Re: VMWare Tools for FreeBSD

2008-04-02 Thread Shawn Barnhart

David Robillard wrote:

Basically the only reason I have for using VM Tools is for the ability
of Vmotion and such with our ESX Server farm. It's really the only
benefit that the VM tools will give me on FreeBSD as all my virtual
machines which are running FreeBSD are servers and don't use any GUI's
either.

Currently there is nothing that doesn't run correctly under VMWare and I
have not seen any lack of performance or anything compared to a physical
machine. Maybe if enough of us push to have the VMWare Tools developed
and certified for use with VMWare that they might actually get started.

I might develop some sort of E-Petition for it, what you think?


Why not? I'm in the exact same position as you are with ESX & FreeBSD.
Hence I'd love to have VMWare Tools developed and certified for use
with FreeBSD. Actually, I'd really like to see VMWare Server and
Player certified for FreeBSD i386 and amd64.


VMWare is great stuff, I use and support all of it, but as a company they have 
a bit of Fortune 500 tunnel vision.  Their pricing is geared towards nickel 
and diming large enterprises and their software support is geared too much 
towards Windows.


Hyper-V will cut deeply into their market and they might regret being too 
MS-centered.  I already have potential customers asking "Gee, what about 
Hyper-V.  It's a lot cheaper than ESX and we don't really care about non-MS..."


Pricing ESX enterprise (with all the bells & whistles, including vmotion, 
virtual center, HA, etc) at around US$500 per node and providing better 
support for FreeBSD and other alternative OSes would go a LONG way towards 
long-term competitiveness.

___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"


What periodic process strips executable permissions?

2008-02-25 Thread Shawn Barnhart
I wrote a shell script to email me the output of ntpdc -p and put it in my 
crontab.


It works for a week, and at some point over the weekend my script loses its 
executable permissions for me (user) but not for group or other.


Is there a FreeBSD periodic job that runs periodically and removes executable 
permissions?  I'm pretty sure I didn't make it SUID.  It was rwxr-xr-x and 
there are other scripts in the same directory rwxr--r-- that don't lose their 
permissions.


I ran through the scripts in /etc/periodic but didn't see anything that made 
sense as a culprit.



___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"


Re: Clock stabilization in VMWare hosted machines?

2008-02-10 Thread Shawn Barnhart

Gelsema, P (Patrick) wrote:

I am having problems with the vmware tools. It seems to die, everytime
when I start it or rebooted.

Feb 10 13:22:25 wolverine kernel: pid 566 (vmware-checkvm), uid 0: exited
on sig
nal 12 (core dumped)

$ uname -a
FreeBSD wolverine.superhero.nl 7.0-RC2 FreeBSD 7.0-RC2 #0: Sun Feb 10
13:01:07 CET 2008
[EMAIL PROTECTED]:/usr/obj/usr/src/sys/WOLVERINE  i386


  
I'm running 6-STABLE and thusfar I haven't had an issue, but I've only 
had them installed for a brief period.


My long-term experience with VMWare has been that FreeBSD is in the 
"barely supported" category by VMWare.  The guest OS tools are designed 
to help the VMWare host snoop on the OS internals to better manage 
resources like memory, so there's a chance that the tools you're running 
are referencing internals changed from 6.x to 7.x.


That'd be my guess.  You may want to post over on VMWare's forums and 
find out if others have this problem or if 7.x is even considered usable 
with the tools.



___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"


Re: Clock stabilization in VMWare hosted machines?

2008-02-09 Thread Shawn Barnhart

Peter Boosten wrote:

In my /boot/loader.conf:

kern.hz="100"



I rebuilt the kernel with options HZ=100 and this seems to "fix" it -- 
ntpd sync'd immediately and the clock does not appear to drift.


When you installed the tools, did they require X to be operational?  I'm 
not planning on running X.


___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"


Clock stabilization in VMWare hosted machines?

2008-02-09 Thread Shawn Barnhart
I know this is a major nuisance, but I can't remember how I dealt with 
it in the past.  My most recent stab at using ntpd with minpoll 4 
polling of a local ntp time source isn't working, the clock drift 
prevents any sync from happening (but I'll admit not trying some of the 
more aggressive time adjustment options to ntpd).


VMWare's documentation and support leans pretty heavily toward Linux and 
I'm not finding a decent recommendation from them.


I rebuilding a kernel with options HZ=100 to see if that makes a 
difference, not sure why I remember that helping, but any other 
strategies known to work?


___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"


Re: Buildworld for slow system on faster system

2008-01-23 Thread Shawn Barnhart

Brent Jones wrote:

What you propose works fine, but you should have the /usr/src and
/usr/obj directories locally stored on your fast machine for the builds.
Then have your slow machine nfs mount /usr/src and /usr/obj from your
fast machine, and simply do your make installs from your slow system.
  
Thanks.   I'm assuming on the slow install system I have to keep 
/usr/src and /usr/obj mounted from the fast build box on their "correct" 
directories (ie, over the local system's /usr/src and /usr/obj) 
otherwise I'll have problems.


As the other person who replied pointed out, nfs mounting /usr/src and 
/usr/obj from the slow system to the fast system would nullify a lot of 
the speed advantage (although it is 100% GigE), but would let me keep 
the working /usr/src and /usr/obj on the slow machine, which is the one 
that is actually getting used.


On a somewhat related question, which FreeBSD install choice gives me 
enough development tools to "make installworld" if I have /usr/src and 
/usr/obj nfs mounted from a full source system but doesn't actually 
burden the system in question with unnecessary components?





___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"


Re: Buildworld for slow system on faster system

2008-01-23 Thread Shawn Barnhart

[EMAIL PROTECTED] wrote:

It is /usr/obj you want to mount across, though it
will work.  Assuming they're similar enough
versions.  Also assure yourself that their respective
/etc/make.conf (and/or /etc/src.conf) files are
essentially identical, or you will great sorrows
have.

  
I'm not that much of a system hacker, so I doubt either will get 
edited.  I don't even have a src.conf on my 6_STABLE system, but 
regardless, it won't be an issue to keep them in sync.



If /usr/obj is nfs mounted on the _build_ machine
it will slow down your build times absurdly, unless
your network (minus overhead) is nearly as fast as
your HDD controller.
  


It's gigabit, so it'd be about as good as it can get.  It sounds like 
the thing to do is just do the builds on the fast machine, and then 
mount THAT machine's /usr/src and /usr/obj on the slow machine when it 
comes time to installworld.


___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"


Buildworld for slow system on faster system

2008-01-23 Thread Shawn Barnhart
My primary FreeBSD box is a Dual P3 700 Mhz, which is dandy for my 
console mode server usage but kind of blows for buildworld and kernels 
when I want them done a timely fashion.  I'd like to do it in a 
dual-proc VM on my quad core workstation, where it gets done a lot faster.


Is there any documentation for doing buildworld on a faster system for a 
slower system?  Can I just mount the slow system's /usr/src on a 
mountpoint on the faster system, do the buildworld and buildkernel, and 
then run the installworld and installkernel as per normal on the slow 
system?  It sounds too easy, so it probably is..


What about ports?  I can usually tolerate the ports build times, so its 
not a big deal, although sometimes the dependencies and larger packages 
can be toe-tappers as well.


___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"


Re: Trendnet TEG-PCITXR hw rev 3.0 & re driver

2008-01-15 Thread Shawn Barnhart

NetOpsCenter wrote:
I have 5 of the nic cards TEG PCI TXR running on Various  versions of  
7. *  and 8.* of FreeBSD.
There was somebody on the weekind also having trouble with version 6.* 
Same as your problem.

Try switch to 7.*
Now I see why -- the code looks way overhauled recently (like hours ago) 
and the RELENG_6 version doesn't have any of these fixes.


Anyone know if or when they will be MFC'd?
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"


Re: Trendnet TEG-PCITXR hw rev 3.0 & re driver

2008-01-15 Thread Shawn Barnhart

NetOpsCenter wrote:
I have 5 of the nic cards TEG PCI TXR running on Various  versions of  
7. *  and 8.* of FreeBSD.
There was somebody on the weekind also having trouble with version 6.* 
Same as your problem.

Try switch to 7.*
At this point, switching to 7.* isn't really an option and its loads 
easier to just switch to a better NIC.


Out of curiosity, what does the 7.* driver do that the 6.* doesn't and 
why hasn't it been backported?


___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"


Re: Trendnet TEG-PCITXR hw rev 3.0 & re driver

2008-01-14 Thread Shawn Barnhart

Shawn Barnhart wrote:
Went looking for a cheap GigE PCI card to use with a 6.2 
(6.3-PRERELEASE) stable system, and brought this home due to 
Microcenter not stocking Intel cards and me not wanting to wait for 
Newegg to ship one to me.
Card works after switching slots on the motherboard, but I got "re0: 
watchdog timer" errors on reboot and no traffic on the card.  I 
unplugged and replugged the NIC cable at the switch end and I get 
traffic now, but I have a proper Intel card on order from Newegg.  I 
don't trust this one.





___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"


Trendnet TEG-PCITXR hw rev 3.0 & re driver

2008-01-14 Thread Shawn Barnhart
Went looking for a cheap GigE PCI card to use with a 6.2 
(6.3-PRERELEASE) stable system, and brought this home due to Microcenter 
not stocking Intel cards and me not wanting to wait for Newegg to ship 
one to me.


I have the re driver built into the kernel, but this particular card is 
not recognized even though its supposed to be based on the Realtek 8169 
chipset.  Realtek has a linux driver download (of unknown quality) and 
that source has a mod time of 2/06, so I'm guessing it's not a 
reinvention of the wheel or a super-significant change.


Anyone track these things close enough to know if there's something 
significant changed in this chipset family?



___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"


Re: Postfix with Cyrus SASL

2008-01-10 Thread Shawn Barnhart

Paul Schmehl wrote:

It should, because it calls this:

.if defined(WITH_SASL2)
LIB_DEPENDS+=   sasl2.2:${PORTSDIR}/security/cyrus-sasl2
POSTFIX_CCARGS+=-DUSE_SASL_AUTH -DUSE_CYRUS_SASL 
-I${LOCALBASE}/include -I${LOCALBASE}/include/sasl

POSTFIX_AUXLIBS+=   -L${LOCALBASE}/lib -lsasl2 -lpam -lcrypt
.endif

Yes, you need to install saslauthd, however, if you checked the OPTION 
when you installed Postfix, it's most likely already installed.  You 
*also* need to enable saslauthd in /etc/rc.conf:


[EMAIL PROTECTED] /usr/ports/mail/postfix]# grep sasl /etc/rc.conf
saslauthd_enable="YES"
saslauthd_flags=" -a pam -n 2"

(This uses /etc/passwd through pam, btw.)

Look at /usr/local/etc/rc.d/saslauthd.sh for the options and flags 
available or read man (8) saslauthd.




Either I'm totally fubar, or the ports snapshot I have is braindead as I 
did select the SASL option when I built postfix and I have sasl libs in 
/usr/local/lib and /usr/local/lib/sasl2 but none of the other sasl 
components are installed.  No saslauthd in /usr/local/etc/rc.d, no 
manpage, just libraries mentioned above, and my postfix smtpd does 
appear to have a sasl library run-time dependency per ldd.


Is the better fix to manually re-install the same Cyrus sasl port or 
deinstall both it and postfix and rebuild postfix with the sasl option 
and hope I get a complete build?





___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"


Re: Postfix with Cyrus SASL

2008-01-10 Thread Shawn Barnhart

Michal F. Hanula wrote:

Your postfix is trying to use saslauthd, which usually listens on
/var/run/saslauthd/mux. The right way to fix this depends on whether
you want to use saslauthd and the place you store your e-mail user data.
  
I want authentication against /etc/passwd (ultimately), not using 
sasldb2.db.


There is no /var/run/saslauthd/mux, and saslauthd doesn't appear 
installed -- I'm getting the impression that selecting "Cyrus-SASL" in 
the make config dialog box for the Postfix port doesn't completely 
install cyrus-sasl components.


I'm guessing the solution is to completely install the cyrus-sasl2 port 
to enable the use of saslauthd.  Yes?  Or am I way off?

___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"


Postfix with Cyrus SASL

2008-01-09 Thread Shawn Barnhart
I'm running a recently built 6.2 stable build (which uname calls 
6.3-PRERELEASE) and
Postfix built from ports with the Cyrus SASLv2 option.  Postfix works 
fine, including TLS but SASAL authentication logs "file not found" errors.


Jan  9 17:14:10 postfix postfix/smtpd[48488]: warning: SASL 
authentication failure: cannot connect to saslauthd server: No such file 
or directory
Jan  9 17:14:10 postfix postfix/smtpd[48488]: warning: 
unknown[192.168.1.152]: SASL LOGIN authentication failed: generic failure


I'm not sure which file or directory is missing.  The docs on this are 
sketchy, most of what is listed is way out of date, and the most up to 
date docs, http://www.postfix.org/SASL_README.html isn't terribly 
platform specific.


Its not entirely clear if I need other SASL components; the entire Cyrus 
SASL package appears installed.



___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"