gmane.os.openbsd.newbies

Re: varieties of hostname

Jim Barchuk — 2012-04-28T14:04:08

HiHi Dave!

I *knew* this had to be an easy Q and didn't require a rocket science answer, so that's why I posted it to -newbies. :)

BINGO!!!!!!!!!! Ding! Ding! Ding! Ding! Ding! Ding! Ding! Ding! This was the one last detail I -missed-, moving from AT&T DSL to Uverse. I had this same issue with last two previous AT&T DSL installations. I actually did used to run my own bind years ago but for me it was way too much work and one of the few non-fun internet features. Since netsol started doing DNS I never had to bother with it again, just have to nag AT&T to install the reverse. But I've dealt with this issue only maybe 3x in 10 years and that's why I *forgot* about it.

(Added later...) Well, yes that definitely had to be part of it. Took a few minutes but they added it immediately.

But there was another problem, because outbound still didn't work. Spent another fwe hours with tech support. After it was identified and fixed (in their configs) I called AT&T and got a real manager on to listen to me *RANT* for half an hour, and it went about like this: When I first ordered the UV I was ADAMANT about making SURE that I could do 'anything I wanted,' that there were no restrictions, blocks, limits, nothing. I was reassured by several people that yes that was all completely true. And during the whoooole six weeks of trying to get this email working no one mentioned this little tidbit to me --> one tech finally asks, hey, are you using port 25 for outbound email? I said sure, why would I use anything else. He says well, it's blocked. [[!!!!!!!!!!!!!]] I WENT ABSOLUTELY BALLISTIC!!! SIX WEEKS!!!!!!!!!!!!! THE ONLY PROBLEM I HAVE IS OUTBOUND EMAIL AND YOU ***FAIL*** TO MENTION THAT PORT 25 IS BLOCKED!!! That was only the third time in 30
years that -I- ever hung up on tech support. Then I waited an hour to get my thoughts in a row, got a manager on the fone, and did a puuuuure rant for a solid half hour. She gave me a refund on a month's service.

But... now I have no inbound service. Weird. Still looking at it.

The AT&T tech/hardware/service side of things has always been 100% awesome and rock solid, it just keeps on ticking. But the administrivia is always a pure disaster. I have a hundred nightmare/horror stories, this one topped them all, this is one for the highlight reel.

/added

I did run http, ftp, irc server through literally dialup 20 yrs ago. When Worldnet unlimited meant exactly that. :) :) Not true name based hosting, but a script automatically detected dynamic IP and uploaded page with links to server to geocities. It was a great learning experience.

I -could- have had uverse maybe 5 years ago, but the TOS said 'thou shalt run HTTPP and SMTP server and that's it nothing more.' Which annoyed me. New TOS doesn't have any such limitations. I went as fast as to talk with sales and tech support to make -sure- what I read. Oddly, their DSL never had any server limits.

I've always gotten excellent tech support with AT&T all the way back to dialup. Sometimes it takes a -while- to explain and even prove a problem, but they are willing to listen, and when a problem is proved they're willing to fix it. Unfortunately this time when they moved my account from DSL to UV the right hand knoweth not what the left doeth, so they didn't bother. But now I'm not surprised because it's not a 'move' or a 'port' for them, it's two separate companies and -I- close one account and open another. And I didn't dot every 'T' and cross every 'I'.

Oh OK I did not realise that. (Note to self: locate alpine |xargs rm -rf.) That sucks I'm really gonna miss it. :(

It's not ignorance as in 'should know but don't' as much as uninformed. The vast majority of unwashed masses simply have no interest in anything beyond the PC other than entertainment. I don't fault anyone for that, they're just not true geeks. They have their place too because they -can- do the things that I can't and/or don't -want- to do such as ummm... 'tech support.' LOL!

Not necessarily, but again they're uninformed. The folks I've told are *ASTOUNDED* that Firefox has a web server add-on. Maybe 'non-informed' is a better way to say it The add-on list calls it a 'Browser Server.' Eventually they mention the word webserver, but the word httpd isn't in there.

I knew that none of those were the issue. I've been around those many times.

Thanks a ton for the RDNS clue. I'll do another few hours looking at receipt failure and maybe ask here. I'm looking at tcpdump to try to decipher what's connecting/not connecting. SSH from the outside world works, there're very few problems.

Have a :) day!

Jim

Re: varieties of hostname

Woodchuck — 2012-04-26T19:51:42

Speaking of DNS, I see this: <snipped a little>

[djv< at >aemilia ~ 1:393]$ dig jbarchuk.com   


;; ANSWER SECTION:
jbarchuk.com.           7200    IN      A       99.28.54.137

;; AUTHORITY SECTION:
jbarchuk.com.           172800  IN      NS      ns74.worldnic.com.
jbarchuk.com.           172800  IN      NS      ns73.worldnic.com.

...for the forward and this for the reverse:

[djv< at >aemilia ~ 1:394]$ dig -x 99.28.54.137

; <<>> DiG 9.4.2-P2 <<>> -x 99.28.54.137
;; global options:  printcmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 34608
;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 3, ADDITIONAL: 3

;; QUESTION SECTION:
;137.54.28.99.in-addr.arpa.     IN      PTR

;; ANSWER SECTION:
137.54.28.99.in-addr.arpa. 7097 IN      PTR     99-28-54-137.uvs.dybhfl.sbcglobal.net.

;; AUTHORITY SECTION:
54.28.99.in-addr.arpa.  172697  IN      NS      ns1.swbell.net.
54.28.99.in-addr.arpa.  172697  IN      NS      ns3.sbcglobal.net.
54.28.99.in-addr.arpa.  172697  IN      NS      ns2.swbell.net.

So this isn't going to work for email in the general case, since
many email receivers will refuse email unless the forward and reverse
DNS match, and then some will refuse connections from anything they
think is a "consumer", i.e. a dynamically assigned IPA or the widely
feared "dialup" account.  

Yeah, this steps on all kinds of assumptions about TCP/IP being a
network of equal hosts, all kinds of RFCs about email, and so on,
but these "anti-spam" zealots will just turn up their noses and
refuse the connections, and give you a condescending brush-off.
Some ISPs will callously block port 25 both ways.  Isn't that
special?

So OP needs "better" DNS, or needs to do the inconvenient work-around,
which is to configure his MTA (sendmail) to do "MASQUERADE AS" and
to do it through a "SMARTHOST", and those are not in the default
OpenBSD sendmail.mc.  Complicating this are the "security" "precautions"
that the smarthost (i.e. relay) will demand ... encryption, SSL,
authentication in some inconvenient way, and possibly some other
stuff.  Some additional header munging may be necessary, either
in sendmail or the MUA (mutt, pine, or, <gag> thunderbird). 

See, you're supposed to use google/gmail and shut up, not do it
yourself.  Your supposed to use DHCP, blah blah blah.  300 million
Windoze users already believe that their home computer is, netwise,
fundamentally different from a "server", and this ignorant misconception
is being morphed into the truth.  This is being done to facilitate
surveillance, to control k-pr0n, money laundering and the theft of
"intellectual property", to protect "medical data", "fight terrism",
stop "identity theft", and a thousand other fronts and excuses
for what amount basically to police/nanny motives.  The net has
changed.  Wait till the nannies get their hooks into IPV6; they're
salivating for that.  But I digress.

OP didn't give any specifics about his configuration.  It might help
to have the exact contents of /etc/myname, /etc/hosts, whether BIND/named
is being used, if so, how.  Remarks about ssh configuration needs to
be more specific, too.  We can *guess* about those things, but that's
just lost motion at best if the truth is available. One of OP's symptoms
sounded like the result of a typo, but can't say until I see the type.

OP: you might have a look in /etc/rc* to see the actual fate of "myname".
This is not as sticky nor pervasive as you think.

Dave

Re: varieties of hostname

Daniel Staal — 2012-04-26T17:23:45


Typically email it should (nearly) all be in the sendmail 
configuration.  I wouldn't touch anything else, except for perhaps 
resolv.conf.  (And even there I'd only touch it if the box doesn't have 
any DNS lookups going on, which for a mail server I'd consider 
*extremely* unlikely.)  I haven't worked with sendmail (I just replace 
it), but Postfix I know can receive email for multiple domains as well 
as the local box, using nothing more than it's own config and the 
correct entries in DNS.

Daniel T. Staal

---------------------------------------------------------------
This email copyright the author.  Unless otherwise noted, you
are expressly allowed to retransmit, quote, or otherwise use
the contents for non-commercial purposes.  This copyright will
expire 5 years after the author's death, or in 30 years,
whichever is longer, unless such a period is in excess of
local copyright law.
---------------------------------------------------------------

varieties of hostname

Jim Barchuk — 2012-04-26T15:07:54

Hi All!
New OBSD 5.0 install, and by -several- symptoms I think I'm having a lower level 'server identity' issue.

With all the years I used RH (since 4.1) and then moved to OBSD 4.4, I was *never* truly 'forced' to conform to an individual specific 'hostname.' About all I've ever run is one PC running all server stuff with not much LANning. This was the first install that absolutely insisted that I not leave it either blank, or not be able to later -easily- change it from host.domain.com to domain.com. OK so I can go along with that and set it to srv01. (I have a few LAN PCs around here but no other true servers.)

By 'easily change it' I mean that I notice that install uses hostname to create several /etc/ssh configuration files. I hesitate to -just- edit /etc/myname because I think that might interfere with authentication. Not sure, but I don't want to mess with anything I don't know a bunch about.

My main concern at the moment is email, which I can neither send nor receive. Within the LAN I can SSH to the server, so I know a lot of things are basically functional.

I *think* that my email issue is that I'm identifying myself as srv01.jbarchuk.com, and that's not what the outside world is expecting to find here.

I added that hostname to my DNS at netsol, but with no change in effect.

I added jbarchuk.com to /etc/hosts, and similarly no change.

Should I be looking at pf.conf, sendmail.mc? Or some other configuration?

Are there effects in email where a 'primary' name takes such precedence that I have to jump through certain hoops to make other names work properly?

I'm also fairly new to this uverse 'intelligent' modem, but I think that's not the issue here. At the server I see connection attempts but everything inbound and outbound always reports simply 'connection timed out.'

I'm not annoyed by any of this. I understand that OBSD out of the box has basically nearly everything turned off and I have to figure out the right ways to turn things on, but this one has me stymied.

I'm also not annoyed that I must (or definitely should) use the 'true hostname.' I don't necessarily -want- to change /etc/myname to jbarchuk.com because I do intend to add other server hardware, or at least have the option to do it more easily.

TYVM. Have a :) day!

Jim

Re: varieties of hostname

Jim Barchuk — 2012-04-26T15:46:04

More clues.

From a LAN PC, 'telnet 99.28.54.137 25' fails with 'Connecting To 99.28.54.137...Could not open connection to the host, on port 25,' but I'm less concerned with that right now.

On the plus side, at the console telnet plus sending an email locally worked fine.

However when I logged in it replied:

250- jbarchuk.com Hello jb2localhost... which is -not- /etc/myname, so unsure what it meant.

Then I recalled that as part of trying to identify myself properly I added several names to /etc/hosts. I listed jbarchuk.com as the first name. Previously that had no effect (in terms of accepting/sending email.) Now I see that the -first- name after 99.28.54.137 is how it identified itself. I can change the EHLO response at will.

But again I don't know if that can cause authentication problems if it 'disagrees' with /etc/myname, and the ssh configs. Maybe I need to create a new SSH config for -each- name I use???

I'm trying to swim with the current, not fight it. :) :)


Thanks much. Have a :) day!

Jim

Re: mount old drive problem with mount/disklabel/fdisk

Nick — 2012-03-16T19:17:44

SATA drive? It should be sd* then. Is your BIOS in fake-IDE mode? Check
your dmesg, or equivalently (but somewhat more usefully) do tail
-f /var/log/messages.log and plug the drive in and see if it spits out
anything when you do that.


The disklabel and the mbr are separate records. If there isn't a proper
disklabel then OpenBSD generates a fake in-kernel disklabel based on the
MBR (e.g. for thumbdrives) but that's the exception, not the rule. It
looks like your disklabel got overwritten, and since there's only an
OpenBSD partition in the MBR, OpenBSD created that fake disklabel for
you. I'm unsure how to check for if a disklabel actually resides on a
disk or not (without resorting to low-level scanning and parsing of C
structs), and disklabel(8) just says
  "The kernel's in-core copy of the label is displayed; if the disk has
no label, or the partition types on the disk are incorrect, the kernel
may have constructed or modified the label."
 which is somewhat unhelpful.

..that sucks. If switching your BIOS to proper SATA mode doesn't
magically help, do you have a backup of your disklabel? You should
probably read http://www.openbsd.org/faq/faq14.html#OhBugger.

Does it still mount on 4.9 (set up a VM with qemu and point it at the
real disk)?

Good luck,
-Nick

Re: mount old drive problem with mount/disklabel/fdisk

Woodchuck — 2012-03-16T14:53:35

I assume you think the drive is empty.  The disklabel suggests
that the drive has been wiped clean of data.  If you're expecting
that there is data on the drive, say old files, then something
is wrong, and the kernel cannot read the disklabel and is using
its own notions.

Look in your dmesg output, see if the kernel probed it.

Get the output of   dmesg | grep wd

You should see all your drives.  (Three or more, right?)

Assuming this is a wiped drive and you want to use it, then
the rest of the email applies.  If, however, you think this
drive should have existing filesystems and files on it, you
have, as they say, a Problem.

/dev/wd2c is not a file system, it is a reference to the entire drive.

You need to create one or more partitions with disklabel on that
drive, and then run newfs on those partitions.  I recommend the
-E option to disklabel.  

If you wish to just use the drive for data, you can make
a partition, say "a" to cover the drive.  Make it of fstype
"4.2BSD", which is the default anyway.

Then newfs on that partition and then mount it.

But on a drive that large, I might want to make several partitions.
(don't mess with b or c.) This is a matter of personal preference.

If, oddly, you actually used to use that drive as it is now, with
partition 'c' as an actual file system [Don't Do This] then try
mounting it  # mount /dev/wd2c /mnt   If that succeeds, then
come back for more info.  It ought to fail.

I'm sorta-kinda confused.

Dave

On Sat, Mar 17, 2012 at 12:11:58AM +1100, kim wrote:

mount old drive problem with mount/disklabel/fdisk

kim — 2012-03-16T13:11:58

Hi

I just updated to 5.0, and I'm trying to mount an old sata drive which 
use to hold my home directories and keep getting the following;

# mount -t ffs /dev/wd2a /home/kim/mount_point/
mount_ffs: /dev/wd2a on /home/kim/mount_point: Device not configured


looking at it with disklabel I get the following, which seems rather odd?


# disklabel -h wd2
# /dev/rwd2c:
type: ESDI
disk: ESDI/IDE disk
label: ST3500641AS
duid: 0000000000000000
flags:
bytes/sector: 512
sectors/track: 63
tracks/cylinder: 255
sectors/cylinder: 16065
cylinders: 60801
total sectors: 976773168 # total bytes: 465.8G
boundstart: 64
boundend: 976768065
drivedata: 0

16 partitions:
#                size           offset  fstype [fsize bsize  cpg]
   c:           465.8G                0  unused

looking at the drive with fdisk I can see the partition

# fdisk wd2
Disk: wd2       geometry: 60801/255/63 [976773168 Sectors]
Offset: 0       Signature: 0xAA55
             Starting         Ending         LBA Info:
  #: id      C   H   S -      C   H   S [       start:        size ]
-------------------------------------------------------------------------------
  0: 00      0   0   0 -      0   0   0 [           0:           0 ] 
unused
  1: 00      0   0   0 -      0   0   0 [           0:           0 ] 
unused
  2: 00      0   0   0 -      0   0   0 [           0:           0 ] 
unused
*3: A6      0   1   2 -  60800 254  63 [          64:   976768001 ] OpenBSD

So how do you go about mounting a partition which fdisk can see but 
nothing else seems to find, or is the drive just cactus?

cheers for any insights

kim

Re: Epiphany crashes reliably

Stuart Henderson — 2012-02-11T15:03:38

ulimit -f is a file size limit, see ksh(1) for a description of ulimit
options. You need ulimit -n. If it's too low, edit login.conf and find
the relevant openfiles-cur for the class your userid is in. Logout and
back in again to use the new limits.

If this isn't enough, you could also be running into system limits.

$ sysctl kern.{n,max}files
kern.nfiles=315
kern.maxfiles=7030

kern.maxfiles can be raised temporarily (sysctl kern.maxfiles=NN) or at
boot in /etc/sysctl.conf.

fstat will show you what's open. Try something like "fstat | grep
-c username" for a quick estimate of what you're using.

Wireless bridge Question

Ed D. — 2011-11-21T17:28:58

Hi Everybody.

I hope I can clearly explain my problem.
I want to solve this with on hand hardware.

I have a firewall/router in a locked cabinet in the basement.
It's running OpenBSD 5.0 Current.
I also have some 2-Wire brand 1700 and 2700 series DSL
routers lying around.

I need a wireless access point upstairs, out of range of a
wireless card in my firewall/router.

I want to run a wired network connection from my firewall/router
upstairs to one of my 2-Wire routers to act as a wireless access
point.

BUT...
I want the DHCP leases to be handled by the wired NIC in my
firewall/router, and have the 2-Wire router ONLY act as nothing more
than a "bridge" (if that's the correct term).

Is this possible?
                              Thanks for any input,  Ed

test

kaspop — 2011-08-22T17:12:42

test

Re: Kernel pppoe STILL broken my OpenBSD 4.9 and64

Woodchuck — 2011-06-28T03:20:14

Just out of curiosity, are these errors on boot-up, or later in the
game?  If on boot-up, you might consider not starting named
until later.

216.239.38.10  -- this is a number at google.  I wouldn't give a
nickel for the reliability of google's services of any kind.  Why is
your named trying to get hold of google's?

No help, I know, just questions.

You're scaring me from installing 4.9.

Dave
--
"All money nowadays seems to be produced with a natural
homing instinct for the Treasury."  -- Prince Phillip

Re: Kernel pppoe STILL broken my OpenBSD 4.9 and64

Ed D. — 2011-06-28T03:09:30


On 27 Jun 2011 at 23:10, Ed  D. wrote

With this latest fix, kernel mode pppoe is improved
but still broken in my case.
having worked without problem until 4.9 stable.


When I got back in town, I upgraded to amd64 -current
on Friday 6-24-11.

The assertwaitok panic doesn't occur anymore on bootup.

However, when the DSL connection drops, the error
still occurs.

I'm still having to use usermode pppoe to keep my DSL
connection going.

Example from /var/log/mesages

Jun 25 00:21:59 meenon /bsd: pppoe0: LCP keepalive timeout
Jun 25 00:21:59 meenon /bsd: splassert: assertwaitok: want -1 have 1
Jun 25 00:22:00 meenon named[13407]: 
/usr/src/usr.sbin/bind/lib/isc/unix/socket.c:1218: unexpected error:
Jun 25 00:22:00 meenon named[13407]: internal_send: 216.239.38.10#53: 
Network is down
Jun 25 00:22:01 meenon named[13407]: 
/usr/src/usr.sbin/bind/lib/isc/unix/socket.c:1218: unexpected error:
Jun 25 00:22:01 meenon named[13407]: internal_send: 202.12.27.33#53: 
Network is down
Jun 25 00:22:01 meenon named[13407]: 
/usr/src/usr.sbin/bind/lib/isc/unix/socket.c:1218: unexpected error:
Jun 25 00:22:01 meenon named[13407]: internal_send: 202.12.27.33#53: 
Network is down
Jun 25 00:22:01 meenon named[13407]: 
/usr/src/usr.sbin/bind/lib/isc/unix/socket.c:1218: unexpected error:
Jun 25 00:22:01 meenon named[13407]: internal_send: 202.12.27.33#53: 
Network is down
Jun 25 00:22:01 meenon named[13407]: 
/usr/src/usr.sbin/bind/lib/isc/unix/socket.c:1218: unexpected error:
Jun 25 00:22:01 meenon named[13407]: internal_send: 202.12.27.33#53: 
Network is down
Jun 25 00:22:02 meenon named[13407]: 
/usr/src/usr.sbin/bind/lib/isc/unix/socket.c:1218: unexpected error:
Jun 25 00:22:02 meenon named[13407]: internal_send: 198.41.0.4#53: 
Network is down
Jun 25 00:22:02 meenon named[13407]: 
/usr/src/usr.sbin/bind/lib/isc/unix/socket.c:1218: unexpected error:
Jun 25 00:22:02 meenon named[13407]: internal_send: 198.41.0.4#53: 
Network is down

The system just keeps on generating this same "network is down" error
until thr system is rebooted.

Re: FIXED? Kernel pppoe broken my OpenBSD 4.9 and64

Ed D. — 2011-06-21T18:03:16

Was just checking the -Current Changelog and
noticed:
"Fixed an assertwaitok panic in sppp(4). "

When I get back in town I'll upgrade my
firewall/router to -Current and report
back the results.
                                     Ed

On 15 Jun 2011 at 10:11, Stuart Henderson wrote:

Re: Kernel pppoe broken my OpenBSD 4.9 and64

Stuart Henderson — 2011-06-15T09:11:24

It's not totally broken, but it is broken in your particular
situation. If it was easier to reproduce it would already be fixed.

$ shmux -c 'ifconfig pppoe|grep ^pppoe;uptime' jodrell mylor2-gw
  jodrell: pppoe0: flags=8851<UP,POINTOPOINT,RUNNING,SIMPLEX,MULTICAST> mtu 1492
  jodrell: 10:05AM  up 6 days, 17:51, 1 user, load averages: 0.13, 0.13, 0.08
mylor2-gw: pppoe3: flags=8851<UP,POINTOPOINT,RUNNING,SIMPLEX,MULTICAST> mtu 1492
mylor2-gw: pppoe4: flags=8851<UP,POINTOPOINT,RUNNING,SIMPLEX,MULTICAST> mtu 1492
mylor2-gw: pppoe6: flags=8851<UP,POINTOPOINT,RUNNING,SIMPLEX,MULTICAST> mtu 1492
mylor2-gw: pppoe7: flags=8851<UP,POINTOPOINT,RUNNING,SIMPLEX,MULTICAST> mtu 1492
mylor2-gw:  9:06AM  up 36 days, 20:57, 0 users, load averages: 0.13, 0.11, 0.22

These machines are totally stable, despite some of the pppoe's on
mylor2-gw here having dropped the connection hundreds of times since
the machine booted.

$ ssh mylor2-gw 'ifconfig pppoe7|grep -e ^pppoe -e time'
pppoe7: flags=8851<UP,POINTOPOINT,RUNNING,SIMPLEX,MULTICAST> mtu 1492
        sid: 0xd3d PADI retries: 3520 PADR retries: 0 time: 15:28:25

Re: Kernel pppoe broken my OpenBSD 4.9 and64

Ed D. — 2011-06-14T06:51:30

So I infer from the lack of any further responses that
kernel mode pppoe is broken for OpenBSD 4.9 Stable,
and the only way to have a working pppoe is to go back
to using the user mode pppoe (8)?
(which I had done when I first posted this thread)
                                       Thanks,  Ed

On 7 Jun 2011 at 12:47, Ed  D. wrote:

Re: boot.conf on mac ppc

Bryan — 2011-06-09T20:57:16

He has a macppc...  do a man for boot.conf, and one doesn't exist.
_______________________________________________
Openbsd-newbies mailing list
Openbsd-newbies< at >sfobug.theapt.org
http://mailman.theapt.org/listinfo/openbsd-newbies

Re: boot.conf on mac ppc

Bryan Irvine — 2011-06-09T18:38:33

http://www.openbsd.org/cgi-bin/man.cgi?query=boot.conf&apropos=0&sektion=0&manpath=OpenBSD+Current&arch=i386&format=html

:-)

boot.conf on mac ppc

Satyriasis — 2011-06-09T15:40:39

I'm running 4.9 STABLE on a PPC Mac Powerbook 15".

OpenFirmware is set for autoboot and the boot device is hd:,ofwboot.

The box boots correctly: a start or reboot drops me into an OBSD boot 
prompt that lasts for five seconds.  I'd like to control the length of 
time that the boot prompt is on the screen with the boot.conf "timeout" 
keyword.

This works, of course, on the i386 port (you can set "timeout" to 0, and 
override it by holding down one of the keys during boot), but macppc 
throws out any timeout boot.conf configuration with an error message.

Does boot.conf work on macppc?  If so, is there a difference in the 
syntax from i386?

Alternatively, can the length of time the secondary boot prompt 
(ofwboot) is on the screen be controlled from OpenFirmware? (I 
understand they're two different processes.)

Thanks for any help.

Re: Kernel pppoe broken my OpenBSD 4.9 and64

Ed D. — 2011-06-07T16:47:50

Thanks for the reply.

Actually the thread you gave me is one of the ones I found when
I Googled.
But when I read it through to the end, there doesn't appear to be
a clear agreed on solution to this.

But something has changed from OpenBSD 4.8 Stable and 4.9
Stable that makes kernel pppoe not work right anymore.for me.

I can reproduce this problem at will by simply unplugging the
phone line or ethernet cable from my DSL modem.

I've been jusing kernel mode pppoe for several versions and
it's been bulletproof for me until 4.9  Stable.
                                                          Thanks, Ed

On 7 Jun 2011 at 7:42, Bryan wrote:

Re: Kernel pppoe broken my OpenBSD 4.9 and64

Bryan — 2011-06-07T12:42:33

Maybe you didn't google hard enough...  always check MARC

http://marc.info/?l=openbsd-tech&m=129542722106887&w=2

On Mon, Jun 6, 2011 at 22:24, Ed  D. <lists< at >rensseltucky.com> wrote:
_______________________________________________
Openbsd-newbies mailing list
Openbsd-newbies< at >sfobug.theapt.org
http://mailman.theapt.org/listinfo/openbsd-newbies