Re: All transmits on txq0 on an ix interface - why no balancing?

2021-04-13 Thread Phil Rosenthal
> On Apr 13, 2021, at 8:22 PM, Kevin Bowling wrote: > > On Tue, Apr 13, 2021 at 5:11 PM Phil Rosenthal wrote: >> >>> On Apr 13, 2021, at 8:07 PM, Olivier Cochard-Labbé >>> wrote: >>> >>> Are you exiting through a tunnel interface (

Re: All transmits on txq0 on an ix interface - why no balancing?

2021-04-13 Thread Phil Rosenthal
> On Apr 13, 2021, at 8:07 PM, Olivier Cochard-Labbé > wrote: > > Are you exiting through a tunnel interface (GRE, GIF, PPPoE, IPsec, OpenVPN, > etc.) ? No. I am running PF/Altq for NAT and Traffic shaping. One port is using a Multirate 2.5G SFP+ (linked at 2.5G) to connect to a cablemodem

All transmits on txq0 on an ix interface - why no balancing?

2021-04-13 Thread Phil Rosenthal
Hello, I have spent a few hours trying to figure this out, and can't seem to make any progress. All transmits on my ix interfaces seem to use a single queue, txq0... which is resulting in occasional drops: sysctl -a|grep -i ix.0|grep -i txq_proc ; sysctl -a|grep -i ix.0|grep tx_packets dev.ix.0

L2TP/IPSec server on FreeBSD

2004-08-20 Thread Phil Rosenthal
Hello, I've been trying to set up racoon on FreeBSD to serve as a VPN Server that I can connect into using the OS X L2TP/IPSec client, or the Windows L2TP/IPSec client, preferably from dynamic ip's, preferably from NAT. The point would be to have secure internet access behind, say, open free h

Re: FXP Bundling on a STL2

2001-12-27 Thread Phil Rosenthal
dmesg says: fxp0: Microcode loaded, int_delay: 1000 usec bundle_max: 6 fxp1: Microcode loaded, int_delay: 1000 usec bundle_max: 6 But it only appears to do anything on the fxp1 (the external one) >From systat -vm: ~3000 interrupts ~500 interrupts from netstat -I fxp0 -in 1: input ~3000 packets

FXP Bundling on a STL2

2001-12-25 Thread Phil Rosenthal
I noticed that the FXP built in to my STL2 motherboards seem to get have no difference on 4.5-PRERELEASE wether bundling is on or off, but an external FXP PCI card does begin to have a difference when receiving ~1000 pps (about half the interrupts). Does the STL2 FXP not support bundling? --Phil

Re: Running out of sockets, TIME_WAIT state

2001-10-29 Thread Phil Rosenthal
Hm. can I use MSG_EOF with a sendfile() in the flags field? man sendfile seems to say no, but perhaps the man page hasnt been updated... --Phil To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-net" in the body of the message

Running out of sockets, TIME_WAIT state

2001-10-29 Thread Phil Rosenthal
I have the kernel configured with 65536 sockets. This system is very fast, connected to a very fast network, with a very fast storage system. It is handling about 5000 connections per second and is maxing out its 65536 sockets (because so many are sitting in TIME_WAIT status). This box is pushin