DSS 6.0.3 is out ! Check this out at http://dss.macosforge.org/

I’m not sure what this new version brings, few things seems to have changed. I hope for the best..
They claim to be 64 bits compatible on OsX, with some speed up changes. Don’t know what this will bring on Solaris or Linux.

Whatever, this version will not compile out of the tar.

Sverker Abrahamsson have made a patch for Linux here.
He also provide a pre-patched source code there.

This will not compile on Solaris 10
So here are 2 patches for Solaris 10. I compiled using the coolthread Sun GCC, but it should be OK with whatever GCC you use.
Note that I had to so something dirty with the Timezone detection, and I still don’t know what the impact can be on a production server. I’ll come back here if I see any trouble one on production.

I also have few errors (warnings, in fact), that should be OK when running but detected by the pre-processor :

RTCPUtilitiesLib/RTCPAPPNADUPacket.h: In destructor `NaduList::~NaduList()':
RTCPUtilitiesLib/RTCPAPPNADUPacket.h:205: warning: comparison between signed and unsigned integer expressions

FYI, I’m not using DSS on production right now, and DSS 6 will have to be heavily tested before it is validated.

So…

[EDIT]
Thanks to Stefan Parvu, I added in the patch the Solaris x86 things. So DSS now also compile on it (tested on Sun Fire X4100 Server with AMD Opteron(tm) Processor – gcc version 3.4.3 (csl-sol210-3_4-branch+sol_rpath))

[EDIT]

• dss-603-orig-to-solaris10-v3patch.gz
  Use this patch if you just download the plain source.
  gunzip the patch.
  untar the DSS source code and rename the directory « DarwinStreamingSrvr6.0.3-Source-Solaris10″.
  At the same level as the source directory, use « patch < dss-603-orig-to-solaris10-v3patch".
  go inside the DSS directory, then use « ./buildtarball 6.0.3 » to compile it.
  Your install package will be in something like DarwinStreamingSrvr6.0.3-SunOS.tar.gz
• dss-603-linux-to-solaris-v2.patch
  Use this if you start from the modified source code for Linux, provided by Sverker. Compile it the same way as above.

Then, Copy the tar file on your production server, untar it, and use the « ./Install » script for install in /usr/local.

I’m working on a modified installer script where you can change the root installation dir. Soon to come…

Of course many people are talking about how to set up a DHCP server… like :

apt-get dhcpd
server dhcpd start

And that’s it.

I do need more, first because I work on Solaris (10) and because, heh, I’m not doing basic stuffs.

First of all, you have 3 « stable » versions online you can download. Just tell me if you can find the differences between them. I just found that the latest is IPv6, not the other. I chosed to go with the latest, 4.0.0, but with IPv6 disabled.

Compilation was fine. I just had to add a LDFLAG so the compiled binary knows how to find the libcrypto shared object library, which is in /opt/sfw/lib on Solaris 10.

Then comes the configuration.

I went through some website talking about the failover config parameters. You can also find them in the Man pages. This is pretty simple.
On the master :

failover peer "dhcp-failover" {
# ce serveur est le maitre
primary;
# L'adresse IP local du serveur maitre
address 172.16.8.95;
# Le port IP local du serveur maitre
port 1520;
# L'adresse IP du serveur esclave
peer address 172.16.8.96;
# Le port IP du serveur esclave
peer port 1520;
# Le temps (en secondes) de pooling pour definir quand le serveur d'en face est mort
max-response-delay 30;
# Le nombre de message BNDUPD avant de considerer que le serveur d'en face est mort
max-unacked-updates 10;
# Permet d'eviter l'etat ou un serveur repond aux messages de failover mais plus aux requetes
load balance max seconds 3;
# Pendant combien de temps (en secondes) un serveur DHCP qui a repondu sur l'autre moitie de le plage
# que la sienne continue-t-il a repondre aux requetes
mclt 1800;
# Cela permet de specifier comment couper en deux la plage d'adresses dynamiques
# Il utilise un mecanisme de hashage et 128 permet de mettre autant d'adresses IP sur chaque serveur.
split 128;
}

On the Secondary :

failover peer "dhcp-failover" {
# ce serveur est le maitre
secondary;
# L'adresse IP local du serveur maitre
address 172.16.8.96;
# Le port IP local du serveur maitre
port 1520;
# L'adresse IP du serveur esclave
peer address 172.16.8.95;
# Le port IP du serveur esclave
peer port 1520;
# Le temps (en secondes) de pooling pour definir quand le serveur d'en face est mort
max-response-delay 30;
# Le nombre de message BNDUPD avant de considerer que le serveur d'en face est mort
max-unacked-updates 10;
# Permet d'eviter l'etat ou un serveur repond aux messages de failover mais plus aux requetes
load balance max seconds 3;
}

Then I configured a subnet, options, etc. I also created a special « empty » subnet containing the IP I am serving DHCP on :
# Local Subnet

subnet 172.16.8.0 netmask 255.255.252.0 {
        option routers 172.16.11.254;
        authoritative;
        }

Then you can start it… I used the -d (debug) and -f (foreground) options to debug.
I won’t spend too much time here in explanations, but it wasn’t starting. The error message was : no pool defined for any interface.
I am using Vlan, virtual interfaces, etc. So, the physical interface is e1000g1. The vlan tagged one is e1000g54001. The virtual vlan tagged interface is on e1000g54001:4 (the 4th interface).
For doing this, you just plumb interface e1000g54001, giving it no IP, letting it down. Then you use the « addif » option of ifconfig to add the virtual interface over e1000g54001.

For any reason (a bug ?) dhcpd can’t see the IP’s on virtual interfaces if the physical interface is not up.
Just do a « ifconfig e1000g54001 up ».
Then add the e1000g54001 interface at the end of the dhcpd command line to bind against it.

Everything went find then.
Wow!!!!
and what if I stop one of the node ?

Unfortnatly, stopping one node with a kill command set the other node in a mode where it does not answer any request. The reson is that the living node does not know if the other is still alive or not. To be sure not to answer twice to a request, it stop answering for some time. I havn’t been able to know for how long yet. maybe 24 hours. I read few things about it, and never found how to change this behaviour. I would like to minimize this value to 5 minuts, maybe less, so if you have an idea, come along.

Then I heard of omshell and dhcpctl. This is a way to connect to the dhcp server and give him live commands.
After a lot of researches, I finaly found how to use it. First, configure omshell to start up, in both nodes config file :

key omapiname {
algorithm hmac-md5;
secret "XXXXXXXXXX";
};
omapi-key omapiname;
omapi-port 7911;

« omapiname » can be any name. Replace the XXX with a key, which can be anything, I think.
You can also use dnssec-keygen to creake a key and copy the public part as a key :

dnssec-keygen -a HMAC-MD5 -b 512 -n HOST dhcpd_key_file

Restart dhcpd and you’re all set.
Start omshell, and connect to one of the node, the one you want to stop :

# /opt/dhcp/dhcp-4.0.0/bin/omshell
> server 172.16.248.18
> port 7911
> key rtlomapi "XXXXXX"
> connect

> new control
obj: control
> open
obj: control
state = 00:00:00:00
> set state=2
obj: control
state = 2
> update
obj: control
state = 2

You’ll see in the logs :

 Apr 16 18:15:22 cetus dhcpd: [ID 702911 local7.info] failover peer dhcp-failover: peer moves from normal to shutdown
Apr 16 18:15:22 cetus dhcpd: [ID 702911 local7.info] failover peer dhcp-failover: I move from normal to partner-down
Apr 16 18:15:23 cetus dhcpd: [ID 702911 local7.info] peer dhcp-failover: disconnected

Aclocal was complaining not having the libtool macro.

On this host we have a special folder with every tools, /opt/tools :

apr-util-1.2.8             coreutils-5.94             grep-2.5.1a                libshout-2.2.2             neon                       rsync-2.6.8                tcpdump-3.9.4
IQ-300-SOL.tar.Z           ase1254sol64               gcc                        icecast-2.3.1              libtool-1.5.24             neon-0.26.2                subversion                 tcpflow
apr                        autoconf-2.60              gcc-4.0.3                  lame-3.96.1                lsof                       openserver-12.5.1-x86      subversion-1.4.2           tcpflow-0.21
apr-1.2.8                  automake-1.10              gcc-coolthreads            libpcap                    lsof-4.77-64-100606-sparc  pushconfig                 sybase                     vim
apr-util                   coreutils                  grep                       libpcap-0.9.4              ncurses-5.5                rsync                      tcpdump                    vim-7.0.017

I just changed the PATH env variable, so I can find libtool and automake libraries :

export PATH=/opt/tools/automake-1.10/bin:/opt/tools/autoconf-2.60/bin:/opt/tools/libtool-1.5.24/bin:$PATH

But during the autogen.sh from Icecast, I got errors like this :

# ./autogen.sh
Checking for automake version
found automake
found aclocal
Generating configuration files for icecast, please wait....
aclocal  -I m4
autoheader
libtoolize --automake
automake --add-missing
src/avl/Makefile.am:7: Libtool library used but `LIBTOOL' is undefined
src/avl/Makefile.am:7:   The usual way to define `LIBTOOL' is to add `AC_PROG_LIBTOOL'
src/avl/Makefile.am:7:   to `configure.in' and run `aclocal' and `autoconf' again.
src/avl/Makefile.am:7:   If `AC_PROG_LIBTOOL' is in `configure.in', make sure
src/avl/Makefile.am:7:   its definition is in aclocal's search path.
src/httpp/Makefile.am:5: Libtool library used but `LIBTOOL' is undefined
src/httpp/Makefile.am:5:   The usual way to define `LIBTOOL' is to add `AC_PROG_LIBTOOL'
src/httpp/Makefile.am:5:   to `configure.in' and run `aclocal' and `autoconf' again.
src/httpp/Makefile.am:5:   If `AC_PROG_LIBTOOL' is in `configure.in', make sure
src/httpp/Makefile.am:5:   its definition is in aclocal's search path.
src/log/Makefile.am:7: Libtool library used but `LIBTOOL' is undefined
src/log/Makefile.am:7:   The usual way to define `LIBTOOL' is to add `AC_PROG_LIBTOOL'
src/log/Makefile.am:7:   to `configure.in' and run `aclocal' and `autoconf' again.
src/log/Makefile.am:7:   If `AC_PROG_LIBTOOL' is in `configure.in', make sure
src/log/Makefile.am:7:   its definition is in aclocal's search path.
src/net/Makefile.am:7: Libtool library used but `LIBTOOL' is undefined
src/net/Makefile.am:7:   The usual way to define `LIBTOOL' is to add `AC_PROG_LIBTOOL'
src/net/Makefile.am:7:   to `configure.in' and run `aclocal' and `autoconf' again.
src/net/Makefile.am:7:   If `AC_PROG_LIBTOOL' is in `configure.in', make sure
src/net/Makefile.am:7:   its definition is in aclocal's search path.
src/thread/Makefile.am:7: Libtool library used but `LIBTOOL' is undefined
src/thread/Makefile.am:7:   The usual way to define `LIBTOOL' is to add `AC_PROG_LIBTOOL'
src/thread/Makefile.am:7:   to `configure.in' and run `aclocal' and `autoconf' again.
src/thread/Makefile.am:7:   If `AC_PROG_LIBTOOL' is in `configure.in', make sure
src/thread/Makefile.am:7:   its definition is in aclocal's search path.
src/timing/Makefile.am:7: Libtool library used but `LIBTOOL' is undefined
src/timing/Makefile.am:7:   The usual way to define `LIBTOOL' is to add `AC_PROG_LIBTOOL'
src/timing/Makefile.am:7:   to `configure.in' and run `aclocal' and `autoconf' again.
src/timing/Makefile.am:7:   If `AC_PROG_LIBTOOL' is in `configure.in', make sure
src/timing/Makefile.am:7:   its definition is in aclocal's search path.
autoconf
I am going to run ./configure with no arguments - if you wish
...

After few reserch, I found nothing usefull, until I found a way to get the path aclocal is using to search for macro:

aclocal --print-ac-dir
/opt/tools/automake-1.10/share/aclocal

This directory was not present, butI had a /opt/tools/automake-1.10/share/aclocal-1.10.

My solution was to symlink it :

ln -s /opt/tools/automake-1.10/share/aclocal-1.10 /opt/tools/automake-1.10/share/aclocal

I then had a look into the libtool install. I found there an aclocal directory, with two .m4 file inside.
Once again, my quick solution was to copy those files inside the aclocal directory :

cp  /opt/tools/libtool-1.5.24/share/aclocal/*  /opt/tools/automake-1.10/share/aclocal

Another solution would have been to change the aclocal path to also check the libtool dirs, but I’m too busy to find out how. Give a comment if you know that

• Apache version upgraded to 2.2.6. Added mod_fcgid-2.2, mod_jk-1.2.25, mod_proxy, SMF support.
• PHP version upgraded to 5.2.4. Configure and compiler options changed. Added dtrace extension, FastCGI support. Added a configuration script, setup-ws7-php.sh for use with Sun Web Server 7.0. Default location of php.ini is now in /opt/coolstack/php5/etc irrespective of whether you’re using mod_php, php binary or php-cgi.
• APC version upgraded to 3.0.14.
• gmp library compiler optimization level changed from -fast to -xO4.
• openldap CFLAGS now includes -D_AVL_H_.
• MySQL version upgraded to 5.0.45. Compiler optimization flags changed, added support for ndbcluster and Solaris SMF support.
• Memcached version upgraded to 1.2.2. libevent version upgraded to 1.3d along with some patches which are now in 1.3e. Changed compiler flags to include « -mt » and remove « -xipo » (memcached only).
• Squid version upgraded to 2.6.16. Added SMF support.
• Tomcat version upgraded to 5.5.23. Added SMF support. Tomcat now installs in /opt/coolstack/tomcat.
• Ruby version upgraded to 1.8.6, rails to 1.2.3.
• Added lighttpd 1.4.18.
• Perl extensions DBI-1.59, DBD-mysql-4.005 and Sys-Syslog-0.18 have been added, DB_File removed.
• Added pcre-7.1 as part of CSKruntime package. Currently used by lighttpd only.

 Check it out here. Let’s have a try before the Week-end time comes  

Of course you can have the serial number from the ALOM/ILOM. But what if you are already logged and too lazy to open a new connection to it ?

There comes the IPMItool.
On solaris 10, use « /usr/sfw/bin/ipmitool »
You can have almost everything you have in the ALOM. You can GET values (like the CPU temp) and also SET values.

Get back to our Support problem. The Serial number. It is hidden in the « fru » parameter. Be carefull, you’ll get tons of serial number.

The right one is the « Product Serial » of the « Rack Mount Chassis » :

FRU Device Description : mb.fru (ID 4)
Chassis Type          : Rack Mount Chassis
Chassis Part Number   : 541-4444-44
Chassis Serial        : UNKNOWN
Board Product         : ASSY,MOTHERBOARD,X4100 M2
Board Serial          : 1005LCB-444444444
Board Part Number     : 501-4444-01
Board Extra           : 50
Board Extra           : G1F_MB
Product Manufacturer  : SUN MICROSYSTEMS
Product Name          : SUN FIRE X4100 M2
Product Part Number   : 602-4444-01
Product Serial        : 4444AL44AA          <-------------------- here

Find your equipment dealer, click on its name then chose to download the ASN.1 zip file. It will contail every mib file you choosed.
Thanks !!

Then, if you are running Solaris, copy the mibs file to /etc/sma/snmp/mibs
Then, when doing your snmpwalk,  use the « -m all » option to use all of the mibs files :

 $ snmpwalk -m all -v2c -c public mlx1- 1.3.6.1.4.1.1991.1.1.1.1   
FOUNDRY-SN-AGENT-MIB::snChasSerNum.0 = STRING: "QSFDSF444"
FOUNDRY-SN-AGENT-MIB::snChasPwrSupplyStatus.0 = INTEGER: 268
FOUNDRY-SN-AGENT-MIB::snChasFanStatus.0 = INTEGER: 24
FOUNDRY-SN-AGENT-MIB::snChasEnablePwrSupplyTrap.0 = INTEGER: enabled(1)
FOUNDRY-SN-AGENT-MIB::snChasEnableFanTrap.0 = INTEGER: enabled(1)
FOUNDRY-SN-AGENT-MIB::snChasIdNumber.0 = STRING: "SDSD33DF"
FOUNDRY-SN-AGENT-MIB::snChasEnableTempWarnTrap.0 = INTEGER: enabled(1)
FOUNDRY-SN-AGENT-MIB::snChasFlashCard.0 = INTEGER: 0
FOUNDRY-SN-AGENT-MIB::snChasFlashCardLeds.0 = INTEGER: 0
FOUNDRY-SN-AGENT-MIB::snChasNumSlots.0 = INTEGER: 4
FOUNDRY-SN-AGENT-MIB::snChasArchitectureType.0 = INTEGER: 4
FOUNDRY-SN-AGENT-MIB::snChasProductType.0 = INTEGER: 12

More to come soon (I hope) on Foundry snmp monitoring…

We have no pictures of the rackmount itself, but it is far more intuitive than the one of the Netra 4200. The only strange thing is you have to put some cage nut outside of the rack. Then you screw a metal part in each corner, and you finaly plug a strong metal rack in each side, no screw there. you are done, just put your 130 Kg Sun Blade rackin it (or remove everything you can, like fans and power supply, and you’ll get a 50 kg box).

And this is what you get :

And the back :

Then you can start putting all your things inside the Sun Blade. This is an AMD Opetron Blade, which seems to be a used one, as you can see the dust on the CPU radiators :

Then the SAN fibre channel card :

And the Hard Drives :

Don’t forget to get the right plug as our was delivered with big 16 ampers blue plug like this :

[Edit - 20070913]

We had a « crash » problem many times with the Sunblade. It’s like some (not all) blades were shutdown sometimes after 2 or 3 weeks up, but doing nothing. The Network IO cards and FC cards were also down.
A Sun support guy  went in and ask us for a spacial cable to connect directly to the blade. That was the answer of the small plub in the front of each blade : a multi IO interface.

There you go with the plug picture and the cable offering : dual USB, serial, RJ45 and video (VGA ?) output.

This will enable us to patch some of the RME 32AES soundcard input to some software our developpers just made.

One thing the Dev team forgot, is that we are exclusively working on Solaris 10, on Sun’s hardware. This may sound easy as Jack is known to compile on Linux, OsX and Windows. But unfortunately, Jack does not compile on Solaris 10 !!

Googling to find a similar problem, we just found that FreeBSD had the same problem one year ago. Strange…

Cutting the story short, I had to dig into the code and kind of clean it up to have it compiling. I will just show you the final result as the changes are only trivial sanityse. You just have to edit config/os/generic/time.h and config/os/generic/time.c. It’s like some code in the .h should have been set in the .c file.

These are the final files. Change the original so it look like this :
In config/os/generic/time.h :

#ifndef __jack_time_h__
#define __jack_time_h__

#include 

jack_time_t jack_get_microseconds_from_system ();

inline jack_time_t jack_get_microseconds (void);

/*
inline jack_time_t
jack_get_microseconds (void) {
        return jack_get_microseconds_from_system ();
}
*/

#endif /* __jack_time_h__ */

config/os/generic/time.c :

void jack_init_time ()
{
        /* nothing to do on a generic system - we use the system clock */
}
void jack_set_clock_source (jack_timer_type_t clocksrc)
{
        /* only one clock source on a generic system */
}

inline jack_time_t
jack_get_microseconds (void) {
        return jack_get_microseconds_from_system ();
        }

[EDIT]
After some testing it seems libjack is having a bug when trying to link to jackd. A friend of mine, Daniel, Dev Team Manager, found the bug, and do a solution. You just have to change few lines in libjack/client.c, line 834 :

replace :

        snprintf (req.object_path, sizeof (req.object_path),
                  "%s", va->load_name);
        snprintf (req.object_data, sizeof (req.object_data),
                  "%s", va->load_init);

by :

      snprintf (req.object_path, sizeof (req.object_path),
                  "%s", va->load_name ? va->load_name: "");
      snprintf (req.object_data, sizeof (req.object_data),
                  "%s", va->load_init ? va->load_init : "");

[EDIT]
Even if Jack compiles well, it seems it is still buggy in usage. For the moment we switched back to Linux, waiting to have time to do the whole debug. Maybe the lattency on the Solaris kernel will never enable us to use Jack on Solaris….

First is a « Framework » for helping  Solaris Zones deployment : BrandZ.
BrandZ will help you manage the creation of zones. It will allow you to have a simple Solaris zone, the same but with different software installed, up to a full Linux zone in a minute (or so )

The second one is a (yet to be tested by me) Wiki : TiddlyWiki.
What’s a Wiki ?    It’s a website you can easily change, tune, arrange….
What’s so special with this on ?    It’s HTML, CSS and Javascript based so you won’t need any « server side logic », which means no SSI, no PHP, no damn perl or python !

I had to put some PCI cards in it. One of the main feature of the M2 version is to support 3 full size PCI (PCI-X) cards. You can also have 2 PCI-Express cards. You will find in below the « extension board » inside the server. When you open the server, the top board slides to the back of the server. Then you can access the PCI-E and the RAM slots (under the white cover at the front). This is quite funny to see on a server but the 3 flat cables used to connect this IO board to the motherboard seems week. Don’t play too much with it.

The front fan module also seems to be a little bit lazy. Maybe it won’t move once the top is closed…

On the front you will see the DVD rom and 2 hotswap SAS disks.

The main problem we had concerns the PCI-X slots. As you’ll see in next pictures, there is nothing to « lock » the card once in place. When I say nothing i’m wrong. You have the small green plastic thing you can move  from « open » to « close ». The only problem is that even on the close position they are damn too small to reach the edge of the PCI card.

Only the PCI on the right (when looking from the top, like on pictures) have a locking system, a screw to put from the outside.

I think I will open a case at Sun about this issue. I may also open a case about the shitty rackmount they give with it : instead of the clean rackmount of the T2000, with a slider for cables at the back, it’s a « do it yourself » old russian way rackmount. You get a heavy package with metal parts hidden in it and a thousands screw bag. No docs !
Hopefully with a 15 mins searching on Sun’s website you can find the x4200 rackmount doc. You will then learn you need to use the M4, M5 and M6 screw. Looking to the paper packaged inside the screwbag, you will see that you have 4 types of M4, 4 of M5 and as many as M6. You choose
I spent some time last week, moving T2000 from one room to another, taking about 15 mins a server, unmounting to remounting.
It took me half a day for 2 x4200… and the backside of the rack if a mess with cables everywhere.
Ooops, I forgot : once in the rackmount, there is no locking facility. Be carefull when you plug cables at the back as your new x4200 may jump out ou crash to the front door (you have a locking system when you take the x4200 out).

Pictures of the rackmount to come.
X4200 is an AMD based host so it have an ILOM. You can get the console using the remote display tool which is a java JNLP application. We had issue here when using it on OsX. On my host for example, I have a SSL key problem which disallow me to use it. I had to use another osX… If you have any info on that I can provide the error code.

Whatever, once started I see no differences between this and the past one.
The new challenge will to have QFS working on Linux (Ubuntu).