Re: Diagnosing terrible ixl performance
On 20/4/18 12:03 pm, Garrett Wollman wrote: I'm commissioning a new NFS server with an Intel dual-40G XL710 interface, running 11.1. I have a few other servers with this adapter, although not running 40G, and they work fine so long as you disable TSO. This one ... not so much. On the receive side, it gets about 600 Mbit/s with lots of retransmits. On the *sending* side, though, it's not even able to sustain 10 Mbit/s -- but there's no evidence of retransmissions, it's just sending really really slowly. (Other machines with XL710 adapters are able to sustain full 10G.) There is no evidence of any errors on either the adapter or the switch it's connected to. So far, I've tried: - Using the latest Intel driver (no change) - Using the latest Intel firmware (breaks the adapter) - Disabling performance tweaks in loader.conf and sysctl.conf - Changing congestion-control algorithms replacing card? obvious but you don't list it... Anyone have suggestions while I still have time to test this? (My plan B is to fall back to an X520 card that I have in my spares kit, because I *know* those work great with no faffing about.) Any relevant MIBs to inspect? The test I'm doing here is simple iperf over TCP, with MTU 9120. It takes about 10 seconds for the sending side to complete, but buffers are severely constipated for 20 seconds after that (delaying all traffic, including ssh connections). I'm at the point of trying different switch ports just to eliminate that as a possibility. that is also a possibility, as well as swap cables with a good machine... -GAWollman ___ freebsd-net@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-net To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org" ___ freebsd-net@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-net To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org"
Re: Bad outbound performance with Intel 6205 [Taylor Peak] WiFi (IWN)
On 04/20/18 06:37, Kevin Oberman wrote: > I am seeing poor performance on my WiFi that seems to have deteriorated > since FreeBSD V11.0. I see an average of about 35 errors for every packet > transmitted. > wlan0 1500 a0:88:b4:c6:ad:28 14501195 0 0 367667 > 12886354 0 > > I have always seen lot of errors on transmission on this interface, but I > have recently noticed that uploads were very slow. They seem to max out at > about 20 Mbps (bits, not bytes). For "normal" traffic (as opposed to test > traffic) it is often much worse, often topping out at around 2 or 3 Mbps. > > No input errors at all and god download performance. > > Are others seeing this? Any suggestions for troubleshooting? I don't see > much in the way of statistics or diagnostic information on iwn. the man > page gives me no clues.> iwn0@pci0:3:0:0:class=0x028000 card=0xc2208086 chip=0x00858086 rev=0x96 hdr=0x00 vendor = 'Intel Corporation' device = 'Centrino Advanced-N 6205 [Taylor Peak]' class = network iwn0: mem 0xf0c0-0xf0c01fff at device 0.0 on pci2 NameMtu Network Address Ipkts Ierrs Idrop IbytesOpkts Oerrs Obytes Coll wlan0 1500 84:3a:4b:cb:aa:c0 3914486 2 0 5523440357 7685 2087124 888555 0 The same 6205, I didn't notice bad upload performance though. Numbers of Ipkts and Oerrs are close enough and Opkts is too small compared to Ipkts, so I believe the driver misbehaves a bit and erroneously shows number of output packets as Oerrs. The Obytes also seems to be too small to be correct. A few months ago the association with AP starts sometimes to stall with no traffic in/out at all, reconfig in wpa_cli helps then. Looks like it correlates with upgrade from 10-stable to 11, but I'm not so sure. > Kevin Oberman, Part time kid herder and retired Network Engineer > E-mail: rkober...@gmail.com > PGP Fingerprint: D03FB98AFA78E3B78C1694B318AB39EF1B055683 > ___ > freebsd-net@freebsd.org mailing list > https://lists.freebsd.org/mailman/listinfo/freebsd-net > To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org" > -- Hrant Dadivanyan (aka Ran d'Adi)hrant(at)dadivanyan.net /* "Feci quod potui, faciant meliora potentes." */ ran(at)psg.com signature.asc Description: OpenPGP digital signature
Diagnosing terrible ixl performance
I'm commissioning a new NFS server with an Intel dual-40G XL710 interface, running 11.1. I have a few other servers with this adapter, although not running 40G, and they work fine so long as you disable TSO. This one ... not so much. On the receive side, it gets about 600 Mbit/s with lots of retransmits. On the *sending* side, though, it's not even able to sustain 10 Mbit/s -- but there's no evidence of retransmissions, it's just sending really really slowly. (Other machines with XL710 adapters are able to sustain full 10G.) There is no evidence of any errors on either the adapter or the switch it's connected to. So far, I've tried: - Using the latest Intel driver (no change) - Using the latest Intel firmware (breaks the adapter) - Disabling performance tweaks in loader.conf and sysctl.conf - Changing congestion-control algorithms Anyone have suggestions while I still have time to test this? (My plan B is to fall back to an X520 card that I have in my spares kit, because I *know* those work great with no faffing about.) Any relevant MIBs to inspect? The test I'm doing here is simple iperf over TCP, with MTU 9120. It takes about 10 seconds for the sending side to complete, but buffers are severely constipated for 20 seconds after that (delaying all traffic, including ssh connections). I'm at the point of trying different switch ports just to eliminate that as a possibility. -GAWollman ___ freebsd-net@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-net To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org"
Bad outbound performance with Intel 6205 [Taylor Peak] WiFi (IWN)
I am seeing poor performance on my WiFi that seems to have deteriorated since FreeBSD V11.0. I see an average of about 35 errors for every packet transmitted. wlan0 1500 a0:88:b4:c6:ad:28 14501195 0 0 367667 12886354 0 I have always seen lot of errors on transmission on this interface, but I have recently noticed that uploads were very slow. They seem to max out at about 20 Mbps (bits, not bytes). For "normal" traffic (as opposed to test traffic) it is often much worse, often topping out at around 2 or 3 Mbps. No input errors at all and god download performance. Are others seeing this? Any suggestions for troubleshooting? I don't see much in the way of statistics or diagnostic information on iwn. the man page gives me no clues. Kevin Oberman, Part time kid herder and retired Network Engineer E-mail: rkober...@gmail.com PGP Fingerprint: D03FB98AFA78E3B78C1694B318AB39EF1B055683 ___ freebsd-net@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-net To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org"
Re: max devices in tun
I'm using FreeBSD 11.1 current in a qemu instance. Today I tried it it said no space left on device. So I can say that it happened for some other number. (not 32768) Actually I tried other numbers too. and today it also created a character special device in /dev/tun for some huge number which wasn't shown in ifconfig. Sorry for any lack of info. I'll try to gather more from the bash history if I can. ___ freebsd-net@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-net To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org"
[Bug 218579] Wake on Lan doesn't work for bge NIC driver
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=218579 --- Comment #9 from Cy Schubert --- (In reply to Koen Martens from comment #8) That bug has been fixed in the latest patch I posted here. No worries about drawing too much current when powered off any more. I've been using it and previous versions of this patch on my laptop (only machine I have with bge) for about a year. I suppose I should submit the patch in phabricator for review prior to commit. -- You are receiving this mail because: You are on the CC list for the bug. ___ freebsd-net@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-net To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org"
[Bug 218579] Wake on Lan doesn't work for bge NIC driver
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=218579 --- Comment #8 from Koen Martens --- (In reply to Cy Schubert from comment #6) Hi, thanks for the reply (and the reworked patch). I guess the patch won't go upstream because of the reasons mentioned (ie. potential to blow up boards that can't support enough power when the system is powered down)? I'm also surprised it did work for me. I actually woke up the machine with wake-on-lan to ssh into it and upgrade it to 11.1-RELEASE. -- You are receiving this mail because: You are on the CC list for the bug. ___ freebsd-net@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-net To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org"
Re: svn commit: r332645 - head/sbin/ifconfig
On Tue, Apr 17, 2018 at 12:54:59PM +, Andrew Gallatin wrote: > New Revision: 332645 > URL: https://svnweb.freebsd.org/changeset/base/332645 > > Log: > Make lagg creation more fault tolerant > > - Warn, don't exit, when SIOCSLAGGPORT returns an error. > > When we exit with an error during lagg creation, a single > failed NIC (which no longer attaches) can prevent lagg > creation and other configuration, such as adding an IPv4 > address, and thus leave a machine unreachable. Hey Andrew, Since you're on lagg(4) a bit, could you perhaps have a look at this thread from last year on -net@: https://lists.freebsd.org/pipermail/freebsd-net/2017-June/048284.html I've been asked by several people that were using WiFi-CopperNet failover which worked fine in FreeBSD 10.x but was broken since then. I'm not an src committer nor a network stack expert in this area nor I'm using lagg(4), but perhaps you could expedite on handling this or at least point us to a more appropriate channel (due to lack of interest on -net@). ./danfe ___ freebsd-net@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-net To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org"
Re: max devices in tun
19.04.2018 7:29, Akilan Elango wrote: > Hey guys, > I was fiddling around with the /dev/tun to write a TUN interface > for my app. I wanted to check what is the max amount of interfaces > that can be made. Turned out it was 32768 [0 - 32768). But when I > created (ifconfig tun32768 create), a weird device appeared as shown > in this figure : https://imgur.com/a/0ypbwgg. > > Is this expected? and I'm unable to delete the device too from ifconfig. You should tell what FreeBSD version do you use. It does not do so for my 11.1/i386: # ifconfig tun32768 create ifconfig: SIOCIFCREATE2: No space left on device ___ freebsd-net@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-net To unsubscribe, send any mail to "freebsd-net-unsubscr...@freebsd.org"
