Re: FreeBSD-specific extension of struct mtget breaks dump(8) from Solaris

[Kenneth D. Merry]

On Tue, Feb 01, 2000 at 23:28:17 -0800, Matthew Jacob wrote:
> 
> 
> > rmt(8) returns the result of MTIOCGET filled into struct mtget in
> > response to the 'S' command, issued by dump(8) of some certain OSs
> > including Solaris. (FreeBSD's one seems to not issue the 'S' command)
> > Due to the extention made to struct mtget in FreeBSD, remote dump
> > from, eg Solaris to FreeBSD fails to establish rmt connection.
> > 
> > Maybe we would have to provide something like struct mtget_org for
> > compatibility with other OSs. Also, if we were to issue 'S' in our
> > dump(8), some command extention may be required. (any plans?)
> > 
> > Any comments?
> 
> Yes. This is definitely an issue. Didn't this come up before and was a PR
> filed that I somehow forgot to own? After all, I probably broke things.

I wonder if this has been broken for a long time, likely by the
FreeBSD-specific things in struct mtget.  (They first show up in rev 1.2 of
mtio.h, and were likely from the 386bsd days.)  It would be nice if it
worked between systems, though.

Ken
-- 
Kenneth Merry
[EMAIL PROTECTED]


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: FreeBSD-specific extension of struct mtget breaks dump(8) fromSolaris

[Matthew Jacob]

> > 
> > Yes. This is definitely an issue. Didn't this come up before and was a PR
> > filed that I somehow forgot to own? After all, I probably broke things.
> 
> I wonder if this has been broken for a long time, likely by the
> FreeBSD-specific things in struct mtget.  (They first show up in rev 1.2 of
> mtio.h, and were likely from the 386bsd days.)  It would be nice if it
> worked between systems, though.

Yes, this falls under the "plays well with others" grade.




To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: FreeBSD-specific extension of struct mtget breaks dump(8) from Solaris

[Seigo Tanimura]

[Added Oren Sela, the originator of the PR]

On Tue, 1 Feb 2000 23:28:17 -0800 (PST),
  Matthew Jacob <[EMAIL PROTECTED]> said:

>> rmt(8) returns the result of MTIOCGET filled into struct mtget in
>> response to the 'S' command, issued by dump(8) of some certain OSs
>> including Solaris. (FreeBSD's one seems to not issue the 'S' command)
>> Due to the extention made to struct mtget in FreeBSD, remote dump
>> from, eg Solaris to FreeBSD fails to establish rmt connection.
>> 
>> Maybe we would have to provide something like struct mtget_org for
>> compatibility with other OSs. Also, if we were to issue 'S' in our
>> dump(8), some command extention may be required. (any plans?)
>> 
>> Any comments?

Matthew> Yes. This is definitely an issue. Didn't this come up before and was a PR
Matthew> filed that I somehow forgot to own? After all, I probably broke things.

You mean this one? I saw exactly the same message.

On Wed, 17 Nov 1999 05:50:13 -0800 (PST),
  [EMAIL PROTECTED] said:

>> Number: 14946
>> Category:   i386
>> Synopsis:   rmt - remote magtape protocol
(snip)
>> Description:
osela> I want to use the rmt on bsd vs. solaris 2.7.
osela> When I use ufsdump from the solaris machine to the BSD I get following output:
osela>   DUMP: Writing 32 Kilobyte records
osela>   DUMP: Date of this level 0 dump: Wed Nov 17 15:23:38 1999
osela>   DUMP: Date of last level 0 dump: the epoch
osela>   DUMP: Dumping /dev/rdsk/c0t0d0s7 (hope:/export/home) to
osela> alice:/dev/nrsa0.
osela>   DUMP: Mapping (Pass I) [regular files]
osela>   DUMP: Mapping (Pass II) [directories]
osela>   DUMP: Estimated 436 blocks (218KB).
osela>   DUMP: rmtstatus: expected response size 24, got 76
osela>   DUMP: This means the remote rmt daemon is not compatible.
osela>   DUMP: Lost connection to remote host.
osela>   DUMP: Bad return code from dump: 1
osela> The rmt on Solaris is not compatible with the one on BSD.
osela> I need a solution - can any one help :->

The following patch change mtget to mtget_org, and performs necessary
conversion.



diff -urN --exclude=CVS src.org/sbin/dump/dumprmt.c src/sbin/dump/dumprmt.c
--- src.org/sbin/dump/dumprmt.c Sat Aug 28 09:12:37 1999
+++ src/sbin/dump/dumprmt.c Wed Feb  2 17:08:54 2000
@@ -320,12 +320,28 @@
 {
register int i;
register char *cp;
+#if defined(__FreeBSD__)
+   struct mtget_org mtsorg;
+#endif /* __FreeBSD__ */
 
if (rmtstate != TS_OPEN)
return (NULL);
rmtcall("status", "S\n");
+#if defined(__FreeBSD__)
+   for (i = 0, cp = (char *)&mtsorg; i < sizeof(mtsorg); i++)
+   *cp++ = rmtgetb();
+   /* Convert the compatible struct into FreeBSD's one. */
+   bzero(&mts, sizeof(mts));
+   mts->mt_type = mtsorg->mt_type;
+   mts->mt_dsreg = mtsorg->mt_dsreg;
+   mts->mt_erreg = mtsorg->mt_erreg;
+   mts->mt_resid = mtsorg->mt_resid;
+   mts->mt_fileno = mtsorg->mt_fileno;
+   mts->mt_blkno = mtsorg->mt_blkno;
+#else
for (i = 0, cp = (char *)&mts; i < sizeof(mts); i++)
*cp++ = rmtgetb();
+#endif /* __FreeBSD__ */
return (&mts);
 }
 
diff -urN --exclude=CVS src.org/sys/sys/mtio.h src/sys/sys/mtio.h
--- src.org/sys/sys/mtio.h  Mon Jan 24 11:34:17 2000
+++ src/sys/sys/mtio.h  Wed Feb  2 17:08:11 2000
@@ -136,6 +136,38 @@
u_int32_t mt_comp2; /* compression type for mode 2 */
u_int32_t mt_comp3; /* compression type for mode 3 */
 /* end not yet implemented */
+#endif /* __FreeBSD__ */
+   daddr_t mt_fileno;  /* relative file number of current position */
+   daddr_t mt_blkno;   /* relative block number of current position */
+};
+
+/* The original struct mtget for compatibility of rmt(8) with other OSs */
+
+struct mtget_org {
+   short   mt_type;/* type of magtape device */
+/* the following two registers are grossly device dependent */
+   short   mt_dsreg;   /* ``drive status'' register */
+   short   mt_erreg;   /* ``error'' register */
+/* end device-dependent registers */
+   short   mt_resid;   /* residual count */
+#if notdef /* FreeBSD extention */
+   daddr_t mt_blksiz;  /* presently operating blocksize */
+   daddr_t mt_density; /* presently operating density */
+   u_int32_t mt_comp;  /* presently operating compression */
+   daddr_t mt_blksiz0; /* blocksize for mode 0 */
+   daddr_t mt_blksiz1; /* blocksize for mode 1 */
+   daddr_t mt_blksiz2; /* blocksize for mode 2 */
+   daddr_t mt_blksiz3; /* blocksize for mode 3 */
+   daddr_t mt_density0;/* density for mode 0 */
+   daddr_t mt_density1;/* density for mode 1 */
+   daddr_t mt_density2;/* density for mode 2 */
+   daddr_t mt_density3;/* density for mode 3 */
+/* the following are not yet implemented */
+   u_int32_t mt_comp0; /* compression type for mode 0 */
+   u_int32_t mt_comp1; /* compression type f

NMBCLUSTERS

[Egervary Gergely]

hello,

what size of $SUBJECT should be used on a box with two _extremely_ busy 
100baseTX interfaces?

-- mauzi



To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: Bad memory suspected

[Willem Jan Withagen]

Good hints,

Well actually it is a Cyrix 266 (aka 200Mhz CPU) on an ASUS board.
(I'm not sure about the numbers since it is at home, but it does SCSI, no network)
And I did upgrade the bios, and looked for support, and it seems suported, since the 
BIOS recognises the CPU and speed. I'll again check the voltages

--WjW

Perhaps we should "merge" this thread to [EMAIL PROTECTED] ?

- Oorspronkelijk bericht - 
Van: Matthew D. Fuller <[EMAIL PROTECTED]>
Aan: Chris Dillon <[EMAIL PROTECTED]>
CC: Willem Jan Withagen <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]>
Verzonden: woensdag 2 februari 2000 0:08
Onderwerp: Re: Bad memory suspected


> On Tue, Feb 01, 2000 at 05:01:55PM -0600, a little birdie told me
> that Chris Dillon remarked
> > 
> > The last time I had a problem like this, it was because I put a P54C
> > (Pentium-MMX) into a board only designed for the P53C (a.k.a standard
> 
> ITYM P55C on a P54C board.
> 
> 
> 
> -- 
> Matthew Fuller (MF4839) |[EMAIL PROTECTED]
> Unix Systems Administrator  |[EMAIL PROTECTED]
> Specializing in FreeBSD |http://www.over-yonder.net/
> 
> "The only reason I'm burning my candle at both ends, is because I
>   haven't figured out how to light the middle yet"
> 
> 
N…'²æìr¸›{ûÙšŠ[h™¨è­Ú&£ñkyàRú+ƒ­‡û§²æìr¸›yúÞy»þœ‘êìþ)í…æèw*¶¦zË

Re: downed IP addresses/redundancy

[Benjamin Black]

It's actually slightly more complicated than this as you must be able to
specify *per IP* which MAC should be used, not simply per interface.  That
was the barrier which stopped my VRRP implementation as well.


> On Sun, 30 Jan 2000, Joe Abley wrote:
>
> > On Sat, Jan 29, 2000 at 01:19:53AM +, Tony Finch wrote:
> > > I'd be interested to know of a free implementation of VRRP for the BSD
> > > network stack.
> >
> > I started to look at this a while back, but started to flounder when
> > I looked for an existing interface to allow me to source frames on
> > a local ethernet with a userland-specified MAC address.
> >
> > Actually, I think I looked on OpenBSD, and can't remember whether I
> > looked on FreeBSD too. If anybody has a good idea about how to send
> > and receive frames on a local ethernet interface using one of several
> > possible local MAC addresses (most user-specified) I can probably
> > resurrect the code.
>
> Mhmmm... I'm using the code developed by Bill Paul, to change MAC address
> via special ioctl. It works just fine for me.
>
> http://www.freebsd.org/~wpaul/setmac.tar.gz
>
> Andrzej Bialecki
>
> //  <[EMAIL PROTECTED]> WebGiro AB, Sweden (http://www.webgiro.com)
> // ---
> // -- FreeBSD: The Power to Serve. http://www.freebsd.org 
> // --- Small & Embedded FreeBSD: http://www.freebsd.org/~picobsd/ 
>
>
>
>
> To Unsubscribe: send mail to [EMAIL PROTECTED]
> with "unsubscribe freebsd-hackers" in the body of the message
>



To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: Kernel modules

[Chris D. Faulhaber]

On Tue, 1 Feb 2000, Derek White wrote:

> Hey folks,
> 
> I just wanted to know if anyone could point out a good
> online resource like a tutorial or developer guide
> that explains kernel module development under FreeBSD.

See http://thc.pimmel.com/files/thc/bsdkern.html  Although not necessarily
its indented purpose, this security-related article serves as a fairly
complete kld tutorial.

>  I am new to FreeBSD coming from a Linux/BeOS
> background and am interested in playing around with it
> to maybe contribute back to the FreeBSD source tree. 
> Maybe this is a dumb question but is developing kernel
> modules for FreeBSD the same as it is technically for
> Linux?
> 

-
Chris D. Faulhaber - [EMAIL PROTECTED] - [EMAIL PROTECTED]

FreeBSD: The Power To Serve   -   http://www.FreeBSD.org



To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: NMBCLUSTERS

[Dan Nelson]

In the last episode (Feb 02), Egervary Gergely said:
> what size of $SUBJECT should be used on a box with two _extremely_
> busy 100baseTX interfaces?

Interface type/speed is less important than number of open sockets. 
I've got an NFS server with a 100mbit card in it that is pretty heavily
used the whole day, and after 22 days of uptime, netstat -m shows:

72/596/2112 mbuf clusters in use (current/peak/max)

That's just using whatever value of NMBCLUSTERS a 'maxusers 100' gives
me.

On the other hand, when the Star Wars trailers were released last year,
I was a mirror site, and ran out of mbufs a couple times because I had
a lot of backed-up FTP connections over a T1 link.

Easiest way to determine what you need is to just let the system run
for a while, and then rebuild the kernel with your NMBCLUSTERS set at
your peak value + 50% . 

-- 
Dan Nelson
[EMAIL PROTECTED]


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: Re/Fwd: freebsd specific search

[Michael Bacarella]

> systems have the highest availability rate possible.  Over the last few
> years, I have replaced almost all of our Linux-based servers with FreeBSD,
> due to the quality-control measures that the FreeBSD development team have
> implemented.

Not to start a flame-fest or anything (but who doesn't love em?), I hear
the above quite a lot. 

I'm under the firm belief that a decent sys admin can rub either system to
do whatever they want it to do. Not that I am questioning your abilities.
I just get the "yeah, Linux is good, but just try to use it in a
production environment and you'll understand" a lot.

Granted, a lot of Linux distributions are totally unsuited for a server
environment. Compared to that, I could understand why the 
server-orientedness of FreeBSD is attractive, but I certainly couldn't put
up a reasonable arguement for either side in Slackware Linux vs. FreeBSD.

Could you?

-MB



To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: NMBCLUSTERS

[Egervary Gergely]

> Interface type/speed is less important than number of open sockets. 
> I've got an NFS server with a 100mbit card in it that is pretty heavily
> used the whole day, and after 22 days of uptime, netstat -m shows:
> 
> 72/596/2112 mbuf clusters in use (current/peak/max)
> 
> That's just using whatever value of NMBCLUSTERS a 'maxusers 100' gives
> me.

hmm right. well I've just compiled with 4096, my statistics after 9
minutes uptime:

2308/3182/4096 mbuf clusters in use (current/peak/max)

-- mauzi



To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: NMBCLUSTERS

[Dan Nelson]

In the last episode (Feb 02), Egervary Gergely said:
> hmm right. well I've just compiled with 4096, my statistics after 9
> minutes uptime:
> 
> 2308/3182/4096 mbuf clusters in use (current/peak/max)

I'd say you need to raise it a bit more then if you're at 75% capacity
after only 9 minutes :)  Although if you are already at your peak load
for the day, you might be okay.

-- 
Dan Nelson
[EMAIL PROTECTED]


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: Re/Fwd: freebsd specific search

[Neil Blakey-Milner]

On Wed 2000-02-02 (11:18), Michael Bacarella wrote:
> > systems have the highest availability rate possible.  Over the last few
> > years, I have replaced almost all of our Linux-based servers with FreeBSD,
> > due to the quality-control measures that the FreeBSD development team have
> > implemented.
> 
> Not to start a flame-fest or anything (but who doesn't love em?), I hear
> the above quite a lot. 

(I don't like them.)

> I certainly couldn't put up a reasonable arguement for either side in
> Slackware Linux vs. FreeBSD.
> 
> Could you?

Nope.  I can't say I know Slackware or its quality-control measures,
so could you please answer some questions for me?

What sort of quality-control measures does Slackware have?  Where
do I access their cvs tree?  Where do I access their problem reports?
Where do I subscribe to get every commit message?  How long are
their code freezes?  How many committers do they have?  What
mechanism creates their releases?  Where do I get release-candidates?

(I apologize if that sounds argumentative, it really isn't.  It
just addresses some of the things that I do know about FreeBSD with
regards to quality control, and what I don't know about Slackware
in order to answer your question.)

Neil
-- 
Neil Blakey-Milner
[EMAIL PROTECTED]


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: Re/Fwd: freebsd specific search

[Geoff Buckingham]

On Wed, Feb 02, 2000 at 11:18:23AM -0500, Michael Bacarella wrote:
> 
> > systems have the highest availability rate possible.  Over the last few
> > years, I have replaced almost all of our Linux-based servers with FreeBSD,
> > due to the quality-control measures that the FreeBSD development team have
> > implemented.
> 
> Not to start a flame-fest or anything (but who doesn't love em?), I hear
> the above quite a lot. 
> 
> I'm under the firm belief that a decent sys admin can rub either system to
> do whatever they want it to do. Not that I am questioning your abilities.
> I just get the "yeah, Linux is good, but just try to use it in a
> production environment and you'll understand" a lot.
> 
> Granted, a lot of Linux distributions are totally unsuited for a server
> environment. Compared to that, I could understand why the 
> server-orientedness of FreeBSD is attractive, but I certainly couldn't put
> up a reasonable arguement for either side in Slackware Linux vs. FreeBSD.
> 
I am going to regret this, but 

For production enviroments FreeBSD has two significant advantages.

It's release structure means FreeBSD is a complete operating system (as 
opposed to a kernel and one of several distributions) and machines are 
maintainable and upgradable in production over long periods of time via 
the STABLE branch.

The FreeBSD kernel internals seem to have consistantly scaled further than the
Linux kernel over the last few years (though linux has improved lots recently).
This isn't a problem in most production enviroments however in marginal 
configurations it can be nasty. I had a very bad day six months ago attemting
to patch a linux kernel to have >2048 file descriptors.

-- 
GeoffB



To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: Re/Fwd: freebsd specific search

[Ronald G. Minnich]

On Wed, 2 Feb 2000, Michael Bacarella wrote:
> Granted, a lot of Linux distributions are totally unsuited for a server
> environment. Compared to that, I could understand why the 
> server-orientedness of FreeBSD is attractive, but I certainly couldn't put
> up a reasonable arguement for either side in Slackware Linux vs. FreeBSD.

Linux is definitely a less reliable system for clustering than freebsd.
I've got 5 years of running them both at Sarnoff to back me up. Maybe I
was doing something wrong, but I'm seeing similar problems here at the ACL
on Linux. 

We ran into four classes of problems that linux had that freebsd did not.
These problems are still not fixed as of 2.2.x or 2.3.x.

1) network stack. heavy use of udp can result in a hung kernel. Trivial
   TCP servers that need to take lots of connections cause trouble --
   clients start getting ECONNREFUSED after a while
2) nfs. Hit nfs hard and random clients will hang. The dirty little 
   secret of linux clusters is that 'everyone knows' that you don't run
   client nfs on linux cluster nodes if you want the cluster to stay up. 
   This came out clearly at a cluster conference last spring (JCP4). 
3) vm system. There's still some strange problems in there. 
4) ext2. ext2 does not handle unplanned outages well. There is a 
   reasonable chance that after a power fail you're going to have trouble
   if you have 100 nodes or more. You'll see 2 or 3 in need of help. 


freebsd was just more solid on our clusters. But note that linux isn't
standing still -- it's just not as good as freebsd yet. I had one freebsd
cluster that ran through 5 years of anything we could throw at it -- power
fail, etc. It took disk death to finally halt one node and require me to
hook up a keyboard to it to reload it. 

Our general experience was that NT fails a lot, esp. if you ask it to talk
to a network or run a screensaver. Linux clusters run a long time, but
power outages and other unplanned events will cause it trouble. Freebsd
tolerates very high levels of abuse. The UFS guys really know their stuff.

ron



To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: Re/Fwd: freebsd specific search

[Michael Bacarella]

> What sort of quality-control measures does Slackware have?  Where
> do I access their cvs tree?  Where do I access their problem reports?
> Where do I subscribe to get every commit message?  How long are
> their code freezes?  How many committers do they have?  What
> mechanism creates their releases?  Where do I get release-candidates?

Hmmm. my face is red.

These aren't quite technical limitations as they are political ones.

FreeBSD will naturally be more organized in this respect since the entire
system is under the survey and control of one organization. This certainly
makes it conveniant because you only have one source of information, and
thusly it will be (or is easily made) uniform.

I can't quite subscribe to a Slackware Linux mailing list and get
information on changes to MySQL as well as Apache, but it's not out of the
question for me to listen in on MySQL, Apache, and Slackware information
channels, although their mechanisms obviously are not uniform.

I suppose our needs just vary. :)

But hey, we love FreeBSD too.

-MB



To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: Re/Fwd: freebsd specific search

[Michael Bacarella]

> It's release structure means FreeBSD is a complete operating system (as 
> opposed to a kernel and one of several distributions) and machines are 
> maintainable and upgradable in production over long periods of time via 
> the STABLE branch.

I can agree with you here, as our organization has benefitted from said
features.  Although we aren't having the most pleasant time going from
2.2.8-STABLE to a 3.3.4-STABLE ever since the elfining. But that's
probably our own fault for letting some of them get so out of date. :)

> The FreeBSD kernel internals seem to have consistantly scaled further than the
> Linux kernel over the last few years (though linux has improved lots recently).
> This isn't a problem in most production enviroments however in marginal 
> configurations it can be nasty. I had a very bad day six months ago attemting
> to patch a linux kernel to have >2048 file descriptors.

Useless Anecdote Warning:

For some particular reason (I'll spare the details of what lead up to
this) I ping -f'd a FreeBSD box from a Linux box. The FreeBSD box held up
very well. Despite this, my coworker (who owns said box) put icmp limits
in place to further lessen the load.

This attack was barely noticable on a neighboring Linux machine. (In fact,
I was so annoyed that it had such a meager effect that I went to several
other machines and had them all join in on the ping -f'ing fun). I
experimented with variable packet sizes but I never did quite get to
lock down the Linux machine as badly as I could the FreeBSD.

Maybe I was just equating dropped packets with increased cpu availability
as a plus, whereas FreeBSD put more effort into replying to all of the
requests. Anyway, coworker also modified his system to dedicate more
memory to the mbuf pool as well as some other tweaks which may or may not
have done anything.

Frustrated with my inability to saturate his FreeBSD box, I switched my
attack to UDP. It was a basic UDP packet flood which, evidently,
should've justed pushed as many packets onto the wire as the physical
medium would allow. My machine is a P166 whereas his is a dual PPro200.

The packet flood rendered the machine useless as far as networks go within
10 seconds. He tried it (out of annoyance) on a Linux machine which barely
showed signs of struggle. Then, out of curiousity he turned it towards his
Indy running Irix (don't remember the version). The machine fell right off
the network in less than 3 seconds.  Not only that, but when the flood
stopped, the machine still wouldn't come back.  One week later, the
machine is still dead, despite reboots. Hardware damage? Neat!

We are both using fairly recent versions of our respective systems, but I
don't claim that either of us configured them optimally. Ho hum.

If you can find a moral to this story, let me know. :)

-MB



To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: Re/Fwd: freebsd specific search

[Matthew Dillon]

:On Wed, 2 Feb 2000, Michael Bacarella wrote:
:> Granted, a lot of Linux distributions are totally unsuited for a server
:> environment. Compared to that, I could understand why the 
:> server-orientedness of FreeBSD is attractive, but I certainly couldn't put
:> up a reasonable arguement for either side in Slackware Linux vs. FreeBSD.
:
:Linux is definitely a less reliable system for clustering than freebsd.
:I've got 5 years of running them both at Sarnoff to back me up. Maybe I
:was doing something wrong, but I'm seeing similar problems here at the ACL
:on Linux. 
:
:We ran into four classes of problems that linux had that freebsd did not.
:These problems are still not fixed as of 2.2.x or 2.3.x.
:
:1) network stack. heavy use of udp can result in a hung kernel. Trivial
:   TCP servers that need to take lots of connections cause trouble --
:   clients start getting ECONNREFUSED after a while
:2) nfs. Hit nfs hard and random clients will hang. The dirty little 
:   secret of linux clusters is that 'everyone knows' that you don't run
:   client nfs on linux cluster nodes if you want the cluster to stay up. 
:   This came out clearly at a cluster conference last spring (JCP4). 
:3) vm system. There's still some strange problems in there. 
:4) ext2. ext2 does not handle unplanned outages well. There is a 
:   reasonable chance that after a power fail you're going to have trouble
:   if you have 100 nodes or more. You'll see 2 or 3 in need of help. 
:
:freebsd was just more solid on our clusters. But note that linux isn't
:standing still -- it's just not as good as freebsd yet. I had one freebsd
:cluster that ran through 5 years of anything we could throw at it -- power
:fail, etc. It took disk death to finally halt one node and require me to
:hook up a keyboard to it to reload it. 
:
:Our general experience was that NT fails a lot, esp. if you ask it to talk
:to a network or run a screensaver. Linux clusters run a long time, but
:power outages and other unplanned events will cause it trouble. Freebsd
:tolerates very high levels of abuse. The UFS guys really know their stuff.
:
:ron

Linux has made great strides in the performance area -- the are way ahead
of us on SMP issues, but they are definitely still behind in the 
reliable department.  They almost caught up when we were going through
our 3.0/3.1 fiasco but then fell behind again.  I agree with your general
assessment (though I'm even more rabid about NT, which I consider 
plain and simply to be a piece of crap).

It interesting to note that two years ago it was well known that running
NFSv3 under FreeBSD would destabilize it, so most people ran NFSv2.
Even NFSv2 2 years ago had problems.  Linux is just reaching the point
now with NFSv2 where we were with NFSv3 two years ago.  Thus in regards
to NFS, FreeBSD is about 2 years ahead of Linux.  At this time both
NFSv2 and NFSv3 under FreeBSD are considered stable and reliable.

-Matt
Matthew Dillon 
<[EMAIL PROTECTED]>


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: Re/Fwd: freebsd specific search

[Stephen]

On Wed, Feb 02, 2000 at 10:59:34AM -0800, Matthew Dillon wrote:
> Linux has made great strides in the performance area -- the are way ahead
> of us on SMP issues, but they are definitely still behind in the 
> reliable department.  They almost caught up when we were going through
> our 3.0/3.1 fiasco but then fell behind again.  I agree with your general
> assessment (though I'm even more rabid about NT, which I consider 
> plain and simply to be a piece of crap).
> 
> It interesting to note that two years ago it was well known that running
> NFSv3 under FreeBSD would destabilize it, so most people ran NFSv2.
> Even NFSv2 2 years ago had problems.  Linux is just reaching the point
> now with NFSv2 where we were with NFSv3 two years ago.  Thus in regards
> to NFS, FreeBSD is about 2 years ahead of Linux.  At this time both
> NFSv2 and NFSv3 under FreeBSD are considered stable and reliable.
> 

Re NFS stability.  What version of the 3.x branch contained the updated NFS
code?  3.3?

Thanks,
sk

-- 
[EMAIL PROTECTED]



To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: Re/Fwd: freebsd specific search

[W Gerald Hicks]

> 
> > What sort of quality-control measures does Slackware have?  Where
> > do I access their cvs tree?  Where do I access their problem reports?
> > Where do I subscribe to get every commit message?  How long are
> > their code freezes?  How many committers do they have?  What
> > mechanism creates their releases?  Where do I get release-candidates?
> 
> Hmmm. my face is red.
> 
> These aren't quite technical limitations as they are political ones.
[snip]

Not having an orderly development process has caused quite a few
technical limitations for Linux.  Not having any sort of standardized
and reproducible release process is notable; lack of version control
is one of the more serious problems I have with Linux.

Not really political, but (dis)organizational




To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Removing interfaces

[Archie Cobbs]

With all the PCMCIA card stuff going on, is it now possible to
remove a networking interface in FreeBSD (from within the kernel)?

If so could someone show me an example how. I'd like to implement
this in the ng_iface(8) netgraph node type.

Thanks,
-Archie

___
Archie Cobbs   *   Whistle Communications, Inc.  *   http://www.whistle.com


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

IBM releases JFS for Linux.

[Thomas David Rivers]

This came across the Linux/390 mailing list today, I thought it
might be interesting for people:

>"IBM makes JFS technology available for Linux - Technology based on OS/2
>Warp Journaled File System goes open source". See
>http://oss.software.ibm.com/developerworks/opensource/features/jfs_feature.h
>tml
>


The URL there is incorrect - the correct one is:

  http://oss.software.ibm.com/developerworks/opensource/jfs


- Dave Rivers -



To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: IBM releases JFS for Linux.

[Greg Lehey]

On Wednesday,  2 February 2000 at 22:18:02 -0500, Thomas David Rivers wrote:
>
> This came across the Linux/390 mailing list today, I thought it
> might be interesting for people:
>
>> "IBM makes JFS technology available for Linux - Technology based on OS/2
>> Warp Journaled File System goes open source". See
>> http://oss.software.ibm.com/developerworks/opensource/features/jfs_feature.html
>
> The URL there is incorrect - the correct one is:
>
>   http://oss.software.ibm.com/developerworks/opensource/jfs

Interesting.  I'm downloading and will take a look.

Greg
--
Finger [EMAIL PROTECTED] for PGP public key
See complete headers for address and phone numbers


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

porting linux app. Syscalls

[Marco van de Voort]

I'm new to the list (and to BSD development in general), and I'm developper
of the FreePascal project (www.freepascal.org) which is a bootstrapping 
compiler, completely written in Pascal.

Currently I started preparations for a port of the linux version to FreeBSD.

FreePascal (FPC) doesn't default link to libc for some reasons, and therefore
directly calls syscalls, about 60 of them.

I started translating the syscalls this week (peeking at the linux "emulator" 
code), and now I arrived at the "getdents" and readdir calls. (I want to create
a native port, the emulator already runs fine)

These are supported by the emulator, but actually are emulated (they don't
simply change parameters and do a FBSD syscall)

However when I looked into sys/syscall.h, I found out that the getdents call 
seems to exist? So I have a few questions:-) 

1. Is there documentation for the syscalls other than the code? I know there 
are some manpages, but these seem to be outdated (e.g. the named 
getdents call)

2. Can I use the getdents call to emulate the linux getdents? I use the base 
functionality, not all details have to be fully the same, just the principle. And
why does the emulator not use this call?

3. If I can't use the call, the emulator code seems to call VOP_ functions. Can
 I access these functions from the application?

--- not related to the other 3
4. In the rare event that I get it finished this week; The ports howto doesn't seems to
apply to bootstrapping stuff. (It assumes the apps are gcc based or created by 
a gcc compiled app (like perl).

Where does such an app fit into the BSD project?

- several megs of Pascal source (entire project + sideprojects + texdocs is 25 MB 
source)
- gmake makefiles.
- binary required to start the compile(either native FBSD or linux)

. (compiling the bsd version with the linux 
compiler is also possible I think, if a linux binary can create an ordinary FBSD 
binary) 



Marco van de Voort ([EMAIL PROTECTED])



To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: porting linux app. Syscalls

[Chuck Robey]

On Thu, 3 Feb 2000, Marco van de Voort wrote:

> --- not related to the other 3
> 4. In the rare event that I get it finished this week; The ports howto doesn't seems 
>to
> apply to bootstrapping stuff. (It assumes the apps are gcc based or created by 
> a gcc compiled app (like perl).
> 
> Where does such an app fit into the BSD project?
> 
> - several megs of Pascal source (entire project + sideprojects + texdocs is 25 MB 
>source)
> - gmake makefiles.
> - binary required to start the compile(either native FBSD or linux)
> 
> . (compiling the bsd version with the linux 
> compiler is also possible I think, if a linux binary can create an ordinary FBSD 
> binary) 

This is just answering the last point.  The modula-3 port is about the
same size as yours, and it bootstraps, but (like you said) it does it from
C.  The lang/objc port *used* to require a binary-only bootstrap image,
only.  What I did there was to move the linux image home, use it to
bootstrap a FreeBSD image, stored that on my home page, and made the port
fetch it as one of the distfiles.  In building the port, it just
referenced the FreeBSD-native binary boot image to bootstrap itself, but
it did build a full image of itself from that.

Objc doesn't need that anymore (it has a C boot now) but you could take
that approach.  It's in the archives, probably version 1.5 of the
Makefile, from October 97.

> 
> 
> 
> Marco van de Voort ([EMAIL PROTECTED])
> 
> 
> 
> To Unsubscribe: send mail to [EMAIL PROTECTED]
> with "unsubscribe freebsd-hackers" in the body of the message
> 


Chuck Robey| Interests include C & Java programming, FreeBSD,
[EMAIL PROTECTED]  | electronics, communications, and signal processing.

New Year's Resolution:  I will not sphroxify gullible people into looking up
fictitious words in the dictionary.




To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: porting linux app. Syscalls

[Alfred Perlstein]

* Marco van de Voort <[EMAIL PROTECTED]> [000202 20:02] wrote:
> 
> I'm new to the list (and to BSD development in general), and I'm developper
> of the FreePascal project (www.freepascal.org) which is a bootstrapping 
> compiler, completely written in Pascal.
> 
> Currently I started preparations for a port of the linux version to FreeBSD.
> 
> FreePascal (FPC) doesn't default link to libc for some reasons, and therefore
> directly calls syscalls, about 60 of them.
> 
> I started translating the syscalls this week (peeking at the linux "emulator" 
> code), and now I arrived at the "getdents" and readdir calls. (I want to create
> a native port, the emulator already runs fine)
> 
> These are supported by the emulator, but actually are emulated (they don't
> simply change parameters and do a FBSD syscall)
> 
> However when I looked into sys/syscall.h, I found out that the getdents call 
> seems to exist? So I have a few questions:-) 
> 
> 1. Is there documentation for the syscalls other than the code? I know there 
> are some manpages, but these seem to be outdated (e.g. the named 
> getdents call)

see: http://www.freebsd.org/cgi/man.cgi
you can view linux syscalls from the slackware docs.

> 
> 2. Can I use the getdents call to emulate the linux getdents? I use the base 
> functionality, not all details have to be fully the same, just the principle. And
> why does the emulator not use this call?

It seems that freebsd supports this call but it's hard to tell if the
actual implementation is the same, I would just try some simple 
"hello worlds" to see if you can either use the FreeBSD version
(possibly with a wrapper) or if you'll need to emulate it via a library
call.

It also seems that freebsd's version supports returning more information
to the user application, and it's possible that the linux emulation
code for getdents() actually pre-dates the FreeBSD emulation (it's happened
before)

It'd be interesting to know what you find out.

> 
> 3. If I can't use the call, the emulator code seems to call VOP_ functions. Can
>  I access these functions from the application?

no, those are kernel interfaces into the VFS system, you may not access
them directly, they don't even exist from the userland's perspective.
(mostly) :)

what confuses me is that you don't support bootstrapping from the 
system C compiler.

also, can you wrap lines at 70 characters?

-Alfred


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: Re/Fwd: freebsd specific search

[Matthew Dillon]

:On Wed, Feb 02, 2000 at 10:59:34AM -0800, Matthew Dillon wrote:
:
:Re NFS stability.  What version of the 3.x branch contained the updated NFS
:code?  3.3?
:
:Thanks,
:sk
:
:-- 
:[EMAIL PROTECTED]

3.3 got a big chunk of it but 3.4 has even more.  4.0 has all the bug
fixes (there were some that we simply couldn't backport, mainly related
to garbage beyond the file EOF in mmap() and things like that).

-Matt
Matthew Dillon 
<[EMAIL PROTECTED]>


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: Removing interfaces

[Warner Losh]

In message <[EMAIL PROTECTED]> Archie Cobbs writes:
: With all the PCMCIA card stuff going on, is it now possible to
: remove a networking interface in FreeBSD (from within the kernel)?
: 
: If so could someone show me an example how. I'd like to implement
: this in the ng_iface(8) netgraph node type.

if_detach() is supposed to do this, but there are theoretical problems
with our implementation of it.  Also, there needs to be some way to
propigate the "gone"ness of the interface generically rather than the
ad-hoc way we do it now.

Warner


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: Re/Fwd: freebsd specific search

[Alex Zepeda]

On Wed, 2 Feb 2000, Michael Bacarella wrote:

> Not to start a flame-fest or anything (but who doesn't love em?), I hear
> the above quite a lot. 
> 
> I'm under the firm belief that a decent sys admin can rub either system to
> do whatever they want it to do. Not that I am questioning your abilities.
> I just get the "yeah, Linux is good, but just try to use it in a
> production environment and you'll understand" a lot.

Needless to say I think that FreeBSD makes a great desktop environment
too. What contributes to server sanity also makes things much less
confusing for a desktop user too :)

- alex



To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: Suggestions for Gigabit cards for -CURRENT

[Wes Peters]

"Kenneth D. Merry" wrote:
> 
> On Wed, Feb 02, 2000 at 13:03:09 -0500, Thomas Stromberg wrote:
> > We're currently looking at upgrading several of our FreeBSD servers
> > (dual PIII-600's, 66MHz PCI) and some Sun Ultra's to Gigabit Ethernet.
> > We plan to hook these machines into our Cisco Catalyst 5000 server. They
> > will most likely move to be running FreeBSD 4.x by the time that we
> > actually get our budget approved. What experiences do you guys have with
> > the cards?
> 
> I would recommend getting Alteon boards.  It is likely that the Sun boards
> are Alteon OEM, although I'm not positive.
> 
> One thing to keep in mind is that both Netgear and 3Com are OEMing Alteon
> boards, and you'll get them much cheaper that way.  The boards are pretty
> much identical to the Alteon branded boards (which have no identifying
> marks on them).  The performance is the same, at least for the Netgear
> boards.  (I don't have any 3Com boards.)
> 
> The Netgear GA620 is a 512K Tigon 2 board, and generally goes for around
> $300 or so.  The 3Com boards have 1MB of SRAM, but I'm not sure whether
> they're Tigon 1 or Tigon 2.  You really want a Tigon 2 board.  Maybe
> someone who has one can comment.

I have a couple of GA620's here, too, and they work quite well.  They support
both 66Mhz and 64-bit PCI bus.  We paid $329 from warehouse.com about 4 months
ago, quite a reasonable price.

You might want to look around and see if anyone has Solaris drivers for 
the Tigon-2, if you're using PCI Sun machines.

-- 
"Where am I, and what am I doing in this handbasket?"

Wes Peters Softweyr LLC
[EMAIL PROTECTED]   http://softweyr.com/


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: Suggestions for Gigabit cards for -CURRENT

[Kenneth D. Merry]

On Wed, Feb 02, 2000 at 15:11:16 -0700, Wes Peters wrote:
> "Kenneth D. Merry" wrote:
> > 
> > On Wed, Feb 02, 2000 at 13:03:09 -0500, Thomas Stromberg wrote:
> > > We're currently looking at upgrading several of our FreeBSD servers
> > > (dual PIII-600's, 66MHz PCI) and some Sun Ultra's to Gigabit Ethernet.
> > > We plan to hook these machines into our Cisco Catalyst 5000 server. They
> > > will most likely move to be running FreeBSD 4.x by the time that we
> > > actually get our budget approved. What experiences do you guys have with
> > > the cards?
> > 
> > I would recommend getting Alteon boards.  It is likely that the Sun boards
> > are Alteon OEM, although I'm not positive.
> > 
> > One thing to keep in mind is that both Netgear and 3Com are OEMing Alteon
> > boards, and you'll get them much cheaper that way.  The boards are pretty
> > much identical to the Alteon branded boards (which have no identifying
> > marks on them).  The performance is the same, at least for the Netgear
> > boards.  (I don't have any 3Com boards.)
> > 
> > The Netgear GA620 is a 512K Tigon 2 board, and generally goes for around
> > $300 or so.  The 3Com boards have 1MB of SRAM, but I'm not sure whether
> > they're Tigon 1 or Tigon 2.  You really want a Tigon 2 board.  Maybe
> > someone who has one can comment.
> 
> I have a couple of GA620's here, too, and they work quite well.  They support
> both 66Mhz and 64-bit PCI bus.  We paid $329 from warehouse.com about 4 months
> ago, quite a reasonable price.
> 
> You might want to look around and see if anyone has Solaris drivers for 
> the Tigon-2, if you're using PCI Sun machines.

Alteon includes a Solaris driver in their driver development kit.  It might
be possible to modify the probe routine to recognize other Tigon boards
like the Netgear or 3Com boards.

Ken
-- 
Kenneth Merry
[EMAIL PROTECTED]


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Suggestions for Gigabit cards for -CURRENT

[Thomas Stromberg]

We're currently looking at upgrading several of our FreeBSD servers
(dual PIII-600's, 66MHz PCI) and some Sun Ultra's to Gigabit Ethernet.
We plan to hook these machines into our Cisco Catalyst 5000 server. They
will most likely move to be running FreeBSD 4.x by the time that we
actually get our budget approved. What experiences do you guys have with
the cards?

Currently we're looking at the ~$1000 range,  specifically at Alteon
512k's ($1000) for the FreeBSD servers and Sun Gigabit 2.0's ($2000) for
the Sun servers. I was interested in the Myrinet cards (for obvious
reasons), but they appear to require a Myrinet switch (though I found
myself slightly confused so I may be wrong) rather then being able to
hook into our Catalyst 5000. The Intel PRO/1000 Gigabit cards look
rather nice too, but I haven't seen drivers yet for FreeBSD (Linux yes).

I'm pretty much purchasing on marketing and reputation rather then any
experience here, so any help would be much appreciated. 

-- 
===
Thomas R. StrombergAsst. IS Manager / Systems Guru
FreeBSD Contrib, Security Geek, etc.   Research Triangle Commerce, Inc.
http://www.afterthought.org/   http://www.rtci.com/
[EMAIL PROTECTED]   [EMAIL PROTECTED]
---
 MCSE: McDonald's Certified Service Engineer
===


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message

Re: Suggestions for Gigabit cards for -CURRENT

[Kenneth D. Merry]

On Wed, Feb 02, 2000 at 13:03:09 -0500, Thomas Stromberg wrote:
> We're currently looking at upgrading several of our FreeBSD servers
> (dual PIII-600's, 66MHz PCI) and some Sun Ultra's to Gigabit Ethernet.
> We plan to hook these machines into our Cisco Catalyst 5000 server. They
> will most likely move to be running FreeBSD 4.x by the time that we
> actually get our budget approved. What experiences do you guys have with
> the cards?
> 
> Currently we're looking at the ~$1000 range,  specifically at Alteon
> 512k's ($1000) for the FreeBSD servers and Sun Gigabit 2.0's ($2000) for
> the Sun servers. I was interested in the Myrinet cards (for obvious
> reasons), but they appear to require a Myrinet switch (though I found
> myself slightly confused so I may be wrong) rather then being able to
> hook into our Catalyst 5000. The Intel PRO/1000 Gigabit cards look
> rather nice too, but I haven't seen drivers yet for FreeBSD (Linux yes).
> 
> I'm pretty much purchasing on marketing and reputation rather then any
> experience here, so any help would be much appreciated. 

I would recommend getting Alteon boards.  It is likely that the Sun boards
are Alteon OEM, although I'm not positive.

One thing to keep in mind is that both Netgear and 3Com are OEMing Alteon
boards, and you'll get them much cheaper that way.  The boards are pretty
much identical to the Alteon branded boards (which have no identifying
marks on them).  The performance is the same, at least for the Netgear
boards.  (I don't have any 3Com boards.)

The Netgear GA620 is a 512K Tigon 2 board, and generally goes for around
$300 or so.  The 3Com boards have 1MB of SRAM, but I'm not sure whether
they're Tigon 1 or Tigon 2.  You really want a Tigon 2 board.  Maybe
someone who has one can comment.

The Intel cards may look nice, and there is a FreeBSD driver for them, but
I wouldn't get one.  The first problem with the Intel boards is that there
are no docs for them.  Supposedly they're using a Cisco chip, and the specs
for the chip are top secret.

The FreeBSD driver (written by Matt Jacob) is based on the Linux driver,
which Intel wrote, and he hasn't yet managed to get decent throughput
through the cards.  (Maybe Matt will comment.)  They also only have 64K of
memory on board, which is insufficient for a heavily loaded server, IMO.

Even with the 512K Alteon boards, you have a minimum of about 200K, and
probably more like 300K of cache for transmit and receive.

The Intel boards also don't have the features necessary to really support
zero copy TCP receive.

The Alteon boards, on the other hand, have most of the features necessary,
and if I get some time, I may add the last feature (header splitting) to
the firmware.

The other alternative is SysKonnect, and that might actually be a good
alternative.  I haven't seen the boards, don't know how much they cost,
etc. etc.  You might want to ask Bill Paul about them, he wrote the driver.

Ken
-- 
Kenneth Merry
[EMAIL PROTECTED]


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-hackers" in the body of the message