1. I have a realtek network card and am using a cable modem router. Does anyone know if fixes for problems with these (see below) have been backported from HEAD to RELENG_7? 2. Is this a good time to upgrade to RELENG_7? 3. I read a post last week that advised that 7.1 is a long way off. Appreciate any advice on use of ULE 3.0 and ZFS prior to 7.1 and a comment on what is holding up 7.1. Thanks>From FreeBSD 7.0 errata notes:[20080229] Instances of packet corruption and instability have been observed with the re(4) network driver. Some users have reported that using a newer version of this driver (on HEAD) has solved their problems, or at least mitigated them. [20080229] A change in the way that FreeBSD sends TCP options has been reported to cause odd interactions with some cable modem routers. While this issue is still under investigation, a change has been committed to HEAD that returns the option processing to that of FreeBSD 6. So far, this change has shown some promising results. Send instant messages to your online friends http://uk.messenger.yahoo.com
On May 11, 2008, at 9:34 PM, John Daniels wrote:> 1. I have a realtek network card and am using a cable modem > router. Does anyone know if fixes for problems with these (see > below) have been backported from HEAD to RELENG_7?The most useful thing to do would be to test the fixes out yourself and report back whether they help, assuming you encounter problems.> 2. Is this a good time to upgrade to RELENG_7?On a test machine, sure. It's been a long time since I've run a production environment off of a .0 release at all, much less without testing it first. :-)> 3. I read a post last week that advised that 7.1 is a long way off. > Appreciate any advice on use of ULE 3.0 and ZFS prior to 7.1 and a > comment on what is holding up 7.1.Respectively: ULE seems to be a solid improvement. ZFS is still beta. Nothing...be patient, or make more active contributions to help move things forwards. -- -Chuck
On Mon, May 12, 2008 at 05:34:47AM +0100, John Daniels wrote:> 1. I have a realtek network card and am using a cable modem router. > Does anyone know if fixes for problems with these (see below) have > been backported from HEAD to RELENG_7?I'd recommend downloading a RELENG_7 bootonly image, or possibly the livefs image (which will give you utilities and a working read-only filesystem), and see how your hardware looks on that. Meaning, if you have a test scenario which can be easily reproduced and can be done from livefs, give that a try. RELENG_7 snapshots (meaning 7.0-STABLE vs. 7.0-RELEASE) are here: ftp://ftp.freebsd.org/pub/FreeBSD/snapshots/200804> 2. Is this a good time to upgrade to RELENG_7?That's entirely up to you. RELENG_6 will be decomissioned quicker than RELENG_5 was, from what I've read, so testing compatibility now would be a good idea.> 3. I read a post last week that advised that 7.1 is a long way off. > Appreciate any advice on use of ULE 3.0 and ZFS prior to 7.1 and a > comment on what is holding up 7.1.Regarding ULE 3.0, I think you're referring to the ULE that's actually in RELENG_7 presently, which actually works (compared to the ULE that was in RELENG_5 and RELENG_6, which had major bugs). Or are you referring to an even newer scheduler (something -CURRENT might have)? ZFS is in a state of... well... my opinion is: use it at your own risk. The "experimental!" warning is very applicable. There are known problems with ZFS when performing lots of I/O between a ZFS pool and a UFS fs.> [20080229] Instances of packet corruption and instability have been observed with the re(4) network driver. Some users have reported that using a newer version of this driver (on HEAD) has solved their problems, or at least mitigated them.I believe this has been fixed, based on commits I've seen to the driver. Pyun YongHyeon would be able to confirm/deny this.> [20080229] A change in the way that FreeBSD sends TCP options has been reported to cause odd interactions with some cable modem routers. While this issue is still under investigation, a change has been committed to HEAD that returns the option processing to that of FreeBSD 6. So far, this change has shown some promising results.Do you have more details on this? Open PR, or a thread I can read? I'm thinking it might be related to the net.inet.tcp.rfc1323 sysctl, but I'm not sure. -- | Jeremy Chadwick jdc at parodius.com | | Parodius Networking http://www.parodius.com/ | | UNIX Systems Administrator Mountain View, CA, USA | | Making life hard for others since 1977. PGP: 4BD6C0CB |
On Mon, 12 May 2008 05:34:47 +0100 (BST) John Daniels <daniels_intrepid@yahoo.ie> wrote:> 1. I have a realtek network card and am using a cable modem > router. Does anyone know if fixes for problems with these (see > below) have been backported from HEAD to RELENG_7?FWIW, I had major troubles with re(4) around 7.0-release, and a while later (I had to use patches). After upgrading to 7-stable on 2008-04-12, re(4) is working for me without patches. HTH -- Regards, Torfinn Ingolfsen
Hi all,
I only just now subscribed, so my apologies if this has been covered
and my cursory search missed it.
I am seeing a similar issue to the one described in the email below
(pasted from a web listing of this list).
That is, my ssh connection will occasionally be dropped with the
message
Disconnecting: Bad packet length 3190815048.
(the number is not always the same).
I can reliably reproduce this with `cat < /dev/urandom | od -c`
from a FreeBSD-current box (a few weeks old, actually, since
the build was broken when I last tried to update) as well
as from a mac running OS X 10.4 .
The system that is killing my connections is:
FreeBSD periphrasis.mit.edu 7.0-STABLE FreeBSD 7.0-STABLE #3: Tue Mar
25 14:53:51 EDT 2008
kaduk@periphrasis.mit.edu:/usr/obj/usr/src/sys/PERIPHRASIS amd64
Any thoughts would be appreciated.
-Ben Kaduk
------------------ random message from this thread, for some context
-------------
* From: Jeremy Chadwick <koitsu@xxxxxxxxxxx>
* Date: Mon, 12 May 2008 05:32:07 -0700
On Mon, May 12, 2008 at 01:44:22PM +0200, Torfinn Ingolfsen wrote:
On Mon, 12 May 2008 11:37:53 +0200
Torfinn Ingolfsen <torfinn.ingolfsen@xxxxxxxxxxxx> wrote:
FWIW, I had major troubles with re(4) around 7.0-release, and a while
later (I had to use patches). After upgrading to 7-stable on
2008-04-12, re(4) is working for me without patches.
(sigh).. it seems that I spoke to soon. Murphy just showed up. I still
get ssh disconnects (see below) on connections _to_ the machine when
transferring largish amounts of data (like when upgrading a port).
Here is one example (portupgrading the jdk port):
/usr/bin/gcc -fno-strict-aliasing -fPIC -W -Wall -Wno-unused
-Wno-parentheses -pipe -Damd64 -DARCH='"amd64"'
-DRELEASE='"1.6.0_03-p4"'
-DFULL_VERSION='"1.6.0_03-p4-root_12_may_2008_13_25-b00"'
-D_GNU_SOURCE -D_REENTRANT -D_THREAD_SAFE -D_ALLBSD_SOURCE -D_LP64=1
-I.
-I/usr/ports/java/jdk16/work/control/build/bsd-amd64/tmp/java/java.lang.management/management/CClassHeaders
-I../../../src/solaris/javavm/export
-I../../../src/share/javavm/export -I../../../src/share/javavm/include
-I../../../src/solaris/javavm/include
-I../../../src/share/native/sun/management
-I../../../src/solaris/hpi/include -I../../../src/share/native/common
-I../../../src/solaris/native/common
-I../../../src/share/native/java/lang/management
-I../../../src/solaris/native/java/lang/management -c -o
/usr/ports/java/jdk16/work/control/build/bsd-amd64/tmp/java/java.lang.management/management/obj64/ClassLoadingImpl.o
../../../src/share/native/sun/management/ClassLoadingImpl.c
Disconnecting: Bad packet length 3601161525.
The machine is running:
tingo@kg-vm$ uname -a
FreeBSD kg-vm.kg4.no 7.0-STABLE FreeBSD 7.0-STABLE #10: Sat Apr 12
21:42:55 CEST 2008 root@xxxxxxxxxxxx:/usr/obj/usr/src/sys/GENERIC
amd64
and has been up for about 14 days:
tingo@kg-vm$ uptime
1:36PM up 14 days, 17:42, 7 users, load averages: 2.15, 1.85, 1.34
I see that if_re.c for RELENG_7 has been updated on April 22nd, so
I'll upgrade the machine to latest -stable and see if that works
better.
Is this machine using pf(4) at all?