ssic-linux-users — User discussion

[SSI-users] Is www.openssi.org dead?

[John H. <jo...@Ca...>]

I can't get to it.

OpenSSI.eu doesn't work either.

(Before I rewrote the http://en.wikipedia.org/Single-sytem_image page 
someone (128.193.146.131 at Oregon State U.) marked OpenSSI as 
"inactive" because the website hadn't been updated for 2 years.  I undid 
their edit on the grounds that "the website is inactive doesn't mean 
that the project is inactive").

[SSI-users] Using ALSA DMIX plugin on an OpenSSI cluster

[John H. <jo...@Ca...>]

If you want to use ALSA in an OpenSSI cluster you may run across the 
problem that the "DMIX" plugin - software mixer - uses a fixed semaphore 
ID, which can cause conflicts between different nodes.

My quick and dirty fix is to:

1. move the /usr/share/alsa/alsa.conf file to one side:

    mv /usr/share/alsa/alsa.conf /usr/share/alsa/alsa.conf-openssi
      

(on a Debian system a better way to do this is:

    dpkg-divert --local --rename --divert /usr/share/alsa/alsa.conf-openssi /usr/share/alsa/alsa.conf
      

)

2. Make a new alsa.conf:

    # cat <<'EOF' > /usr/share/alsa/alsa.conf
    </usr/share/alsa/alsa.conf-openssi>
    defaults.pcm.!ipc_key 5678293
    EOF
      

3, make this a node-specific file:

    # mklocalfile /usr/share/alsa/alsa.conf
      

4. edit the /cluster/nodeXX/usr/share/alsa/alsa.conf files so the 
"ipc_key" is unique per node.  Prefix it with the node number for example.

Re: [SSI-users] Can OpenSSI Be Used for General-Purpose Computing?

[John H. <jo...@Ca...>]

Robert Wachinger wrote:
> On Mon, Sep 01, 2008 at 12:17:37PM +0200, John Hughes wrote:
>   
>> An OpenSSI cluster tries to look as much like one big multi-processor 
>> machine as possible.
>>     
>
> As far as I know, with one big difference:
>
> On a multiprocessor machine the OS' scheduler chooses automatically,
> on which processor a process will run.
> In OpenSSI the default is, that a process (with all of its threads)
> runs on the node, where it is started, unlesse you say explicitely
> otherwise (e.g. started in a "bash-ll", your application is made known
> to be able to migrate or so ...).
>
> Or is that not correct (any more)?
>   
Yup.  If /proc/pid/loadlevel is zero a process will not be moved from 
one node to another by the loadlevel logic.

If /proc/pid/pin is zero you can still migrate the process by hand (by 
writing a node number to /proc/pid/goto).

(As you say, the normal way of setting /proc/pid/loadlevel is to run 
/bin/bash-ll, or any other program mentioned in 
/proc/cluster/loadlevellist).

Re: [SSI-users] Can OpenSSI Be Used for General-Purpose Computing?

[Robert W. <no...@ro...>]

On Mon, Sep 01, 2008 at 12:17:37PM +0200, John Hughes wrote:
> Will Keaney wrote:
> > Hello.
> > I came across your project while doing some research on Linux 
> > clustering.  One question that I don't find clearly answered in the 
> > available documentation is, can unmodified applications be run on an 
> > OpenSSI cluster?  The other clustering solutions I have found all 
> > require specially modified code to use their libraries (Beowulf, PVM, 
> > et al), and I am really looking for something that is more versatile.
> An OpenSSI cluster tries to look as much like one big multi-processor 
> machine as possible.

As far as I know, with one big difference:

On a multiprocessor machine the OS' scheduler chooses automatically,
on which processor a process will run.
In OpenSSI the default is, that a process (with all of its threads)
runs on the node, where it is started, unlesse you say explicitely
otherwise (e.g. started in a "bash-ll", your application is made known
to be able to migrate or so ...).

Or is that not correct (any more)?


> The limitations are all related to the (much) higher latency/lower 
> throughput for inter-node memory access.
> 
> So:
> 
>    1. Cross-node shared memory is slower than normal shared memory.
>    2. Other forms of cross-node IPC are slower (Unix domain sockets,
>       SysV semaphores, SysV message queues).
>    3. Multi-threaded applications run with all threads on one node.
>    4. Accesses to filesystems mounted on other nodes are slower.
> 
> If your application can be written as many mostly independent processes 
> then OpenSSI should be very easy to use and give acceptable 
> performance.  (A good example: Linux kernel compiles with "-j" set to 
> number of nodes times number of cores).

Ack.


Regards, Robert

Re: [SSI-users] Can OpenSSI Be Used for General-Purpose Computing?

[John H. <jo...@Ca...>]

Will Keaney wrote:
> Hello.
> I came across your project while doing some research on Linux 
> clustering.  One question that I don't find clearly answered in the 
> available documentation is, can unmodified applications be run on an 
> OpenSSI cluster?  The other clustering solutions I have found all 
> require specially modified code to use their libraries (Beowulf, PVM, 
> et al), and I am really looking for something that is more versatile.
An OpenSSI cluster tries to look as much like one big multi-processor 
machine as possible.

The limitations are all related to the (much) higher latency/lower 
throughput for inter-node memory access.

So:

   1. Cross-node shared memory is slower than normal shared memory.
   2. Other forms of cross-node IPC are slower (Unix domain sockets,
      SysV semaphores, SysV message queues).
   3. Multi-threaded applications run with all threads on one node.
   4. Accesses to filesystems mounted on other nodes are slower.

If your application can be written as many mostly independent processes 
then OpenSSI should be very easy to use and give acceptable 
performance.  (A good example: Linux kernel compiles with "-j" set to 
number of nodes times number of cores).

Re: [SSI-users] Can OpenSSI Be Used for General-Purpose Computing?

[Mulyadi S. <mul...@gm...>]

Hi

On Sat, Aug 30, 2008 at 2:19 AM, Will Keaney <ke...@gm...> wrote:
> Hello.
> I came across your project while doing some research on Linux clustering.
>  One question that I don't find clearly answered in the available
> documentation is, can unmodified applications be run on an OpenSSI cluster?
>  The other clustering solutions I have found all require specially modified
> code to use their libraries (Beowulf, PVM, et al), and I am really looking
> for something that is more versatile.

Up to some degrees, you can run unmodified application to take
advantage of openSSI. The kind of such apps is the one who forks some
numbers of child processes (multi thread or multi processing). Please
remember openSSI can only migrate process. Threads will be migrated
all together.

Or, you can do it with so called "parametric execution". For example
run 10 instance of application A, each is fed with differrent
parameters. Run them then collect the results. One of the example of
such scenario is batch scheduler (e.g running distributed rendering).

regards,

Mulyadi.

[SSI-users] Can OpenSSI Be Used for General-Purpose Computing?

[Will K. <ke...@gm...>]

Hello.
I came across your project while doing some research on Linux 
clustering.  One question that I don't find clearly answered in the 
available documentation is, can unmodified applications be run on an 
OpenSSI cluster?  The other clustering solutions I have found all 
require specially modified code to use their libraries (Beowulf, PVM, et 
al), and I am really looking for something that is more versatile.

Thanks,

Will

Re: [SSI-users] QM400 and open-ssi

[John H. <jo...@Ca...>]

Vincent Diepeveen wrote:
> If i add a quadrics QM400 driver (patch to kernel),
> is it possible then somehow to get to work open-ssi 1.92 over the  
> qm400 cards?
>
> When needed i'm prepared to do effort there, if i knew what.
>   

You should have a look at Stan Smith's OpenSSI over infiniband stuff.

In outline the work to be done involves porting the internode 
communications code (ICS, part of CI) to your Quadrics stuff and making 
a initial ram disk that sets up the communications.

Maybe try contacting Stan directly?

> Even if it would require first boot over tcp/ip. I'm interesting in  
> getting it to work somehow.
>
> My chessprogram Diep uses shared memory. Ideally each process  
> allocates its own shared memory and attaches to each
> other process the shared memory. Each process is basically  
> calculating within its own chunk of shared memory,
> so that shared memory must get allocated at the node where the  
> process was started.
>   
Shared memory "just works" on OpenSSI.

When a page is mapped read-write by a node then is is marked "absent" on 
all the other nodes.

The r/w node just accesses it as memory.

If another node tries to read it then it gets marked read-only and 
copied over to the other node as a read-only page.

If another node tries to write it then it gets copied to the other node 
as a read/write page and marked absent on the other nodes.

So:

1 writer, no readers - no problem.

many readers, no writers - no problem.

1 writer many readers - some latency.  Has anyone measured this lately?

> Sometimes processes reference shared memory at remote nodes. The  
> important thing is that open-ssi gets that reference from
> the remote nodes over the QM400 network and not over the tcp/ip  
> network. Latency is important. Then i can get world champion
> computerchess next month in China with quadrics.
>   
Next month!
> I see by the way it has kernel 2.6.11. This might be a problem unless  
> when this is a vanilla kernel.
> I see quadrics has support for vanilla kernel 2.6 to integrate its  
> driver into the kernel (with a source code patch and
> building modules after that).
>   

Well, it's obviously not "vanilla", it has the CI and OpenSSI code!

To see if the patch applies suck it and see,
> Especially i see it support vanilla 2.6.18 kernel.
>   

No ETA on 2.6.18 at the moment.

[SSI-users] QM400 and open-ssi

[Vincent D. <di...@xs...>]

hi all,

These developments at Open-SSI past few years are really very  
promising. Great stuff!

I have a rather embarrassing question (revealing my layman ship  
regarding open-ssi despite being for years
at the mailing list now).

If i add a quadrics QM400 driver (patch to kernel),
is it possible then somehow to get to work open-ssi 1.92 over the  
qm400 cards?

When needed i'm prepared to do effort there, if i knew what.

Even if it would require first boot over tcp/ip. I'm interesting in  
getting it to work somehow.

My chessprogram Diep uses shared memory. Ideally each process  
allocates its own shared memory and attaches to each
other process the shared memory. Each process is basically  
calculating within its own chunk of shared memory,
so that shared memory must get allocated at the node where the  
process was started.

Sometimes processes reference shared memory at remote nodes. The  
important thing is that open-ssi gets that reference from
the remote nodes over the QM400 network and not over the tcp/ip  
network. Latency is important. Then i can get world champion
computerchess next month in China with quadrics.

I see by the way it has kernel 2.6.11. This might be a problem unless  
when this is a vanilla kernel.
I see quadrics has support for vanilla kernel 2.6 to integrate its  
driver into the kernel (with a source code patch and
building modules after that).

Especially i see it support vanilla 2.6.18 kernel.

Is that a problem?

Best regards,
Vincent

Re: [SSI-users] building cluster

[John H. <jo...@Ca...>]

Richard Cesar wrote:
> Hi there,
>   
>> I read a lot of information about OpenSSI and I must say that I really 
>> like this project.
>> I have an idea how to transform our existing infrastructure to the 
>> cluster but I have some questions.
>> We have few servers (7) which are used for several applications (web 
>> server, DNS, mail server, many game servers) and all together forms 
>> the gaming portal. Each server has its own public IP and each server 
>> has its own hard drive with Debian OS and installed services which it 
>> runs.
>> I am not sure if I understood the right functions of OpenSSI but my 
>> idea is this.
>>
>> I want to create one cluster from these servers. I would have one 
>> master server running OpenSSI and other servers would be cluster 
>> nodes. I would like to have all data on the master server and I want 
>> the cluster to act as one computer with one IP address running all 
>> server software. This should do some process based load-balancing 
>> between nodes.
>> After reading some documentation I believe there is a chance that this 
>> will work.
>>     
OpenSSI can give you exactly that.  You don't need one "master node", 
just one or more nodes that have access to the root filesystem and can 
run init.  If you have more than one OpenSSI gives you high availability 
- if the init node goes down init and the root filesystem will fail over 
to one of the other init nodes.

To make the root filesystem accessible from multiple nodes either use 
dual ported disks - SCSI, SAS, FC - or DRBD (cross-network mirroring in 
effect).
>> BUT ...
>>
>> Our machines are delivered from different sources (mostly as a sponsor 
>> gift) and they have different motherboards, CPUs, network cards (some 
>> are onboard some are PCI) etc. and some of them has multicore CPUs.
>>     
Not a problem, as long as the machines all have a common instruction set 
(i.e. you can't mix and match Itaniums and Alphas!).
>> So one problem is that some network cards probably :-) doesnt support 
>> the PXE protocol, but we will probably buy new gigabit cards and 
>> gigabit switch for cluster`s internal network. Can we use Etherboot 
>> with standard network card , for example can we boot Etherboot from 
>> flashcard ?
>>     
You only need network booting for the initial installation of a node.  
After the node has been configured into the cluster it can boot off it's 
local disk.  I've used etherboot off floppies in the past.

I'd certainly recommend gigabit ethernet for the cluster interconnect - 
100mbit is a bit too slow.

>> But I think bigger problem is ... if we use multicore CPU machine as a 
>> cluster node .. the machine will boot from network image which is 
>> (according to the documentation) Syslinux image ... I dont know then 
>> ... will it use both CPUs (cores) for the cluster ?? will it work at 
>> all ??
>>
>>     
The OpenSSI kernel is an SMP kernel.  It can run on as many cores as you 
want.

OpenSSI doesn't care if different nodes have different numbers of cores.

[SSI-users] building cluster

[Richard C. <sp...@ga...>]

Hi there,
> I read a lot of information about OpenSSI and I must say that I really 
> like this project.
> I have an idea how to transform our existing infrastructure to the 
> cluster but I have some questions.
> We have few servers (7) which are used for several applications (web 
> server, DNS, mail server, many game servers) and all together forms 
> the gaming portal. Each server has its own public IP and each server 
> has its own hard drive with Debian OS and installed services which it 
> runs.
> I am not sure if I understood the right functions of OpenSSI but my 
> idea is this.
>
> I want to create one cluster from these servers. I would have one 
> master server running OpenSSI and other servers would be cluster 
> nodes. I would like to have all data on the master server and I want 
> the cluster to act as one computer with one IP address running all 
> server software. This should do some process based load-balancing 
> between nodes.
> After reading some documentation I believe there is a chance that this 
> will work.
>
> BUT ...
>
> Our machines are delivered from different sources (mostly as a sponsor 
> gift) and they have different motherboards, CPUs, network cards (some 
> are onboard some are PCI) etc. and some of them has multicore CPUs.
>
> So one problem is that some network cards probably :-) doesnt support 
> the PXE protocol, but we will probably buy new gigabit cards and 
> gigabit switch for cluster`s internal network. Can we use Etherboot 
> with standard network card , for example can we boot Etherboot from 
> flashcard ?
>
> But I think bigger problem is ... if we use multicore CPU machine as a 
> cluster node .. the machine will boot from network image which is 
> (according to the documentation) Syslinux image ... I dont know then 
> ... will it use both CPUs (cores) for the cluster ?? will it work at 
> all ??
>
> thanks for answer
>

Re: [SSI-users] drbd ha-lvs dr problems

[Karsten N. <ka...@ww...>]

Hi John

You work realy fast :) Thanks a lot! 
I will test it when i get back from my hollydays in 3 weeks time.

- Karsten

-----Oprindelig meddelelse-----
Fra: John Hughes [mailto:jo...@Ca...]
Sendt: fr 7/25/2008 5:14
Til: Karsten Nielsen
Cc: ssi...@li...
Emne: Re: [SSI-users] drbd ha-lvs dr problems
 
John Hughes wrote:
> John Hughes wrote:
>> The following packages have unmet dependencies:
>>   openssi-drbd: Depends: drbd0.7-module-2.6.11-ssi-8k-686-smp but it 
>> is not installable
>> E: Broken packages
>>   
> Ok, I've fixed at least the install for drbd on openssi-etch.  You can 
> now do a update && install openssi-drbd.
>
> Still no idea whether it works.
>
> Still got to fix drbd install for Lenny.
Ok, drbd install fixed and drbd tested for OpenSSI Etch and Lenny.

It works for me.

Not yet tested ha-lvs.

Re: [SSI-users] drbd ha-lvs dr problems

[John H. <jo...@Ca...>]

John Hughes wrote:
> John Hughes wrote:
>> The following packages have unmet dependencies:
>>   openssi-drbd: Depends: drbd0.7-module-2.6.11-ssi-8k-686-smp but it 
>> is not installable
>> E: Broken packages
>>   
> Ok, I've fixed at least the install for drbd on openssi-etch.  You can 
> now do a update && install openssi-drbd.
>
> Still no idea whether it works.
>
> Still got to fix drbd install for Lenny.
Ok, drbd install fixed and drbd tested for OpenSSI Etch and Lenny.

It works for me.

Not yet tested ha-lvs.

Re: [SSI-users] drbd ha-lvs dr problems

[John H. <jo...@Ca...>]

John Hughes wrote:
> The following packages have unmet dependencies:
>   openssi-drbd: Depends: drbd0.7-module-2.6.11-ssi-8k-686-smp but it is not installable
> E: Broken packages
>   
Ok, I've fixed at least the install for drbd on openssi-etch.  You can 
now do a update && install openssi-drbd.

Still no idea whether it works.

Still got to fix drbd install for Lenny.

Re: [SSI-users] drbd ha-lvs dr problems

[Karsten N. <ka...@ww...>]

I am using your etch based stuff.

- karsten

On Thursday 24 July 2008 14:08:34 Karsten Nielsen wrote:
> I have build the drbd0.7 modules from source and installed drbd0.7 utils as
> far as i recall that was the only dependencies of that package.
>
> - karsten
>
> On Thursday 24 July 2008 13:47:03 John Hughes wrote:
> > Karsten Nielsen wrote:
> > > On Monday 21 July 2008 15:09:20 you wrote:
> > >> Karsten Nielsen wrote:
> > >>> I have setup opessi 1.9.3 from john's debian etch packages. And I
> > >>> have setup drbd failover on the to initnodes.
> >
> > Do you mean my sarge based 1.9.3 or my new etch based stuff?
> >
> > I don't see how you got that far:
> >
> > # apt-get install openssi-drbd
> > Reading package lists... Done
> > Building dependency tree... Done
> > Some packages could not be installed. This may mean that you have
> > requested an impossible situation or if you are using the unstable
> > distribution that some required packages have not yet been created
> > or been moved out of Incoming.
> >
> > Since you only requested a single operation it is extremely likely that
> > the package is simply not installable and a bug report against
> > that package should be filed.
> > The following information may help to resolve the situation:
> >
> > The following packages have unmet dependencies:
> >   openssi-drbd: Depends: drbd0.7-module-2.6.11-ssi-8k-686-smp but it is
> > not installable E: Broken packages
>
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's
> challenge Build the coolest Linux based applications with Moblin SDK & win
> great prizes Grand prize is a trip for two to an Open Source event anywhere
> in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> Ssic-linux-users mailing list
> Ssi...@li...
> https://lists.sourceforge.net/lists/listinfo/ssic-linux-users

Re: [SSI-users] drbd ha-lvs dr problems

[Karsten N. <ka...@ww...>]

I have build the drbd0.7 modules from source and installed drbd0.7 utils as 
far as i recall that was the only dependencies of that package.

- karsten

On Thursday 24 July 2008 13:47:03 John Hughes wrote:
> Karsten Nielsen wrote:
> > On Monday 21 July 2008 15:09:20 you wrote:
> >> Karsten Nielsen wrote:
> >>> I have setup opessi 1.9.3 from john's debian etch packages. And I have
> >>> setup drbd failover on the to initnodes.
>
> Do you mean my sarge based 1.9.3 or my new etch based stuff?
>
> I don't see how you got that far:
>
> # apt-get install openssi-drbd
> Reading package lists... Done
> Building dependency tree... Done
> Some packages could not be installed. This may mean that you have
> requested an impossible situation or if you are using the unstable
> distribution that some required packages have not yet been created
> or been moved out of Incoming.
>
> Since you only requested a single operation it is extremely likely that
> the package is simply not installable and a bug report against
> that package should be filed.
> The following information may help to resolve the situation:
>
> The following packages have unmet dependencies:
>   openssi-drbd: Depends: drbd0.7-module-2.6.11-ssi-8k-686-smp but it is not
> installable E: Broken packages

Re: [SSI-users] drbd ha-lvs dr problems

[John H. <jo...@Ca...>]

Karsten Nielsen wrote:
> On Monday 21 July 2008 15:09:20 you wrote:
>   
>> Karsten Nielsen wrote:
>>     
>>> I have setup opessi 1.9.3 from john's debian etch packages. And I have
>>> setup drbd failover on the to initnodes.
>>>       
Do you mean my sarge based 1.9.3 or my new etch based stuff?

I don't see how you got that far:

# apt-get install openssi-drbd
Reading package lists... Done
Building dependency tree... Done
Some packages could not be installed. This may mean that you have
requested an impossible situation or if you are using the unstable
distribution that some required packages have not yet been created
or been moved out of Incoming.

Since you only requested a single operation it is extremely likely that
the package is simply not installable and a bug report against
that package should be filed.
The following information may help to resolve the situation:

The following packages have unmet dependencies:
  openssi-drbd: Depends: drbd0.7-module-2.6.11-ssi-8k-686-smp but it is not installable
E: Broken packages

Re: [SSI-users] drbd ha-lvs dr problems

[Karsten N. <ka...@ww...>]

On Monday 21 July 2008 15:09:20 you wrote:
> Karsten Nielsen wrote:
> > I have setup opessi 1.9.3 from john's debian etch packages. And I have
> > setup drbd failover on the to initnodes.
> >
> > I am planing on running a mailscanner cluster on the setup but I have run
> > into an issue with the setup and that.
> >
> > I have setup ha-lvs to serve the cluster and started the services on all
> > the nodes. The setup works perfectly until I make a hard fail of the
> > running master. The master sits in the ip pool and does not get removed.
> > My cvip.conf is below.
> > Any suggestions ?
>
> I've not done much (any!) testing of ha-lvs yet.  I'll see what I can find.

Thanks that sounds great. 

>
> > I have another problem and that is if I cut the power on initnode 1
> > initnode 2 takes over and the cluster continues as planned. But when i
> > start initnode 1 again i sits waiting to take over the master settings
> > from initnode 2 during boot and just waiting and waiting. Is there any
> > thing I have to do in order for it to become master again ?
>
> Normaly it won't become the master, it'll join as a secondary.

ok I did not know that :)

>
> When you say it's waiting and waiting what are the last messages displayed.

I cannot recreate that at the moment because the cluster broke as in it cannot 
start any more. So I have to start over with a clean install.

>
> I haven't tried drbd yet, my tests have all been with shared disk.

Re: [SSI-users] drbd ha-lvs dr problems

[John H. <jo...@Ca...>]

Karsten Nielsen wrote:
> I have setup opessi 1.9.3 from john's debian etch packages. And I have setup 
> drbd failover on the to initnodes.
>
> I am planing on running a mailscanner cluster on the setup but I have run into 
> an issue with the setup and that.
>
> I have setup ha-lvs to serve the cluster and started the services on all the 
> nodes. The setup works perfectly until I make a hard fail of the running 
> master. The master sits in the ip pool and does not get removed. My cvip.conf 
> is below.
> Any suggestions ?
>   
I've not done much (any!) testing of ha-lvs yet.  I'll see what I can find.
> I have another problem and that is if I cut the power on initnode 1 initnode 2 
> takes over and the cluster continues as planned. But when i start initnode 1 
> again i sits waiting to take over the master settings from initnode 2 during 
> boot and just waiting and waiting. Is there any thing I have to do in order 
> for it to become master again ?
>   
Normaly it won't become the master, it'll join as a secondary.

When you say it's waiting and waiting what are the last messages displayed.

I haven't tried drbd yet, my tests have all been with shared disk.

[SSI-users] drbd ha-lvs dr problems

[Karsten N. <ka...@ww...>]

hi,

I have setup opessi 1.9.3 from john's debian etch packages. And I have setup 
drbd failover on the to initnodes.

I am planing on running a mailscanner cluster on the setup but I have run into 
an issue with the setup and that.

I have setup ha-lvs to serve the cluster and started the services on all the 
nodes. The setup works perfectly until I make a hard fail of the running 
master. The master sits in the ip pool and does not get removed. My cvip.conf 
is below.
Any suggestions ?

I have another problem and that is if I cut the power on initnode 1 initnode 2 
takes over and the cluster continues as planned. But when i start initnode 1 
again i sits waiting to take over the master settings from initnode 2 during 
boot and just waiting and waiting. Is there any thing I have to do in order 
for it to become master again ?

- Karsten
---- snip
<?xml version="1.0"?>
<cvips>
        <routing>DR</routing>
        <cvip>
                <ip_addr>172.22.0.21</ip_addr>
                <service>
                        <tcp_port>10024</tcp_port>
                        <scheduler>wlc</scheduler>
                </service>
                <director_node>
                        <node_num>1</node_num>
                        <garp_interface>eth0</garp_interface>
                        <sync_interface>eth0</sync_interface>
                </director_node>
                <director_node>
                        <node_num>2</node_num>
                        <garp_interface>eth0</garp_interface>
                        <sync_interface>eth0</sync_interface>
                </director_node>

                <real_server_node>
                        <node_num>3-10</node_num>
                        <service>
                                <tcp_port>10024</tcp_port>
                        </service>
                </real_server_node>
        </cvip>
</cvips>
---- snip

[SSI-users] OpenSSI Debian preview packages moved to deb.openssi.org

[John H. <jo...@Ca...>]

Since my company webserver is down I've taken advantage of Ivan Krstic's 
kind offer to host the packages on deb.openssi.org.

You can now find the instructions at:

    http://deb.openssi.org/alpha/openssi-etch

and

    http://deb/openssi.org/alpha/openssi-lenny

I've not bothered with the Ubuntu Hardy Heron version as that is pretty 
flaky, but if anyone needs/wants it give me a shout.

Re: [SSI-users] problem with udev and drbd

[John H. <jo...@Ca...>]

John Hughes wrote:
> Karsten Nielsen wrote:
>   
>> master01:~#  dpkg -i udev_0.080-1.ssi5_i386.deb 
>> dpkg - warning: downgrading udev from 0.105-4 to 0.080-1.ssi5.
>> (Reading database ... 20119 files and directories currently installed.)
>> Preparing to replace udev 0.105-4 (using udev_0.080-1.ssi5_i386.deb) ...
>> Unpacking replacement udev ...
>>
>>
>> dpkg: dependency problems prevent configuration of udev:
>>  udev depends on libc6 (>= 2.7-1); however:
>>   Version of libc6 on system is 2.3.6.ds1-13etch5-ssi2.
>>  udev depends on libselinux1 (>= 2.0.59); however:
>>   Version of libselinux1 on system is 1.32-3.
>> dpkg: error processing udev (--install):
>>  dependency problems - leaving unconfigured
>> Errors were encountered while processing:
>>  udev
>>   
>>     
> ai ai ai!
>
> I've got my etch & lenny configs confused.
>
> What a bummer.
>
> I'll try and sort this mess out as soon as I can.
>
>   
Should be fixed now

apt-get update && apt-get dist-upgrade

Re: [SSI-users] problem with udev and drbd

[John H. <jo...@Ca...>]

Karsten Nielsen wrote:
> On Tuesday 15 July 2008 14:28:48 John Hughes wrote:
>   
>> Karsten Nielsen wrote:
>>     
>>> Hi,
>>>
>>> I have installed openssi 1.9.3 from
>>> http://www.atlantech.com/~john/openssi-etch/ and set up drbd on the root
>>> device.
>>> I have a problem.
>>> I have to force udev installation from the mirror with dpkg -i
>>> --force-depends in order to get udev to work.
>>>       
>> Why did you need the --force-depends?  What was it missing?
>>     
>
> master01:~#  dpkg -i udev_0.080-1.ssi5_i386.deb 
> dpkg - warning: downgrading udev from 0.105-4 to 0.080-1.ssi5.
> (Reading database ... 20119 files and directories currently installed.)
> Preparing to replace udev 0.105-4 (using udev_0.080-1.ssi5_i386.deb) ...
> Unpacking replacement udev ...
>
>
> dpkg: dependency problems prevent configuration of udev:
>  udev depends on libc6 (>= 2.7-1); however:
>   Version of libc6 on system is 2.3.6.ds1-13etch5-ssi2.
>  udev depends on libselinux1 (>= 2.0.59); however:
>   Version of libselinux1 on system is 1.32-3.
> dpkg: error processing udev (--install):
>  dependency problems - leaving unconfigured
> Errors were encountered while processing:
>  udev
>   
ai ai ai!

I've got my etch & lenny configs confused.

What a bummer.

I'll try and sort this mess out as soon as I can.

Re: [SSI-users] problem with udev and drbd

[Karsten N. <ka...@ww...>]

On Tuesday 15 July 2008 14:28:48 John Hughes wrote:
> Karsten Nielsen wrote:
> > Hi,
> >
> > I have installed openssi 1.9.3 from
> > http://www.atlantech.com/~john/openssi-etch/ and set up drbd on the root
> > device.
> > I have a problem.
> > I have to force udev installation from the mirror with dpkg -i
> > --force-depends in order to get udev to work.
>
> Why did you need the --force-depends?  What was it missing?
>

master01:~#  dpkg -i udev_0.080-1.ssi5_i386.deb 
dpkg - warning: downgrading udev from 0.105-4 to 0.080-1.ssi5.
(Reading database ... 20119 files and directories currently installed.)
Preparing to replace udev 0.105-4 (using udev_0.080-1.ssi5_i386.deb) ...
Unpacking replacement udev ...


dpkg: dependency problems prevent configuration of udev:
 udev depends on libc6 (>= 2.7-1); however:
  Version of libc6 on system is 2.3.6.ds1-13etch5-ssi2.
 udev depends on libselinux1 (>= 2.0.59); however:
  Version of libselinux1 on system is 1.32-3.
dpkg: error processing udev (--install):
 dependency problems - leaving unconfigured
Errors were encountered while processing:
 udev

> >  That breaks the system.
>
> Not a suprise!
>
> > I cannot ssh to the cluster without the downgrade of udev but with the
> > downgrade the system acts very wird.
> >
> > The original etch version of udev cannot be used do to the fact that it
> > requires kernel version > 2.6.15 in order for it to start
>
> Yup, that's why we have our own out-of-date udev.
>
> > Any help/pointers will bee appreciated
>
> Things may have bit-rotted due to some work I've been doing.  I'll see
> if I can check it all this weekend.

Re: [SSI-users] problem with udev and drbd

[John H. <jo...@Ca...>]

Karsten Nielsen wrote:
> Hi,
>
> I have installed openssi 1.9.3 from 
> http://www.atlantech.com/~john/openssi-etch/ and set up drbd on the root 
> device. 
> I have a problem.
> I have to force udev installation from the mirror with dpkg -i --force-depends 
> in order to get udev to work.
>   
Why did you need the --force-depends?  What was it missing?
>  That breaks the system.
>   
Not a suprise!
> I cannot ssh to the cluster without the downgrade of udev but with the 
> downgrade the system acts very wird.
>
> The original etch version of udev cannot be used do to the fact that it 
> requires kernel version > 2.6.15 in order for it to start
>   
Yup, that's why we have our own out-of-date udev.
> Any help/pointers will bee appreciated
>   
Things may have bit-rotted due to some work I've been doing.  I'll see 
if I can check it all this weekend.