headers
Octave Orgeron | 2 May 2007 07:36
Picon
Favicon

Re: Developing an Open source data center automation toolset. (Initially an inventory system)

Hi Everyone,

I think this could be a very interesting project. Please outline the
scope of the project and some goals. I'll take a look at the link below
to see how to get a project going under our community. 

Octave
--- Eric Boutilier <Eric.Boutilier@...> wrote:

> On Fri, 27 Apr 2007, Brian Gupta wrote:
> > ...
> > 
> > What is invloved in starting a project?
> 
> This is in transition, but a new process is almost final (as
> of today actually): `OGB/2007/001 Project Instantiation Policy'
>
http://mail.opensolaris.org/pipermail/ogb-discuss/2007-April/000414.html
> 
> In short, here's all you do. (Note, this presumes viability of the
> Sys Admin
> Community Group in the new structure, which might not be a given[1])
> You
> publicly propose your project to the Sys Admin Community Group
> leaders,
> they publicly review/approve, and then they submit the project for
> instantiation. (If the Sys Admin Community Group can't sponsor, due
> to lack
> of viability or whatever, you'd have to seek another logical
> Community
(Continue reading)

Permalink | Reply | 3 comments |
headers
Brian Gupta | 2 May 2007 09:46
Picon

Re: Developing an Open source data center automation toolset. (Initially an inventory system)

>
> I think this could be a very interesting project. Please outline the
> scope of the project and some goals. I'll take a look at the link below
> to see how to get a project going under our community.

I now know what is required to get a project started. The issue now is that
I don't want to propose a project until we have assembled a small project
team. (As I explained earlier, I can't do this on my own).

Please post to this thread if you are willing to help out with coding and
design.

As for the goals, check out: http://docs.google.com/Doc?id=df6hcjb2_3dfq7t3

-Brian
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.opensolaris.org/pipermail/sysadmin-discuss/attachments/20070502/3eaac08b/attachment.html
Permalink | Reply | 2 comments |
headers
Dagobert Michelsen | 2 May 2007 10:05
Picon
Favicon
Gravatar

Re: Developing an Open source data center automation toolset. (Initially an inventory system)

Hi,

Von "Brian Gupta" <brian.gupta@...> (Wed, 2 May 2007 10:46:49 -0400):
> I now know what is required to get a project started. The issue now is that
> I don't want to propose a project until we have assembled a small project
> team. (As I explained earlier, I can't do this on my own).
> 
> Please post to this thread if you are willing to help out with coding and
> design.

I have done some programming against libdevinfo and libhbaapi paired
with device identification with obp pathes (like in Infodoc 21216). If you
think this would be helpful I'll contribute to the project.

Best regards

  -- Dagobert Michelsen
Permalink | Reply | 1 comment |
headers
Brian Gupta | 2 May 2007 11:09
Picon

Re: Developing an Open source data center automation toolset. (Initially an inventory system)

This might prove useful down the road, but what we really need to get
started is a Ruby and Ruby/Rails developer(s). (Or a programmer who is
familiar with OO programming, and has time and inclination to learn Ruby).

For now I plan to use basic Solaris executables to gather as much
information as I can. It is quite possible that at some point we will run
into an issue where we won't be able to get the information we need, and may
need to develop a remote agent. At that time your experience would be
invaluable. You are welcome to join in the mean time as a collaborator for
design.

The reason I am choosing Ruby, is that one of the key design goals is to be
fully object oriented. (Ruby programming is 100% OO) Also, I would like to
avoid using multiple inheritance. Ruby doesn't support it, so that enforces
this. (I was thinking about using Python, but lean towards ruby for this
reason) I am also going for an agile development methodology, and plan to
allow design goals to change rapidly as we go along. (I expect as people
roll this out into testing, we will have to continually update as we go
along.) I also have the feeling that there are many Ruby developers eager to
join a project.

Cheers,
-Brian

On 5/2/07, Dagobert Michelsen <dam@...> wrote:
>
> Hi,
>
> Von "Brian Gupta" <brian.gupta@...> (Wed, 2 May 2007 10:46:49
> -0400):
(Continue reading)

Permalink | Reply | 0 comments |
headers
Vic Engle | 7 May 2007 08:29
Picon

Sol 10 x86 Update 3 boot hangs following installation

I have been having a problem installing and configuring Solaris 10 update 3 on Sun x86 servers. The problem
has occurred on v40z, X2100 and x4200 systems.

Basically the install itself goes smoothly and following the initial reboot, everything looks good. The
system will boot up and configure networking with no problems. The problem seems to occur when the system
is rebooted following some post install configuration which is routine configuration of nis, the
automounter and some additional nfs mounts. The system continues to work as expected during this
configuration but when the systems is rebooted it doesn't proceed into the boot process far enough to
allow even a console login.

In troubleshooting I have discovered that if I remove /etc/defaultrouoter the system will boot. Once
booted if I try to manually add the default route the route command hangs. 

Has anyone seen this? I'll include the output of truss captured while trying to manually add the route. I
follow the same procedure to configure Sol 10 sparc update 3 with no problem.

Thanks,
Vic

[root <at> sunx4200-shu01--->]truss -f route add default 10.60.181.1
725:    execve("/sbin/route", 0x08047A80, 0x08047A94)  argc = 4
725:    resolvepath("/lib/ld.so.1", "/lib/ld.so.1", 1023) = 12
725:    resolvepath("/sbin/route", "/sbin/route", 1023) = 11
725:    sysconfig(_CONFIG_PAGESIZE)                     = 4096
725:    xstat(2, "/sbin/route", 0x08047878)             = 0
725:    open("/var/ld/ld.config", O_RDONLY)             Err#2 ENOENT
725:    xstat(2, "/lib/libsocket.so.1", 0x08047110)     = 0
725:    resolvepath("/lib/libsocket.so.1", "/lib/libsocket.so.1", 1023) = 19
725:    open("/lib/libsocket.so.1", O_RDONLY)           = 3
725:    mmap(0x00010000, 4096, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_ALIGN, 3, 0) = 0xFEFC0000
(Continue reading)

Permalink | Reply | 1 comment |
headers
David Blasingame | 7 May 2007 09:16
Picon

Sol 10 x86 Update 3 boot hangs following installation

It is kind of difficult to see what you are hanging on with the truss. 
You could potentially use dtrace to provide you more insight.

example script:
************ hang.d ******************

#!/usr/sbin/dtrace -s

#pragma D option flowindent

pid$target:::{
    self->trace=1;
}
fbt:::
/self->trace/
{
}
*****************************************

hang.d -c route add default 10.60.181.1

This will produce a tone of output, so you will need to capture the 
console with script.  Note when you hang and the area of the functions 
you are in.  It may provide you better insight into the problem.

I've done a similiar experiment with NFS hangs.  Check out my example in 
my blog.  http://blogs.sun.com/blaze/entry/dtrace_to_find_what_is

Dave
(Continue reading)

Permalink | Reply | 0 comments |
headers
Atul Vidwansa | 7 May 2007 11:06
Picon

Booting OpenSolaris from Linux PXE server?

Hi,
   I have a DHCP server running on linux(machine A) in my lab and an
x86 opensolaris(b54) machine(B). I wanted another x86 machine(C) to
boot in Solaris over the network. Is it possible:
1) To create solaris PXE network install server on Linux(A)?
2) To make the x86 machine(C) boot from available Solaris install
server(B) keeping DHCP still on linux server(A)?
3) If yes, how do I redirect machine C to install from machine B?

All the three machines are in the same subnet. I tried all the
techniques mentioned in "Solaris 10 11/06 Installation Guide:
Network-Based Installations", but could not get machine C to boot from
B. On initiating a network boot, machine C hangs at error signal "Bad
PBR", "boot file not found".
 Any clues?

Thanks,
_Atul

--

-- 
Atul Vidwansa
Cluster File Systems Inc.
http://www.clusterfs.com
Permalink | Reply | 0 comments |
headers
Atul Vidwansa | 7 May 2007 12:35
Picon

Re: Booting OpenSolaris from Linux PXE server?

Never mind, I found the simple solution to redirect machine being
installed to correct server by modifying dhcpd.conf on the DHCP
server. All I needed to do is provide correct values for following
directives in /etc/dhcpd.conf

next-server<>;
option tftp-server-name "<>";
filename "<>";

Cheers,
-Atul

On 5/7/07, Atul Vidwansa <atulvid@...> wrote:
> Hi,
>    I have a DHCP server running on linux(machine A) in my lab and an
> x86 opensolaris(b54) machine(B). I wanted another x86 machine(C) to
> boot in Solaris over the network. Is it possible:
> 1) To create solaris PXE network install server on Linux(A)?
> 2) To make the x86 machine(C) boot from available Solaris install
> server(B) keeping DHCP still on linux server(A)?
> 3) If yes, how do I redirect machine C to install from machine B?
>
> All the three machines are in the same subnet. I tried all the
> techniques mentioned in "Solaris 10 11/06 Installation Guide:
> Network-Based Installations", but could not get machine C to boot from
> B. On initiating a network boot, machine C hangs at error signal "Bad
> PBR", "boot file not found".
>  Any clues?
>
> Thanks,
(Continue reading)

Permalink | Reply | 0 comments |
headers
Doug | 7 May 2007 16:26
Picon
Favicon

Risks involved with having mixed disks in RAID-5

Hi All,

I am in the process of adding 6 more disks to my existing 4 disk RAID-5. Unfortunately 4 out of the 6 new disks
are SEAGATE instead of the existing SUN72G. Are there any known issues/concerns with mixing the type of
disks contained in a RAID-5? I see that the cylinder block constraints vary so copying the vtoc from the SUN
disks to the SEAGATE disks will not work, but I'm assuming that I will be ok since I manually formatted the
disks so technically their vtocs are there.

All suggestions and/or comments are appreciated.

DY

 
This message posted from opensolaris.org
Permalink | Reply | 1 comment |
headers
Albert Widjaja Tedjadiputra | 7 May 2007 21:42
Picon
Favicon

Installing Plasmon UDO30D-SE with QStar HSM Software on Solaris 5.8

Dear OpenSolaris member,

I'm having difficulties in Installing Plasmon UDO30D-SE SCSI Device (BluRay Drive) under my Solaris OS,
and it should be able to be managed by using QStar HSM for the storage & archival software.

I have successfully installed the QStar software, however when trying to verify the daemons are running,
the following errors are shown:

here's the error log:

  root <at> audi # ./qscsi_ping
  audi: SCSI server is running
  root <at> audi # ./jbping
  audi: JB server is not running, Rpc portmapper error
  root <at> audi # ./vlping
  audi: VL server is not running, Rpc portmapper error
  root <at> audi # ./mmping
  audi: MM server is not running, Rpc portmapper error
  root <at> audi # ./jbschedcheck -a
  jbschedcheck: localhost: JB: create_client:
  q_get_cached_service_port:
  Rpc portmapper error

I'm looking forward for the reply.

thanks

 
This message posted from opensolaris.org
(Continue reading)

Permalink | Reply | 0 comments |

Gmane