AF:
NF:0
PS:10
SRH:1
SFN:
DSR:
MID:<20090724113316.70b32fe5@ripper.onstor.net>
CFG:
PT:0
S:andy.sharp@onstor.com
RQ:
SSV:mail.onstor.net
NSV:
SSH:
R:<jan.seidel@onstor.com>,<Arnaldo.Roldan@onstor.com>,<larry.scheer@onstor.com>,<maxim.kozlovsky@onstor.com>,<Ron.Whitham@onstor.com>,<chris.vandever@onstor.com>,<e.Jin@onstor.com>,<henry.lau@onstor.com>,<Ram.Upendran@onstor.com>
MAID:1
X-Sylpheed-Privacy-System:
X-Sylpheed-Sign:0
SCF:#mh/Mailbox/sent
RMID:#imap/andys@onstor.net@exch1.onstor.net/INBOX	0	2779531E7C760D4491C96305019FEEB52AD707965A@exch1.onstor.net
X-Sylpheed-End-Special-Headers: 1
Date: Fri, 24 Jul 2009 11:35:14 -0700
From: Andrew Sharp <andy.sharp@onstor.com>
To: Jan Seidel <jan.seidel@onstor.com>
Cc: Arnaldo Roldan <Arnaldo.Roldan@onstor.com>, Larry Scheer
 <larry.scheer@onstor.com>, Maxim Kozlovsky <maxim.kozlovsky@onstor.com>,
 Ron Whitham <Ron.Whitham@onstor.com>, Chris Vandever
 <chris.vandever@onstor.com>, Danqing Jin <e.Jin@onstor.com>, Henry Lau
 <henry.lau@onstor.com>, Ram Upendran <Ram.Upendran@onstor.com>
Subject: Re: engineering workstation remix
Message-ID: <20090724113514.3edfb0bc@ripper.onstor.net>
In-Reply-To: <2779531E7C760D4491C96305019FEEB52AD707965A@exch1.onstor.net>
References: <BB375AF679D4A34E9CA8DFA650E2B04E0C7DB1A6@onstor-exch02.onstor.net>
	<2779531E7C760D4491C96305019FEEB52AD707965A@exch1.onstor.net>
Organization: Onstor
X-Mailer: Sylpheed-Claws 2.6.0 (GTK+ 2.8.20; x86_64-pc-linux-gnu)
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

We need to sign up a guinea pig, Jan do you want the honor?  Rendell?
Adding some additional folks to the list in case they might want to
volunteer.  The duties require that you whine if something makes you
frown.  About the workstation usage.

The NFS thing is either fixed already and we just need the latest
package(s), or it will be soon, so we should just hang tight on that
one.  Jan thanks for tracking it down.


 On Fri, 24 Jul 2009 11:21:18 -0700 Jan Seidel
<jan.seidel@onstor.com> wrote:

> Hi,
> 
> The tests pass now, however there were some modifications required:
> 
> FYI: I had to modify the following in the sudoers file to keep all
> environment variables when running "runtests" with sudo:
> Defaults            !env_reset Defaults            env_delete -=
> "PERL5LIB"
> 
> I also modified the default nfs mount options for the tests to
> include the "sec=sys" option. All tests that don't use
> "testutils->mount_share" still need to be modified to include the
> option. I don't know why it's required; Larry mentioned this is a bug
> in the Debian version. The question is: Do we want to resolve this
> problem or just modify all tests to explicitly use "sec=sys"?
> 
> Regards,
> Jan
> 
> ________________________________
> From: Arnaldo Roldan
> Sent: Friday, July 24, 2009 7:22 AM
> To: Jan Seidel; Larry Scheer; Maxim Kozlovsky
> Cc: Ron Whitham; Andy Sharp
> Subject: engineering workstation remix
> Importance: High
> 
> Morning Gentlemen,
> 
> Due to pending changes we need to put this on the front burner. What
> I need from you is a) the list of items that need to be fixed and who
> can fix them and b) a prioritized list of engineers that need their
> systems upgraded, starting with a candidate who is willing to be the
> guinea pig. This person will need to be willing to give us feedback
> on a timely basis so that we may fine tune the image before roll out.
> Unless things change we have two weeks in which to do this or else
> you will have to continue this project with the new IT organization.
> 
> Thanks,
> 
> Arnaldo
> 
> 
> 
> ________________________________
> From: Jan Seidel
> Sent: Tuesday, July 21, 2009 6:43 PM
> To: Larry Scheer; Maxim Kozlovsky; Arnaldo Roldan
> Cc: Ron Whitham; Andy Sharp
> Subject: RE: engineering workstation
> 
> I fixed the compiler warnings and installed all missing Perl modules.
> Most of them were available as debian packages, only Test::Most and
> it's dependencies had to be installed using CPAN
> into /usr/local/share/perl/5.10.0/. The only modules that are now
> installed into $PERL5LIB are the ONStor packages which are required
> to startup the test environment:
> 
> engtest0:~# echo $PERL5LIB
> /n/lib/Linux/deb5.0/2.6.26-2-amd64/x86_64/perl/5.10.0/lib
> engtest0:~# ls -la $PERL5LIB
> total 24
> drwxr-xr-x 3 root root 1024 2009-07-21 16:48 .
> drwxr-xr-x 3 root root 1024 2009-07-21 16:48 ..
> drwxr-xr-x 4 root root 1024 2009-07-21 16:48 ONStor
> 
> What is still missing is an entry for QE_ROOT. It's required so that
> the tests can access the start script in $QE_ROOT/tst-labman/start.
> On the other clients we always checked it out to: /home/p4/nfx-test/
> 
> Shall we do it the same way on engtest0?
> 
> Apart from that most of the tests are running now (I temporarily set
> QE_ROOT to my home dir): # [Tue Jul 21 18:12:00 2009]
> _process_queue(): Running test file
> `/n/users/jseidel/private/jans-dev/nfx-tree/test/t/all/features/filesystem/conformance/functional/snapshot/001_create.t'
> 1..80 ok 1 - successfully retrieve info: vsvr=[G12R204-T1],
> vol=[g12r204-t1-vol1], vsvr_ip=[10.3.204.221] ok 2 - setting up the
> mount point at
> [/mnt/jseidel,gw--g12r204,vsvr--10.3.204.221,vol--g12r204-t1-vol1,test---t-all-features-filesystem-conformance-functional-snapshot-001_create.t,test_run_from--engtest0]
> ok 3 - snapshot [g12r204-t1-vol1_snap_001_create_t] is created ok 4 -
> snapshot [g12r204-t1-vol1_snap_001_create_t] is removed
> 
> I still have a few problems with other tests. I still need to figure
> this out.
> 
> Regards,
> Jan
> 
> ________________________________
> From: Larry Scheer
> Sent: Thursday, July 16, 2009 7:40 PM
> To: Jan Seidel; Maxim Kozlovsky; Arnaldo Roldan
> Cc: Ron Whitham; Andy Sharp
> Subject: RE: engineering workstation
> 
> Hi Jan,
>     Crypt is resolved by linking in /usr/lib/libcrypt.so. Somehow
> your build has redefined the LIBCRYPT macro so that it is no longer
> set to -lcrypt.
> 
> I will help you sort this out tomorrow.  You should not have gotten
> this error.
> 
> Larry
> 
> ________________________________
> From: Jan Seidel
> Sent: Thursday, July 16, 2009 6:22 PM
> To: Maxim Kozlovsky; Larry Scheer; Arnaldo Roldan
> Cc: Ron Whitham; Andy Sharp
> Subject: RE: engineering workstation
> 
> 
> I was fixing some of the compilation failures on the 64 bit machine,
> but now I got this error:
> 
> 
> 
> make[3]: Leaving directory
> `/homes/jseidel/jans-dev/nfx-tree/code/ssc-cluster'
> 
> cc is gcc
> 
> CFLAGS is -fPIC -fms-extensions -D__long64 -DEXT_UTIL
> -DDEFAULT_VENDOR_NAME="ONStor" -DDEFAULT_PROD_NAME="EXT_UTIL" -Wall
> -Wmissing-prototypes -Wno-pointer-sign -Werror -DEXTENDED_SECURITY
> -DNFX_KERBEROS -DNETEEE_FRAGMENT
> 
> gcc -fPIC -fms-extensions  -D__long64  -DEXT_UTIL
> -DDEFAULT_VENDOR_NAME=\"ONStor\" -DDEFAULT_PROD_NAME=\"EXT_UTIL\"
> -Wall -Wmissing-prototypes -Wno-pointer-sign   -Werror
> -DEXTENDED_SECURITY -DNFX_KERBEROS -DNETEEE_FRAGMENT
> -o ../../Build/test/opt/bin/rmc_client \
> 
>         ../../Build/test/opt/Objects/SSC/ssc-rmc/rmc_client.o ../../Build/test/opt/lib/librmc.so
> \
> 
>         -lrmc -lldap -Wl,-Bstatic -lssl -Wl,-Bdynamic -lcrypto -lm
> -lnsl  \
> 
>         -L../../Build/test/opt/lib -lcluster -lncm -lanpssc -le_log
> -lutils -lauthen -licucommon -licudata -lnfxnis -lacrypt -levent
> -lbsdcompat -lgenlib \
> 
>          -Xlinker -rpath
> -Xlinker /homes/jseidel/jans-dev/nfx-tree/Build/test/opt/lib
> 
> ../../Build/test/opt/lib/libauthen.so: undefined reference to `crypt'
> 
> collect2: ld returned 1 exit status
> 
> make[2]: *** [../../Build/test/opt/bin/rmc_client] Error 1
> 
> 
> 
> Is there a missing library that needs to be installed on this
> machine? I don't know what this "crypt" is.
> 
> 
> 
> Thanks,
> 
> Jan
> 
> 
> 
> -----Original Message-----
> From: Maxim Kozlovsky
> Sent: Thursday, July 16, 2009 10:59 AM
> To: Larry Scheer; Arnaldo Roldan
> Cc: Ron Whitham; Brian Stark; Andy Sharp; Jonathan Goldick; Jan Seidel
> Subject: RE: engineering workstation
> 
> 
> 
> Jan will work on the compilation failures. Could you please tell us
> what is the address of the test workstation again?
> 
> 
> 
> Max
> 
> 
> 
> >-----Original Message-----
> 
> >From: Larry Scheer
> 
> >Sent: Thursday, July 16, 2009 10:04 AM
> 
> >To: Arnaldo Roldan
> 
> >Cc: Maxim Kozlovsky; Ron Whitham; Brian Stark; Andy Sharp; Jonathan
> 
> >Goldick; Jan Seidel
> 
> >Subject: RE: engineering workstation
> 
> >
> 
> >It would be nice to get Jan to verify the perl installation is
> >complete,
> 
> >meaning  it will work with the automated test suite, or at least
> >work with
> 
> >me to get it verified.
> 
> >
> 
> >I don't know if the fact that we are unable to compile the x86 test
> 
> >utilities on this system preclude verifying the compatibility of
> >this box
> 
> >with automated test harness. There should be a suite of tests that
> >can be
> 
> >run to check out the perl components. I think this is the last major
> >thing
> 
> >development needs to do to certify this configuration.
> 
> >
> 
> >Jan, do you have some tests that can be run on engtest0 to shake out
> >the
> 
> >perl and tap installation?
> 
> >
> 
> >Once that is done I think we can give this project back to IT for
> 
> >deployment. As Andy pointed out the compile errors in the x86 test
> 
> >utilities just need to be fixed by development and should not stop
> 
> >deployment of this system. IMO.
> 
> >
> 
> >Larry
> 
> >
> 
> >________________________________________
> 
> >From: Arnaldo Roldan
> 
> >Sent: Thursday, July 16, 2009 9:16 AM
> 
> >To: Andy Sharp; Larry Scheer
> 
> >Cc: Maxim Kozlovsky; Ron Whitham; Brian Stark
> 
> >Subject: RE: engineering workstation
> 
> >
> 
> >Gentlemen,
> 
> >
> 
> >I wanted to get an update on where we stand with this. Is someone
> >working
> 
> >on the fixes Andy mentioned in his email? Are there other packages,
> >changes
> 
> >we need to make at the IT end?
> 
> >
> 
> >I would like to get this in working shape by the end of the month so
> >that
> 
> >we can start deploying stations in early Aug. After all Jonathan did
> 
> >indicate this as being a high priority for Engineering.
> 
> >
> 
> >Thanks,
> 
> >
> 
> >Arnaldo
> 
> >
> 
> >-----Original Message-----
> 
> >From: Andy Sharp
> 
> >Sent: Friday, July 10, 2009 11:05 PM
> 
> >To: Larry Scheer
> 
> >Cc: Maxim Kozlovsky; Arnaldo Roldan; Ron Whitham; Brian Stark
> 
> >Subject: Re: engineering workstation
> 
> >
> 
> >On Fri, 10 Jul 2009 19:40:54 -0700 Larry Scheer
> 
> ><larry.scheer@onstor.com> wrote:
> 
> >
> 
> >> Good news, I found out why a make release for the bobcats was not
> 
> >> working on this new workstation and I have a Makefile fix for it.
> 
> >>
> 
> >> The bad news is make test gets lots of compile errors due to the
> 
> >> nature of the64 bit user space on this system.
> 
> >>
> 
> >> Some compile errors are:
> 
> >>
> 
> >>  sm-eee/eee-desc.h:161: error: size of array
> 
> >> 'nfx_compileTimeAssertion' is negative.
> 
> >>
> 
> >> authen-lib.c:2736: error: passing argument 2 of 'fgetln' from
> 
> >> incompatible pointer type
> 
> >>
> 
> >> fs.c: In function 'ensure_dir_exists':
> 
> >> fs.c:704: error: cast to pointer from integer of different size
> 
> >>
> 
> >> ncm-api.c:900: error: cast from pointer to integer of different
> >> size
> 
> >>
> 
> >> rmc.c:2751: error: format '%d' expects type 'int', but argument 5
> >> has
> 
> >> type 'long unsigned int'
> 
> >>
> 
> >> There are a bunch of other format errors expecting type 'int' but
> >> has
> 
> >> type 'long unsigned int'. I think e_log is one of them plus some
> 
> >> cluster logging functions.
> 
> >
> 
> >These are all bugs that can be fixed easily, and should be fixed.  I
> 
> >went through something similar when building the prom code on a 64
> >bit
> 
> >machine:  the time has come.  Etcetera and so forth.
> 
> >
> 
> >> Just a quick not to let you know the latest status.
> 
> >
> 
> >A quick not?  Knot?
> 
> >
> 
> >Thanks for the update.  I'm very glad that we're making progess on
> 
> >this, largely due to the efforts of yourself and Arnaldo and Ron.
> 
> >With the occassional nudge from Max and me ~:^)
> 
> >
> 
> >Thanks guys.
> 
> >
> 
> >Cheers,
> 
> >
> 
> >a
> 
> >
> 
> >> Larry
> 
> >>
> 
> >> -----Original Message-----
> 
> >> From: Maxim Kozlovsky
> 
> >> Sent: Thursday, July 09, 2009 9:38 AM
> 
> >> To: Larry Scheer; Arnaldo Roldan; Ron Whitham; Andy Sharp
> 
> >> Cc: Brian Stark; Jonathan Goldick
> 
> >> Subject: RE: engineering workstation
> 
> >>
> 
> >> Does "make test" work?
> 
> >>
> 
> >>
> 
> >> >-----Original Message-----
> 
> >> >From: Larry Scheer
> 
> >> >Sent: Wednesday, July 08, 2009 6:49 PM
> 
> >> >To: Arnaldo Roldan; Ron Whitham; Andy Sharp
> 
> >> >Cc: Maxim Kozlovsky; Brian Stark; Jonathan Goldick
> 
> >> >Subject: RE: engineering workstation
> 
> >> >
> 
> >> >Everyone,
> 
> >> >   I just wanted to give you a progress update on the engineering
> 
> >> >workstation.
> 
> >> >
> 
> >> >I have all the packages installed and make install_toolchain
> >> >works on
> 
> >> >Debian systems. I am waiting for Andy to review my changes and I
> 
> >> >will check in code and packages.
> 
> >> >
> 
> >> >I was able to successfully build the cougar product (make
> 
> >> >PROD=cg ... and make release PROD=cg ...)
> 
> >> >
> 
> >> >I was able to build the nfx-tree for bobcat but a make release
> 
> >> >PROD=bc fails. There seem to be problems with running the bsdmake
> 
> >> >found in boot- tools with this hardware and Debian release. I am
> 
> >> >attempting to rebuild bsdmake and boot-tools for this platform
> >> >but I
> 
> >> >ran into compile errors.
> 
> >> >
> 
> >> >I will keep plugging away at this to see if I can get the bobcat
> >> >make
> 
> >> >release to work. I won't attempt to get a build of OpenBSD to
> >> >work.
> 
> >> >There are too many issues to bother trying that. Besides we have
> 
> >> >several systems that can build it successfully.
> 
> >> >
> 
> >> >I just thought you would like a progress report.
> 
> >> >
> 
> >> >Once I get the bobcat build to complete the next step would be
> 
> >> >having Jan or someone verify the automated test tools work. Which
> 
> >> >reminds me, I need to build an amd64 version of the tap library
> 
> >> >package and install it on engtest0.
> 
> >> >
> 
> >> >Larry
> 
> >> >
> 
> >> >-----Original Message-----
> 
> >> >From: Arnaldo Roldan
> 
> >> >Sent: Tuesday, June 30, 2009 4:59 PM
> 
> >> >To: Ron Whitham; Andy Sharp
> 
> >> >Cc: Larry Scheer; Maxim Kozlovsky
> 
> >> >Subject: RE: engineering workstation
> 
> >> >
> 
> >> >Larry,
> 
> >> >
> 
> >> >We await your prognosis. Thanks,
> 
> >> >
> 
> >> >Arnaldo
> 
> >> >
> 
> >> >-----Original Message-----
> 
> >> >From: Ron Whitham
> 
> >> >Sent: Tuesday, June 30, 2009 12:34 PM
> 
> >> >To: Andy Sharp; Arnaldo Roldan
> 
> >> >Cc: Larry Scheer; Maxim Kozlovsky
> 
> >> >Subject: RE: engineering workstation
> 
> >> >
> 
> >> >engtest0 has been reinstalled with an amd64 build.
> 
> >> >same login info as before:
> 
> >> >ssh://10.0.0.61
> 
> >> >root // onstor
> 
> >> >
> 
> >> >
> 
> >> >
> 
> >> >-----Original Message-----
> 
> >> >From: Andy Sharp
> 
> >> >Sent: Friday, June 26, 2009 6:18 PM
> 
> >> >To: Arnaldo Roldan
> 
> >> >Cc: Larry Scheer; Ron Whitham; Maxim Kozlovsky
> 
> >> >Subject: engineering workstation
> 
> >> >
> 
> >> >Hi folks,
> 
> >> >
> 
> >> >I just noticed that engtest0 is i386, but it needs to be amd64,
> 
> >> >which is the arch flavor that should be used on all x86_64
> >> >hardware,
> 
> >> >either from AMD or Intel.
> 
> >> >
> 
> >> >Not sure why I didn't notice this before, my bad really.  It will
> 
> >> >have to be reinstalled, sorry.
> 
> >> >
> 
> >> >a
> 
> >> >
> 
> >> >engtest0:~$ dpkg --print-architecture
> 
> >> >i386
> 
> >> >engtest0:~$
