.
I think here is only one problem. problem to 'RUN FAST FSCK'
No need to, already.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd-net
10.04.2012 01:36, Коньков Евгений пишет:
Hi.
Apr 9 19:51:58 fsck: /dev/ad8s1e: UNEXPECTED INCONSISTENCY, CANNOT RUN FAST
FSCK
Apr 9 19:51:58 fsck:
Apr 9 19:51:58 fsck:
Apr 9 19:51:58 fsck: /dev/ad8s1e: UNEXPECTED INCONSISTENCY; RUN fsck
MANUALLY.
Apr 9 19:51:58 fsck: /dev/ad8s1e:
No.
network_interfaces is basically historic rudiment
used in 2.2.x FreeBSD version and alike.
In general, you should not use it in modern version at all.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
(that survives reboots)
for an interface in rc.conf is:
ifconfig_em0=dhcp ?
Yes, thats what man rc.conf says.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd
. Is this a problem?
It seems you have lots of non-TCP and non-UDP traffic
like PPPoE or GRE. igb(4) uses Microsoft Receive Side Scaling,
so it cannot compute a hash for such ethernet frames
and they all go to queue zero. That is unavoidable with those cards.
Eugene Grosbein
08.03.2012 03:29, YongHyeon PYUN пишет:
On Tue, Mar 06, 2012 at 07:54:02PM +0700, Eugene Grosbein wrote:
Hi!
Yesterday I've updated old HP ProLiant DL360 G4p to 8.3-PRELELEASE/amd64
running busy icecast2 server in hope it can saturate 1G bge(4) link.
This server has PCI-X connected HP
show 540Mbps on bge(4)?
I'll try to run benchmarks if I find a peer for netperf.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd-net-unsubscr
dev.bge.0.stats.rx.FCSErrors: 341
^
You have multiple FCS and Input errors. Check signal
quality(i.e. UTP cable).
OTOH, should not bge(4) for such errors increase Ierrs counters
that shows netstat -i?
Eugene Grosbein
the patch. Now dmesg shows:
bge0: HP NC7782 Gigabit Server Adapter, ASIC rev. 0x002100 mem
0xfdf7-0xfdf7 irq 25 at device 2.0 on pci2
bge0: CHIP ID 0x2100; ASIC REV 0x02; CHIP REV 0x21; PCI-X 66 MHz
I'll watch on Ierrs counter.
Eugene Grosbein
09.03.2012 02:06, YongHyeon PYUN пишет:
On Wed, Mar 07, 2012 at 11:42:09PM +0700, Eugene Grosbein wrote:
08.03.2012 10:43, YongHyeon PYUN пишет:
Show me the output of sysctl dev.bge.0.stats.
# sysctl dev.bge.0.stats
dev.bge.0.stats.FramesDroppedDueToFilters: 0
dev.bge.0
with some extra buffering at clients side. Drops make TCP flow
control
think the channel is overloaded when it's not. And many TCP peers do not use
SACK.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo
em/igb have.
Should I try to play with constants in the code and if so, what are limits of
this chip?
Or it will never be capable of utilizing full gigabit speed?
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman
25.02.2012 02:00, Andrew Thompson пишет:
2012/2/25 Eugene Grosbein egrosb...@rdtc.ru:
25.02.2012 00:14, Eugene Grosbein пишет:
This problem occurs only when net.isr.direct=0/net.isr.direct_force=0.
And only when lagg1 has both ports up and running. And when I use oversized
pings
25.02.2012 19:17, Eugene Grosbein пишет:
25.02.2012 02:00, Andrew Thompson пишет:
2012/2/25 Eugene Grosbein egrosb...@rdtc.ru:
25.02.2012 00:14, Eugene Grosbein пишет:
This problem occurs only when net.isr.direct=0/net.isr.direct_force=0.
And only when lagg1 has both ports up and running
fine,
no packet loss at all.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd-net-unsubscr...@freebsd.org
25.02.2012 00:14, Eugene Grosbein пишет:
Hi!
I've found that my PPPoE BRAS server gives 80% packet loss
for fragmented pings against itself while no packet loss for transit
fragmented pings same time.
The problem vanishes if I do ONE of following:
1) disable indirect netisr mode, e.g
The following reply was made to PR kern/162558; it has been noted by GNATS.
From: Eugene Grosbein egrosb...@rdtc.ru
To: bug-follo...@freebsd.org
Cc:
Subject: Re: kern/162558: [dummynet] [panic] seldom dummynet panics
Date: Wed, 15 Feb 2012 19:25:43 +0700
Hi!
The source of this problem
for such misbehaviours. That had fixed same problem for me.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd-net-unsubscr...@freebsd.org
The following reply was made to PR kern/161899; it has been noted by GNATS.
From: Eugene Grosbein egrosb...@rdtc.ru
To: Steven Hartland kill...@multiplay.co.uk
Cc: freebsd-net@freebsd.org, bug-follo...@freebsd.org
Subject: Re: kern/161899: [route] ntpd(8): Repeating RTM_MISS packets causing
protocol (number 94 decimal)?
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd-net-unsubscr...@freebsd.org
, you gif is encrypted with ipsec? That changes ip protocol numbers.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd-net-unsubscr...@freebsd.org
or higher with sysctl command,
until this ngctl error disappears.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd-net-unsubscr...@freebsd.org
mpd2 mpd: PPTP: NgMkSockNode: No buffer space available
You should bot increase these setting without increase
of kern.ipc.maxsockbuf. First greatly increase kern.ipc.maxsockbuf
(I use 80M for it) and only then increase net.graph.recvspace
and net.graph.maxdgram.
Eugene Grosbein
27.12.2011 20:11, per...@pluto.rain.com пишет:
Jason Hellenthal jh...@dataix.net wrote:
See siftr(4). This module writes to a file.
Is siftr(4) new since 8.1?
$ man siftr
No manual entry for siftr
$ cd /usr/ports
$ ls -d */*siftr*
ls: */*siftr*: No such file or directory
by default?
Do you use NAT? man ipfw clearly states:
ipfw nat is not compatible with
the TCP segmentation offloading (TSO). Thus, to reliably nat your net-
work traffic, please disable TSO on your NICs using ifconfig(8)
Eugene Grosbein
to bring link down with ifconfig down command:
http://unix.derkeiler.com/Mailing-Lists/FreeBSD/net/2010-12/msg00058.html
http://unix.derkeiler.com/Mailing-Lists/FreeBSD/net/2010-12/msg00081.html
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
, use ipfw nat instead - it uses the same libalias
but completely in kernel and avoids gigantic natd overhead.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail
? I use 8.2-STABLE/amd64 and
quagga-0.99.20 from ports.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd-net-unsubscr...@freebsd.org
Hi!
How do I give out dummynet's CPU usage via SNMP?
Preferable as ever-incrementing raw counter.
I already use bsnmpd and its plugin bsnmp-ucd.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd
Hi!
How do I give out dummynet's CPU usage via SNMP?
Preferable as ever-incrementing raw counter.
I already use bsnmpd and its plugin bsnmp-ucd.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd
.
In the meantime, please try the following fix and let me know if it also work
for you.
http://people.freebsd.org/~qingli/in.c.diff
Thanks,
-- Qing
Thank you for quick responce. This patch works, no more panics and proxyarp
works too.
Eugene Grosbein
Thanks! I'm using 7.4 I will try it with 8.2
And, how the load sharing occurs between the queues?
The card uses more than one interrupt vector while notifying CPUs on incoming
traffic,
allowing operating system driver to bind distinct NIC interrupts to distinct
CPUs.
The card
, however, there are many requirements for multiple queues to
work:
you need MSI-X capable motherboard, your traffic needs to be TCP/UDP inside
plain IP.
No load distribution for PPPoE/GRE/etc. in this hardware (blame Microsoft RSS
specs).
Eugene Grosbein
14.09.2011 15:33, Vladimir Budnev пишет:
So i think there are at least to questions:
1. Have anyone ever met such situation? Or may be something close to
this one with 'hidden' ipfw rules?
Have you tried ipfw -d -e show?
Eugene Grosbein
03.08.2011 19:21, Bjoern A. Zeeb пишет:
Hi,
I have built IPv6-only snapshots for 9.0-BETA1 for everyone interested.
What does it mean, IPv6-only?
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo
interface got removed, then its
traffic
goes out various system queues (netisr, dummynet etc.) and another kind of panic
occurs due to INVARIANTS' references to non-existent ifp.
Please help.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http
11.07.2011 18:45, Vlad Galu пишет:
On Jul 11, 2011, at 1:42 PM, Eugene Grosbein wrote:
Hi!
My FreeBSD 8.2/amd64 routers use dummynet heavily
and keep panic with the *same* KDB backtrace:
dummynet: bad switch -256!
Forgot to mention that I use io_fast dummynet mode
and have increased
disable ISR parallelism for my mpd routers using:
net.isr.direct=1
net.isr.direct_force=1
At the other hand, there are other queues where traffic got delayed, not ISR
only.
Dummynet itself is an example. The router still panices with INVARIANTS too
often.
Eugene Grosbein
-- Closed
Jul 5 13:07:09 lns mpd: [L28-6225] LCP: Down event
Jul 5 13:07:09 lns mpd: [L28-6225] LCP: state change Closed -- Initial
Jul 5 13:07:09 lns mpd: [L28-6225] Link: SHUTDOWN event
Jul 5 13:07:09 lns mpd: [L28-6225] Link: Shutdown
Eugene Grosbein
)
Is this enough, or I need to modify something else ?
It seems, enough. But, are you sure your L2TP client will wait
for overloaded daemon to complete connection? The change will proportionally
increase
responsiveness of mpd - it has not enough CPU horsepower to process requests
timely.
Eugene
timely.
Eugene Grosbein
Actually something else is happening.
I increased the queue in msg.c
#define MSG_QUEUE_LEN 65536
You can't do this blindly, without other changes.
For example, there is MSG_QUEUE_MASK in the next line
that must be equal to MSG_QUEUE_LEN-1 and effectively
limits
already seen.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd-net-unsubscr...@freebsd.org
that in the
log,
as you have already seen.
Eugene Grosbein
___
freebsd-net@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-net
To unsubscribe, send any mail to freebsd-net-unsubscr...@freebsd.org
The following reply was made to PR kern/153497; it has been noted by GNATS.
From: Eugene Grosbein egrosb...@rdtc.ru
To: bug-follo...@freebsd.org
Cc: mike tancsa m...@sentex.net
Subject: Re: kern/153497: [netgraph] netgraph panic due to race conditions
Date: Tue, 10 May 2011 03:15:52 +0700
Hi
The following reply was made to PR kern/153671; it has been noted by GNATS.
From: Eugene Grosbein egrosb...@rdtc.ru
To: bug-follo...@freebsd.org
Cc:
Subject: Re: kern/153671: [em] [panic] 8.2-PRERELEASE repeatable kernel in
if_em(4)
Date: Fri, 29 Apr 2011 15:17:55 +0700
Hi!
All needed
The following reply was made to PR kern/153255; it has been noted by GNATS.
From: Eugene Grosbein egrosb...@rdtc.ru
To: bug-follo...@freebsd.org
Cc:
Subject: Re: kern/153255: [panic] 8.2-PRERELEASE repeatable kernel panic under
heavy network load
Date: Fri, 29 Apr 2011 15:32:37 +0700
Hi
(the box
I'm trying to improve performance on is on 8.2-RELEASE-p1).
Anybody manage to make this card push above 600mbps in ideal network
benchmarks? Any help would be gladly appreciated.
Please show output of vmstat -z|egrep 'ITEM|mbuf' command after the test.
Eugene Grosbein
401 - 500 of 756 matches
Mail list logo