setoolkit-developer

[SEToolkit-developer] Fwd: Patches for ZFS, e1000g/ipge/bnx, UDP

[Alex K. <ale...@gm...>]

Meant to CC this here, but got the wrong email address :|

---------- Forwarded message ----------
From: Alex Kiernan <ale...@gm...>
Date: 17 Oct 2007 14:07
Subject: Patches for ZFS, e1000g/ipge/bnx, UDP
To: orc...@or...
Cc: set...@so...


A number of small patches for orcallator/orca:

http://www.alexk1.demon.co.uk/e1000g-ipge-bnx.patch - add e1000g, ipge
and bnx interfaces to the list of known GbE interfaces

http://www.alexk1.demon.co.uk/zfs.patch - add ZFS to the list of
filesystem types which are tracked

http://www.alexk1.demon.co.uk/watch-udp.patch - add UDP kstats monitoring

--
Alex Kiernan


-- 
Alex Kiernan

Re: [SEToolkit-developer] Do we have any 'build' instructions yet?

[Dagobert M. <da...@ba...>]

Hi Jon,

Am 29.11.2007 um 19:48 schrieb Tankersley, Jon:
> I'm trying to build with the latest SVN release on a clean Solaris 10
> SPARC server.

Alex described this in
>  http://sourceforge.net/mailarchive/message.php?=20
> msg_id=3Dc461c0d10706130407v2a901afexb2be1a94589b84c4%40mail.gmail.com


> If you check out the tree from SVN, then you can generate the build
> framework using autoconf/automake/libtool using:
>
> autoreconf -if
>
> (we've been using autoconf-2.61, automake-1.10, libtool-1.5.22).
>
> then building like this will get you a 64 bit version:
>
> CPPFLAGS=3D-xarch=3Dgeneric64 LDFLAGS=3D-xarch=3Dgeneric64 ./configure
> make
> make install
>
> will get you a 64 bit setoolkit installed/built (assuming you've got
> Sun Studio somewhere in your PATH).


Best regards

   -- Dagobert

--=20
Dagobert Michelsen (Leiter IT)          Baltic Online Computer GmbH
Firmensitz:   Alter Markt 1-2, 24103 Kiel, Telefon: +49 431 54003-0
Gesch=E4ftsf=FChrer:        Erik Cickovskis, Amtsgericht Kiel, HRB 3756
"Of course computer servers don't need thrust, since they generally
don't go anywhere."  -- Comment in TR on new HP server turbine fans

Re: [SEToolkit-developer] Do we have any 'build' instructions yet?

[Alex K. <ale...@gm...>]

On Nov 30, 2007 6:57 AM, Dagobert Michelsen <da...@ba...> wrote:
> Hi Jon,
>
> Am 29.11.2007 um 19:48 schrieb Tankersley, Jon:
> > I'm trying to build with the latest SVN release on a clean Solaris 10
> > SPARC server.
>
> Alex described this in
> >  http://sourceforge.net/mailarchive/message.php?
> > msg_id=c461c0d10706130407v2a901afexb2be1a94589b84c4%40mail.gmail.com
>
>
> > If you check out the tree from SVN, then you can generate the build
> > framework using autoconf/automake/libtool using:
> >
> > autoreconf -if
> >
> > (we've been using autoconf-2.61, automake-1.10, libtool-1.5.22).
> >
> > then building like this will get you a 64 bit version:
> >
> > CPPFLAGS=-xarch=generic64 LDFLAGS=-xarch=generic64 ./configure
> > make
> > make install
> >
> > will get you a 64 bit setoolkit installed/built (assuming you've got
> > Sun Studio somewhere in your PATH).
>
>

Just added a doc/README.SVN so that this is written down somewhere more useful!

-- 
Alex Kiernan

Re: [SEToolkit-developer] Do we have any 'build' instructions yet?

[Tankersley, J. <jon...@ed...>]

Maybe that should be README.build-from-SVN

Here are my notes
Solaris 10, SPARC, Studio12
Dependencies - GNU autoconf, automake, libtool
Optional Dependencies - bison, m4
	autoreconf -if
	CC=3Dcc CXX=3DCC CPPFLAGS=3D-m64 -LDFLAGS=3D-m64 ./configure
--prefix=3D/usr/local/se
	make

The ReST stuff for the documentation is missing from my dependency list.

I had to add the CXX to get it to work.  The compiler was saying the
-xarch=3Dsparcv9 was deprecated.
I added the --prefix just to see what the install would do.

make[2]: Entering directory `/home/sysstat/Orca/svn/setoolkit/trunk/se'
cat doc/toc doc/preface doc/intro doc/one doc/two doc/three doc/four
doc/five doc/six doc/seven doc/eight doc/nine |  - doc/USERS.MANUAL.html
/bin/bash: -: command not found
make[2]: *** [doc/USERS.MANUAL.html] Error 127

So there is a dependency on ReST for the documentation build.=20

Packaging... That will be fun.
Should /usr/local be the default for the SE stuff?  That's what the
current configuration script handles.
Or should it be more like /opt/SEToolkit or /opt/RICHPse?

The doc/USERS.MANUAL.html should probably be last, not second in the
install.

ls -R /usr/local/se
/usr/local/se:
bin    lib    share

/usr/local/se/bin:
pw_example.sh           se                      workollator_example.sh

/usr/local/se/lib:
app-defaults  libsegui.la   libsegui.so   libsenet.la   libsenet.so

/usr/local/se/lib/app-defaults:
Segui

/usr/local/se/share:
doc        setoolkit

/usr/local/se/share/doc:
setoolkit

/usr/local/se/share/doc/setoolkit:
CHANGES            GUI.MANUAL         README.RELEASE
EXAMPLES.DOC       README.INSTALL     REPORTING.BUGS
FAQ                README.orcallator  USERS.MANUAL.html

/usr/local/se/share/setoolkit:
contrib       dio           examples      net_examples  toptool
COPYRIGHT     etc           include       perfmon       workollator

/usr/local/se/share/setoolkit/contrib:
dfstats     iops        kview       perfbar     psrsetstat  watch

/usr/local/se/share/setoolkit/contrib/dfstats:
author       df_stats.se  fs_class.se  mdf.se

/usr/local/se/share/setoolkit/contrib/iops:
author   iops.se

/usr/local/se/share/setoolkit/contrib/kview:
author     kviewx.se

/usr/local/se/share/setoolkit/contrib/perfbar:
author      perfbar.se

/usr/local/se/share/setoolkit/contrib/psrsetstat:
author         psrsetstat.se

/usr/local/se/share/setoolkit/contrib/watch:
author    watch.se

/usr/local/se/share/setoolkit/dio:
dio.se       dio_disk.se

/usr/local/se/share/setoolkit/etc:
init.d       se-mode.el   se-mode.elc  se_defines

/usr/local/se/share/setoolkit/etc/init.d:
mon_cm      monlog      percol      vader       va_monitor

/usr/local/se/share/setoolkit/examples:
aw.se                   mpvmstat.se             pw_log.se
bigbrother.se           mrvmstat.se             rvmstat.se
collisions.se           msacct.se               siostat.se
cpg.se                  msacct_off.se           swap.se
cpuchk.se               multi_meter.se          swapctl.se
cpu_meter.se            net.se                  syslog.se
cpus.se                 netmonitor.se           tcp_monitor.se
cpustat.se              netstat.se              tpry.se
ctlr.se                 netstatx.se             uname.se
darth.se                nfsmonitor.se           uptime.se
data                    nfsstat.se              util.se
disks.se                nproc.se                vader.se
dumpkstats.se           nx.se                   virtual_adrian.se
hostid.se               pea.se                  virtual_adrian_lite.se
infotool.se             peac.se                 vmmonitor.se
iomonitor.se            peer.se                 vmstat.se
iost.se                 percollator.se          webtune.se
iostat.se               perfmeter.se            workollator.se
kvmname.se              pry.se                  xio.se
limits.se               ps-ax.se                xiostat.se
live_test.se            ps-p.se                 xit.se
mini_cpu.se             pure_test.se            xload.se
mon_cm.se               pw.se                   zoom.se
monlog.se               pwatch.se

/usr/local/se/share/setoolkit/examples/data:
adrian          cpu_meter_icon  kmem            richp
xload_icon
adrian.bio      disk            mutex           richp.bio       zoom
blank           dnlc            network         swap
callog.root     drip.au         nfs_client      tpry.desc
cpu             inode           ram             xit_icon

/usr/local/se/share/setoolkit/include:
app_wrap_class.se      nfs_clnt.se            string.se
condvar.se             partinfo.se            sysdepend.se
ctlr_class.se          p_iostat_class.se      sysinfo_class.se
ctype.se               p_netstat_class.se     sys_kstat.se
devinfo.se             poll.se                syslog.se
dirent.se              proc.se                sys_swap.se
diskinfo.se            process_class.se       systeminfo.se
dlfcn.se               processor.se           tapeinfo.se
errno.se               process_rules.se       tcp_class.se
fcntl.se               pure_rules.se          tcp_rules.se
gui_defines.h          pure_rules_dutch.se    tcp_rules_dutch.se
gui_lib.se             pure_rules_swedish.se  time.se
kstat.se               p_vmstat_class.se      tiuser.se
live_rules.se          pwd.se                 types.se
logfile.se             rpc_clnt.se            uio.se
math.se                rstat_class.se         unistd.se
mib.se                 se.se                  uptime_class.se
mnt_class.se           se_trees.se            utmp.se
multi_rstat_class.se   signal.se              utsname.se
mutex.se               socket.se              vfs.se
ndd.se                 stat.se                vmstat_class.se
netdb.se               statvfs.se             wait.se
netif.se               stdio.se               workload_class.se
net_lib.se             stdlib.se

/usr/local/se/share/setoolkit/net_examples:
client.se  sample.se  server.se

/usr/local/se/share/setoolkit/perfmon:
perfmon            perfmon.se         p_vmstat_class.se  vmstat_class.se

/usr/local/se/share/setoolkit/toptool:
toptool     toptool.se

/usr/local/se/share/setoolkit/workollator:
explode.csh  pwgrab       README

-----Original Message-----
From: Alex Kiernan [mailto:ale...@gm...]=20
Sent: Friday, November 30, 2007 6:20 AM
To: Dagobert Michelsen
Cc: Tankersley, Jon; set...@li...
Subject: Re: [SEToolkit-developer] Do we have any 'build' instructions
yet?

On Nov 30, 2007 6:57 AM, Dagobert Michelsen <da...@ba...>
wrote:
> Hi Jon,
>
> Am 29.11.2007 um 19:48 schrieb Tankersley, Jon:
> > I'm trying to build with the latest SVN release on a clean Solaris=20
> > 10 SPARC server.
>
> Alex described this in
> >  http://sourceforge.net/mailarchive/message.php?
> > =
msg_id=3Dc461c0d10706130407v2a901afexb2be1a94589b84c4%40mail.gmail.com
>
>
> > If you check out the tree from SVN, then you can generate the build=20
> > framework using autoconf/automake/libtool using:
> >
> > autoreconf -if
> >
> > (we've been using autoconf-2.61, automake-1.10, libtool-1.5.22).
> >
> > then building like this will get you a 64 bit version:
> >
> > CPPFLAGS=3D-xarch=3Dgeneric64 LDFLAGS=3D-xarch=3Dgeneric64 =
./configure make=20
> > make install
> >
> > will get you a 64 bit setoolkit installed/built (assuming you've got

> > Sun Studio somewhere in your PATH).
>
>

Just added a doc/README.SVN so that this is written down somewhere more
useful!

--
Alex Kiernan

[SEToolkit-developer] Miscellaneous notes

[Tankersley, J. <jon...@ed...>]

I had to 'rebuild' a package so that I could hand it off to someone who
doesn't want to understand anything.....

I didn't start with the full SVN pkg yet, but I did notice some changes
to the old package that need to be made, and that would have to be check
with the new packaging scheme.
Or we COULD just build a separate package for Orca that has a dependency
on the SEToolkit package.

Binaries move to libexec
Bin/se script has that change included

To INCLUDE Orca we should somehow include the Orca and Adrian's
extensions.
Here's what I did to hack the new stuff into the old package.

Trunk/se/adrian copied to /opt/RICHPse/adrian
Trunk/orca/lib/SE copied to /opt/RICHPse/lib along with symlinks to have
3.4.x per the version information.
Orcallator startup scripts modified to have
-I/opt/RICHPse/adrian/include and -I/opt/RICHPse/lib/SE/3.4.x
Orcallator/orcallator.se.490 - svn version 490 of the orcallator script
included - orcallator.se is symlinked to this file.
Various binaries appropriately installed in the libexec directory.

This is enough to get the orcallator script to launch properly.

Any thoughts?

Re: [SEToolkit-developer] Some project questions

[Dagobert M. <da...@ba...>]

Hi Jon,

Von "Tankersley, Jon" <jon...@ed...> (Tue, 8 Jan 2008 23:37:31 -0600):
> How would one update the project/tasks in the project?

I can either enable you for this or you can just ask me ;-)

> I noticed that the mailing lists only show 2007....

Yes, because there was no mail in 2008 until now.

> I've taken some snapshots of the GUI output... Not 'great' pictures, but
> functional.  They are attached if you want to add them/refuse them.

Good to see you picking up open issues. I enabled screenshots and added
the pics you provided. However, the images must have screen dimensions
between 150x150 and 640x480. Maybe some of the smaller ones need
to be put in one file to fit.

> I can re-save them as PNG if you'd prefer.

That would be nice but its okay now too.

> I was looking at the 'fixing the references to /opt/RICHPse' item... I
> could probably work through that.

That would be really great.

> I've figured out what things can be readily tested for a test suite...
> Some that are one-shot or continuous until ^C and some GUI tests.
> I'd think a simple set of 2 'test' scripts - one for the one-shot and
> one for the one-shot X11 gui.  The continuous ones can't really be
> 'tested' well with a script, though with expect it could be possible to
> timeout....

Most continous programs issue warnings or terminate in case of
an error. The test suite should be executable with 'make check'
and should either succeed or return the faulted tests.


A late Happy New Year and best regards

  -- Dagobert

--
Dagobert Michelsen (Leiter IT)          Baltic Online Computer GmbH
Firmensitz:   Alter Markt 1-2, 24103 Kiel, Telefon: +49 431 54003-0
Geschäftsführer:        Erik Cickovskis, Amtsgericht Kiel, HRB 3756
"Of course computer servers don't need thrust, since they generally
don't go anywhere."  -- Comment in TR on new HP server turbine fans

Re: [SEToolkit-developer] Some project questions

[Tankersley, J. <jon...@ed...>]

=20

>-----Original Message-----
>From: Dagobert Michelsen [mailto:da...@ba...]=20
>Sent: Wednesday, January 09, 2008 2:59 AM
>To: set...@li...; Tankersley,=20
>Jon; Alex Kiernan
>Subject: Re: Some project questions
>
>Hi Jon,
>
>Von "Tankersley, Jon" <jon...@ed...> (Tue, 8 Jan=20
>2008 23:37:31 -0600):
>> How would one update the project/tasks in the project?
>
>I can either enable you for this or you can just ask me ;-)
>
>> I noticed that the mailing lists only show 2007....
>
>Yes, because there was no mail in 2008 until now.
Hmm... This was like the 4th email of the year...  I'll check again.

>
>> I've taken some snapshots of the GUI output... Not 'great' pictures,=20
>> but functional.  They are attached if you want to add=20
>them/refuse them.
>
>Good to see you picking up open issues. I enabled screenshots=20
>and added the pics you provided. However, the images must have=20
>screen dimensions between 150x150 and 640x480. Maybe some of=20
>the smaller ones need to be put in one file to fit.
>
>> I can re-save them as PNG if you'd prefer.
>
>That would be nice but its okay now too.

The sizes were 'out of the box' snapshots.... That's how they show up
when I ran them.  I'll see what I can do.

>
>> I was looking at the 'fixing the references to /opt/RICHPse'=20
>item... I=20
>> could probably work through that.
>
>That would be really great.
>
>> I've figured out what things can be readily tested for a=20
>test suite...
>> Some that are one-shot or continuous until ^C and some GUI tests.
>> I'd think a simple set of 2 'test' scripts - one for the=20
>one-shot and=20
>> one for the one-shot X11 gui.  The continuous ones can't really be=20
>> 'tested' well with a script, though with expect it could be possible=20
>> to timeout....
>
>Most continous programs issue warnings or terminate in case of=20
>an error. The test suite should be executable with 'make check'
>and should either succeed or return the faulted tests.

The continuous programs won't ever end....  It has to be killed via a
timer and a kill command.  I think the GNU test suite stuff can do that,
but I've never tried building them from scratch.  That will be fun.  And
then there are the 'monitoring' scripts.  I'm not sure how you could
build a validation test for the continuous ones or for the monitoring
scripts... As an interim having a simple test-cmdline and test-gui would
make sure the basics are working.  Until I can figure out GNU test
environment.

Re: [SEToolkit-developer] Miscellaneous notes

[Dagobert M. <da...@ba...>]

Hi Jon,

Von "Tankersley, Jon" <jon...@ed...> (Tue, 8 Jan 2008 16:12:27 -0600):
> I had to 'rebuild' a package so that I could hand it off to someone who
> doesn't want to understand anything.....
> 
> I didn't start with the full SVN pkg yet, but I did notice some changes
> to the old package that need to be made, and that would have to be check
> with the new packaging scheme.
> Or we COULD just build a separate package for Orca that has a dependency
> on the SEToolkit package.

That is my intention. The plan is to have two packages for SE Toolkit, one with
the old directory scheme and one new package in Blastwave style. The include
files all go into SE Toolkit. The Orca software is packaged in two parts, the
collector and the web presentation package.

If you want to take a look the repository is open at
  http://svn.blastwave.org/trac/browser/csw/trunk/utils/setoolkit
  http://svn.blastwave.org/trac/browser/csw/trunk/utils/orca

> Binaries move to libexec
> Bin/se script has that change included
> 
> To INCLUDE Orca we should somehow include the Orca and Adrian's
> extensions.
> Here's what I did to hack the new stuff into the old package.
> 
> Trunk/se/adrian copied to /opt/RICHPse/adrian
> Trunk/orca/lib/SE copied to /opt/RICHPse/lib along with symlinks to have
> 3.4.x per the version information.
> Orcallator startup scripts modified to have
> -I/opt/RICHPse/adrian/include and -I/opt/RICHPse/lib/SE/3.4.x
> Orcallator/orcallator.se.490 - svn version 490 of the orcallator script
> included - orcallator.se is symlinked to this file.
> Various binaries appropriately installed in the libexec directory.
> 
> This is enough to get the orcallator script to launch properly.
> 
> Any thoughts?

You should rebuild the software with autoconf and prefix set
and let 'make install' install the software in the proper locations.
Everything should work with this, otherwise it is a bug that needs
to be fixed.


Best regards

  -- Dagobert

--
Dagobert Michelsen (Leiter IT)          Baltic Online Computer GmbH
Firmensitz:   Alter Markt 1-2, 24103 Kiel, Telefon: +49 431 54003-0
Geschäftsführer:        Erik Cickovskis, Amtsgericht Kiel, HRB 3756
"Of course computer servers don't need thrust, since they generally
don't go anywhere."  -- Comment in TR on new HP server turbine fans

Re: [SEToolkit-developer] Miscellaneous notes

[Tankersley, J. <jon...@ed...>]

=20

>-----Original Message-----
>From: Dagobert Michelsen [mailto:da...@ba...]=20
>Sent: Wednesday, January 09, 2008 3:07 AM
>To: set...@li...; Tankersley,=20
>Jon; Alex Kiernan
>Subject: Re: Miscellaneous notes
>
>Hi Jon,
>
>Von "Tankersley, Jon" <jon...@ed...> (Tue, 8 Jan=20
>2008 16:12:27 -0600):
>> I had to 'rebuild' a package so that I could hand it off to someone=20
>> who doesn't want to understand anything.....
>>=20
>> I didn't start with the full SVN pkg yet, but I did notice some=20
>> changes to the old package that need to be made, and that would have=20
>> to be check with the new packaging scheme.
>> Or we COULD just build a separate package for Orca that has a=20
>> dependency on the SEToolkit package.
>
>That is my intention. The plan is to have two packages for SE=20
>Toolkit, one with the old directory scheme and one new package=20
>in Blastwave style. The include files all go into SE Toolkit.=20
>The Orca software is packaged in two parts, the collector and=20
>the web presentation package.
>
>If you want to take a look the repository is open at
>  http://svn.blastwave.org/trac/browser/csw/trunk/utils/setoolkit
>  http://svn.blastwave.org/trac/browser/csw/trunk/utils/orca

That was my starting point, but I didn't have time to guess as to the
intent.... I can now go back and try building a package the 'right' way.

>
>> Binaries move to libexec
>> Bin/se script has that change included
>>=20
>> To INCLUDE Orca we should somehow include the Orca and Adrian's=20
>> extensions.
>> Here's what I did to hack the new stuff into the old package.
>>=20
>> Trunk/se/adrian copied to /opt/RICHPse/adrian=20
>Trunk/orca/lib/SE copied=20
>> to /opt/RICHPse/lib along with symlinks to have 3.4.x per=20
>the version=20
>> information.
>> Orcallator startup scripts modified to have=20
>> -I/opt/RICHPse/adrian/include and -I/opt/RICHPse/lib/SE/3.4.x=20
>> Orcallator/orcallator.se.490 - svn version 490 of the orcallator=20
>> script included - orcallator.se is symlinked to this file.
>> Various binaries appropriately installed in the libexec directory.
>>=20
>> This is enough to get the orcallator script to launch properly.
>>=20
>> Any thoughts?
>
>You should rebuild the software with autoconf and prefix set=20
>and let 'make install' install the software in the proper locations.
>Everything should work with this, otherwise it is a bug that=20
>needs to be fixed.
>

Hmm... Regarding this one...
I was trying to build on a Solaris 10 x86....
Automake/autoconf/autoreconf/etc. all were non-cooperative.
Automake --add-missing got me most of the way but it was complaining
about Makefile.in and ltmain.sh.
Finally, after many different tries, I got it to build Makefile.in.  The
ONLY solution I could come up with for ltmain.sh was to copy from my
Solaris 10 SPARC build.  Then everything built and at first blush
appears to be okay.  I've not fully tested it yet.  We need yet another
test suite for JUST the SVN trunk.

Re: [SEToolkit-developer] Do we have any 'build' instructions yet?

[Dagobert M. <da...@ba...>]

Hi Jon,

Am 09.01.2008 um 23:26 schrieb Tankersley, Jon:
> Hmm...
> Didn't find the README.SVN.

Hmm... it is there, in r47:
  http://setoolkit.svn.sourceforge.net/viewvc/setoolkit/trunk/se/doc/ 
README.SVN?revision=47&view=markup

> I can't use the https, so I'm using the http link.  It appears to  
> work,
> but it definitely is missing some stuff.  I cleaned off my repository
> and re-checked out a copy.  Showed up this time.

You did update your repository, did you?

Best regards

   -- Dagobert

Re: [SEToolkit-developer] Should dumpkstat.se be about the same as kstat?

[Dagobert M. <da...@ba...>]

Hi Jon,

Am 14.01.2008 um 22:11 schrieb Tankersley, Jon:
> I'm working on the test portion and was trying to figure out a good  
> way
> to check the output..
>
> Dumpkstats.se on my sample system generates around 17000 lines.   
> Kstats
> generates 21000 lines.  Should they be closer?

I think it's ok. These numbers are from the old SE Toolkit 3.4.1
on Solaris 8 Sparc:

root@bopack8:/opt/RICHPse/examples > ../bin/se dumpkstats.se | wc -l
     5365
root@bopack8:/opt/RICHPse/examples > kstat | wc -l
     6493


Best regards

   -- Dago

Re: [SEToolkit-developer] Should dumpkstat.se be about the same as kstat?

[Alex K. <ale...@gm...>]

On Jan 14, 2008 10:11 PM, Dagobert Michelsen <da...@ba...> wrote:
> Hi Jon,
>
> Am 14.01.2008 um 22:11 schrieb Tankersley, Jon:
>
> > I'm working on the test portion and was trying to figure out a good
> > way
> > to check the output..
> >
> > Dumpkstats.se on my sample system generates around 17000 lines.
> > Kstats
> > generates 21000 lines.  Should they be closer?
>
> I think it's ok. These numbers are from the old SE Toolkit 3.4.1
> on Solaris 8 Sparc:
>
> root@bopack8:/opt/RICHPse/examples > ../bin/se dumpkstats.se | wc -l
>      5365
> root@bopack8:/opt/RICHPse/examples > kstat | wc -l
>      6493
>

I suspect it could do a better job, but the main job of dumpkstats
seems to be to dump the structures for transcribing to kstat.se.

-- 
Alex Kiernan

[SEToolkit-developer] Patches for ZFS, e1000g/ipge/bnx, UDP

[Tankersley, J. <jon...@ed...>]

=20
I was noticing that the patches are for the orcallator.cfg, not for
orcallator.se.

I'm guessing there's a need for a patch there too?  I've got a system
that someone else supports that has noticed that the e1000gN interfaces
in ifconfig shows are NOT being gathered by SEToolkit/orcallator.

Re: [SEToolkit-developer] Patches for ZFS, e1000g/ipge/bnx, UDP

[Tankersley, J. <jon...@ed...>]

Think I found it... My svn 'connection' was hosed so I wasn't up to
date.=20

-----Original Message-----
From: set...@li...
[mailto:set...@li...] On Behalf Of
Tankersley, Jon
Sent: Thursday, November 15, 2007 2:58 PM
To: set...@li...
Subject: [SEToolkit-developer] Patches for ZFS, e1000g/ipge/bnx, UDP

=20
I was noticing that the patches are for the orcallator.cfg, not for
orcallator.se.

I'm guessing there's a need for a patch there too?  I've got a system
that someone else supports that has noticed that the e1000gN interfaces
in ifconfig shows are NOT being gathered by SEToolkit/orcallator.

------------------------------------------------------------------------
-
This SF.net email is sponsored by: Microsoft Defy all challenges.
Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
SEToolkit-developer mailing list
SET...@li...
https://lists.sourceforge.net/lists/listinfo/setoolkit-developer

[SEToolkit-developer] Interesting problems with 3.4.1 and AMD

[Tankersley, J. <jon...@ed...>]

I'd expressed some concerns with the inclusion of the orcallator.se
file...
Orcallator is starting
Orcallator has started
chrdbt01# "/opt/RICHPse/orcallator/orcallator.se", line 387: Can't find
include file orca_p_vmstat_class.se
"/opt/RICHPse/include/live_rules.se", line 115: error: syntax error
"/opt/RICHPse/include/live_rules.se", line 115: fatal: Errors detected.
Exiting.
"stdin", line 0: error: syntax error
"stdin", line 0: fatal: Errors detected.  Exiting.

The necessary files aren't included, so I'd say not call it
orcallator.se, or we'd have to include more of Orca's patches in the
SEToolkit build.

But here's the REAL issue.
I've got a fairly new Sun system and the most recent package as posted
to Sourceforge doesn't work.
The version I built from an early 3.4.1 without the e1000 and some other
patches didn't work either, but actually RAN for a while before dying
with a Segmentation Fault in the disk section.

Running the test of the kvmname.se, it gets:
host# bin/se examples/kvmname.se
Warning: Cannot init kvm: Illegal seek
Warning: Kvm not initialized: Variables will have invalid values
Segmentation Fault
host# /etc/init.d/percol start
Warning: Cannot init kvm: Illegal seek
Warning: Kvm not initialized: Variables will have invalid values: Near
line 6156

It is acting like the AMD kernel stuff isn't available.
I've attached some information from the old build's run where it
segfaults in the disk section and a compressed copy of the pertinent
/var/adm/messages file from boot to 'up and running'.  Below is some
version info, and a prtdiag.  Unfortunately I don't have access to a
server of this type with a compiler.

host# /opt/RICHPse/bin/se -v
se - Version 3.4.1 (08:20 PM 02/28/07) for i386
host# /usr/platform/`arch -k`/sbin/prtdiag
System Configuration: Sun Microsystems Sun Fire X4600 M2
BIOS Configuration: American Megatrends Inc. 080012  04/19/2007
BMC Configuration: IPMI 1.5 (KCS: Keyboard Controller Style)

=3D=3D=3D=3D Processor Sockets =
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D

Version                          Location Tag
-------------------------------- --------------------------
Dual-Core AMD Opteron(tm) Processor 8220 CPU 1
Dual-Core AMD Opteron(tm) Processor 8220 CPU 2
Dual-Core AMD Opteron(tm) Processor 8220 CPU 3
Dual-Core AMD Opteron(tm) Processor 8220 CPU 4

=3D=3D=3D=3D Memory Device Sockets =
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D

Type    Status Set Device Locator      Bank Locator
------- ------ --- ------------------- --------------------
DDR2    in use 0   DIMM0               BANK0
DDR2    in use 0   DIMM1               BANK1

=3D=3D=3D=3D On-Board Devices =
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
 LSI serial-ATA #1
 Gigabit Ethernet #1
 Gigabit Ethernet #2
 ATI Rage XL VGA

=3D=3D=3D=3D Upgradeable Slots =
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D

ID  Status    Type             Description
--- --------- ---------------- ----------------------------
0   in use    PCI-X            PCIX SLOT0
1   available PCI-X            PCIX SLOT1
2   available other            PCIExp SLOT2
3   available other            PCIExp SLOT3
4   available other            PCIExp SLOT4
5   available other            PCIExp SLOT5
6   available other            PCIExp SLOT6
7   available other            PCIExp SLOT7

Re: [SEToolkit-developer] Do we have any 'build' instructions yet?

[Tankersley, J. <jon...@ed...>]

I'm trying to build with the latest SVN release on a clean Solaris 10
SPARC server.
I had to 'build' GNU m4, autoconf and automake and can't get autoconf to
run properly.

sysstat@tuluelab1 $
aclocal

sysstat@tuluelab1 $
autoconf
configure.ac:34: error: possibly undefined macro: AC_LIBTOOL_DLOPEN
      If this token and others are legitimate, please use
m4_pattern_allow.
      See the Autoconf documentation.
configure.ac:35: error: possibly undefined macro: AC_DISABLE_STATIC
configure.ac:36: error: possibly undefined macro: AC_PROG_LIBTOOL

sysstat@tuluelab1 $
automake
configure.ac: required file `./install-sh' not found
configure.ac: required file `./missing' not found
sex/gui/lib/Makefile.am:25: Libtool library used but `LIBTOOL' is
undefined
sex/gui/lib/Makefile.am:25:
sex/gui/lib/Makefile.am:25: The usual way to define `LIBTOOL' is to add
`AC_PROG_LIBTOOL'
sex/gui/lib/Makefile.am:25: to `configure.ac' and run `aclocal' and
`autoconf' again.
sex/gui/lib/Makefile.am: required file `./depcomp' not found
sex/net/lib/Makefile.am:35: Libtool library used but `LIBTOOL' is
undefined
sex/net/lib/Makefile.am:35:
sex/net/lib/Makefile.am:35: The usual way to define `LIBTOOL' is to add
`AC_PROG_LIBTOOL'
sex/net/lib/Makefile.am:35: to `configure.ac' and run `aclocal' and
`autoconf' again.
configure.ac:22: required file `./config.guess' not found
configure.ac:22: required file `./config.sub' not found
configure.ac:21: required file `config.h.in' not found