bproc-users — General discussion about BProc.

Re: [BProc] Setupfs: further analysis

[Niels W. <Nie...@um...>]

Agree, that was just a test. I have now started looking at the code for 
setupfs.c I have already discovered one quirk:
On line 294 it states
     return -1;
This is an unconditional fail, and it seems to help to comment out this 
line :-) . I keep checking!
Niels

ha...@no... wrote:

>>If I add the line
>>"plugin miscfiles /dev/hda3"
>>before the setupfs line, I get a sensible error message:
>>setupfs   :     mknod("/dev/hda3", 060660, 0x303): File exists
>>    
>>
>
>I guess that "plugin miscfiles /dev/hda3" is not needed, device file
>is normally created in setupfs. Adding this line probably just creates
>another problem, unrelated to the one you are hunting for.
>
>I would try to boot the node anyhow you can (probably without setupfs)
>and then try to do the mount by hand using bpsh, something like:
>
> bpsh 1 mknod /dev/hda3 b 3 3
> bpsh 1 e2fsck /dev/hda3
> bpsh 1 mkdir /scratch
> bpsh 1 mount /dev/hda3 /scratch
>
>and watch for errors.
>
>HTH
>
>Vaclav
>
>  
>

-- 
Dr Niels R. Walet   http://www.phy.umist.ac.uk/Theory/people/walet.html
Dept. of Physics,  UMIST,   P.O. Box 88,    Manchester,  M60 1QD,  U.K.
Phone: +44(0)161-2003693 Fax: +44(0)161-2004303 Nie...@um...

Re: [BProc] Setupfs: further analysis

[<ha...@no...>]

> If I add the line
> "plugin miscfiles /dev/hda3"
> before the setupfs line, I get a sensible error message:
> setupfs   :     mknod("/dev/hda3", 060660, 0x303): File exists

I guess that "plugin miscfiles /dev/hda3" is not needed, device file
is normally created in setupfs. Adding this line probably just creates
another problem, unrelated to the one you are hunting for.

I would try to boot the node anyhow you can (probably without setupfs)
and then try to do the mount by hand using bpsh, something like:

 bpsh 1 mknod /dev/hda3 b 3 3
 bpsh 1 e2fsck /dev/hda3
 bpsh 1 mkdir /scratch
 bpsh 1 mount /dev/hda3 /scratch

and watch for errors.

HTH

Vaclav

[BProc] Setupfs: further analysis

[Niels W. <Nie...@um...>]

I have now pared down my fstab to only one line (with only comments 
setupfs finishes and the node comes up, with either the swap file system 
or the scratch (local) file system I have problems):

setupfs   :     Doing mount: dev=/dev/hda3  mntpt=/scratch  type=ext3  opts=
setupfs   :     Performing modprobe: ext3
setupfs   : Sending: jbd /lib/modules/2.4.19-lanl.22smp/kernel/fs/jbd/jbd.o
setupfs   :     Installing jbd (72444 bytes)
setupfs   :     Installing module "jbd"
setupfs   : Sending: ext3 
/lib/modules/2.4.19-lanl.22smp/kernel/fs/ext3/ext3.o
setupfs   :     Installing ext3 (91517 bytes)
setupfs   :     Installing module "ext3"
nodeup    :   Plugin setupfs returned -1 (error)


If I add the line
"plugin miscfiles /dev/hda3"
before the setupfs line, I get a sensible error message:
setupfs   :     mknod("/dev/hda3", 060660, 0x303): File exists
nodeup    : Node setup returned status 1

I checked that the filesystems are there.

-- 
Dr Niels R. Walet   http://www.phy.umist.ac.uk/Theory/people/walet.html
Dept. of Physics,  UMIST,   P.O. Box 88,    Manchester,  M60 1QD,  U.K.
Phone: +44(0)161-2003693 Fax: +44(0)161-2004303 Nie...@um...

[BProc] Re: Silly errors?

[Niels W. <Nie...@um...>]

I have now solved my "silly" error (interface should be eth1 rather than 
eth1:0 alias).

I run into another one: node_up bails out in setupfs without any 
sensible feedback. In the olden days this was a shell script, and one 
could add debug statements (and I have done that quite a bit).
Is there any way to figure out what goes wrong?

Niels
-- 
Dr Niels R. Walet   http://www.phy.umist.ac.uk/Theory/people/walet.html
Dept. of Physics,  UMIST,   P.O. Box 88,    Manchester,  M60 1QD,  U.K.
Phone: +44(0)161-2003693 Fax: +44(0)161-2004303 Nie...@um...

Re: [BProc] Silly errors?

[Vann H. W. <wa...@pr...>]

Try removing the :0 from the config line (which indicates an alias).  If
you are just using eth1 (probably eth0->public and eth1->cluster), eth1:
should work.

Vann

On Mon, 2003-03-24 at 05:13, Niels Walet wrote:
> I am just upgrading my cluster to rh8.0/clustermatic 3, and I am running 
> into a problem i know I have solved in the past, but the trick used in 
> the previous version of clustermatic doesn't seem to work.
> 
> On the head node, when starting I get the error message:
> 
> Starting bpmaster: eth1:0: Cannot assign requested address
> bpmaster: Could not setup socket on interface eth1:0
> 
> I realized I had used the standard network setup under the previous 
> version (i.e., commented out the lines containing the network start-up 
> in  the beowulf script); but either bringing down eth1 and using the 
> default script, or commenting out the same lines as before with eth1 up, 
> doesn't seem to make any difference?!
> 
> Any suggestions?
> Niels

[BProc] Silly errors?

[Niels W. <Nie...@um...>]

I am just upgrading my cluster to rh8.0/clustermatic 3, and I am running 
into a problem i know I have solved in the past, but the trick used in 
the previous version of clustermatic doesn't seem to work.

On the head node, when starting I get the error message:

Starting bpmaster: eth1:0: Cannot assign requested address
bpmaster: Could not setup socket on interface eth1:0

I realized I had used the standard network setup under the previous 
version (i.e., commented out the lines containing the network start-up 
in  the beowulf script); but either bringing down eth1 and using the 
default script, or commenting out the same lines as before with eth1 up, 
doesn't seem to make any difference?!

Any suggestions?
Niels

-- 
Dr Niels R. Walet   http://www.phy.umist.ac.uk/Theory/people/walet.html
Dept. of Physics,  UMIST,   P.O. Box 88,    Manchester,  M60 1QD,  U.K.
Phone: +44(0)161-2003693 Fax: +44(0)161-2004303 Nie...@um...

Re: [BProc] beoboot w/ jumbo frames?

[steven j. <py...@li...>]

Greetings,

I haven't tried it, but it sounds like it should work.

G'day,
sjames


On Sat, 8 Mar 2003, Jim Phillips wrote:

> Hi,
> 
> I'd like to use beoboot on gigabit ethernet with jumbo frames.  AFAIK, two
> nodes with different MTUs can't talk to each other, so the slaves would
> need to start out using jumbo frames in both boot phases.  It looks like
> this could be done by modifying the beoboot source code to add an ioctl
> call in rarp.c: prepare_interfaces() to set the MTU.  Would this work or
> am I missing something?  Is there another way?  Has anyone tried this?
> 
> Thanks.
> 
> -Jim
> 
> 
> 
> -------------------------------------------------------
> This SF.net email is sponsored by: Etnus, makers of TotalView, The debugger 
> for complex code. Debugging C/C++ programs can leave you feeling lost and 
> disoriented. TotalView can help you find your way. Available on major UNIX 
> and Linux platforms. Try it free. www.etnus.com
> _______________________________________________
> BProc-users mailing list
> BPr...@li...
> https://lists.sourceforge.net/lists/listinfo/bproc-users
> 

-- 
-------------------------steven james, director of research, linux labs
... ........ ..... ....                     230 peachtree st nw ste 701
the original linux labs                             atlanta.ga.us 30303
      -since 1995                              http://www.linuxlabs.com
                                   office 404.577.7747 fax 404.577.7743
-----------------------------------------------------------------------

[BProc] beoboot w/ jumbo frames?

[Jim P. <ji...@ks...>]

Hi,

I'd like to use beoboot on gigabit ethernet with jumbo frames.  AFAIK, two
nodes with different MTUs can't talk to each other, so the slaves would
need to start out using jumbo frames in both boot phases.  It looks like
this could be done by modifying the beoboot source code to add an ioctl
call in rarp.c: prepare_interfaces() to set the MTU.  Would this work or
am I missing something?  Is there another way?  Has anyone tried this?

Thanks.

-Jim

[BProc] Re: Node partitions and permissions

[Miguel D. C. <mc...@fc...>]

Hello again.

Regarding my first question, and in case anyone else is having the same
problem, I just saw that it's bjs that's changing the node permissions,
in routine bjs_clean_node.

In bjs.c,

void bjs_clean_node(int node) {
    int i, nprocs;
    struct bproc_proc_info_t *plist;
    bproc_chown(node, 0);
    bproc_chgrp(node, 0);
    bproc_chmod(node, 0100);

Changing this last line to "bproc_chmod(node, 0111);" does the trick.

miguel


>=20
> Message: 1
> From: Miguel Costa <mc...@fc...>
> To: Bproc mailing list <bpr...@li...>
> Organization: Centro de F=EDsica do Porto
> Date: 04 Mar 2003 17:09:38 +0000
> Subject: [BProc] Node partitions and permissions
>=20
> Hello all.
>=20
> I was just asked to install a small cluster of 3 Dual Athlon machines
> so
> I took another shot at Clustermatic - I'm still running Scyld on
> another
> cluster but I'm close to leaving it - ( I can quit any time I want to,
> really :) ).
>=20
> Now, the cluster is up and running but I have two problems:
>=20
> first, although the documentation says that when the nodes are "up"
> the
> permissions are set to execute for all, they are only execute for user
> -
> I have to run bpctl -S all -m 111 manually. It doesn't work to add
> this
> to /etc/beowulf/node_up because the node status is still "boot" at
> this
> point, I guess. I took a look at master.c and it seems to me that the
> node_up_result routine, which sets node status to "up", is setting the
> right permissions (S_IXUSR | S_IXGRP | S_IXOTH), so what is changing
> them back?
>=20
> the second problem is mounting node partitions: I can't get swap to
> work
> but beoboot's setupfs.c says this is on the todo list. However,
> mounting
> a ext2 or ext3 partition also fails, without explanation on the logs.
> /home is nfs mounted from the server, so I just wanted to know if
> setupfs is broken or I'm still doing something wrong.
>=20
> Gotta take a look at bjs, now :)
>=20
> Miguel




--=20
Miguel Dias Costa <mc...@fc...>
Centro de F=EDsica do Porto

Re: [BProc] Node partitions and permissions

[Miguel C. <mc...@fc...>]

Thanks. It's working, and swap too.

What's funny (and rather embarrassing) is that I had spotted that while
browsing setupfs.c. "This can't be right", I thought, "it must be return
0", which had the remarkably useless effect of removing the error status
but not actually continue and mount the partitions! Oh well..

Thanks again.

Miguel

On Tue, 2003-03-04 at 17:31, Willem Schreuder wrote:
> On 4 Mar 2003, Miguel Costa wrote:
>=20
> > the second problem is mounting node partitions: I can't get swap to w=
ork
> > but beoboot's setupfs.c says this is on the todo list. However, mount=
ing
> > a ext2 or ext3 partition also fails, without explanation on the logs.
> > /home is nfs mounted from the server, so I just wanted to know if
> > setupfs is broken or I'm still doing something wrong.
>=20
> There is a bug in using a local disk drive in Clustermatic 3 that is fi=
xed=20
> by the following patch.  Download beoboot-cm source and apply the patch=
. =20
> Basically, there is a stray "return" that needs to be removed.
>=20
> --- beoboot-cm.1.4/node_up/setupfs.c.0  2002-08-07 10:19:03.000000000=20
> -0600
> +++ beoboot-cm.1.4/node_up/setupfs.c    2003-02-01 21:18:53.000000000=20
> -0700
> @@ -291,7 +291,6 @@
>             return -1;
>         }
>=20
> -       return -1;
>      }
>=20
>      /* More special case filesystem types */
--=20
Miguel Costa <mc...@fc...>
Centro de F=EDsica do Porto

Re: [BProc] Node partitions and permissions

[Willem S. <wi...@pr...>]

On 4 Mar 2003, Miguel Costa wrote:

> the second problem is mounting node partitions: I can't get swap to work
> but beoboot's setupfs.c says this is on the todo list. However, mounting
> a ext2 or ext3 partition also fails, without explanation on the logs.
> /home is nfs mounted from the server, so I just wanted to know if
> setupfs is broken or I'm still doing something wrong.

There is a bug in using a local disk drive in Clustermatic 3 that is fixed 
by the following patch.  Download beoboot-cm source and apply the patch.  
Basically, there is a stray "return" that needs to be removed.

--- beoboot-cm.1.4/node_up/setupfs.c.0  2002-08-07 10:19:03.000000000 
-0600
+++ beoboot-cm.1.4/node_up/setupfs.c    2003-02-01 21:18:53.000000000 
-0700
@@ -291,7 +291,6 @@
            return -1;
        }

-       return -1;
     }

     /* More special case filesystem types */

[BProc] Node partitions and permissions

[Miguel C. <mc...@fc...>]

Hello all.

I was just asked to install a small cluster of 3 Dual Athlon machines so
I took another shot at Clustermatic - I'm still running Scyld on another
cluster but I'm close to leaving it - ( I can quit any time I want to,
really :) ).

Now, the cluster is up and running but I have two problems:

first, although the documentation says that when the nodes are "up" the
permissions are set to execute for all, they are only execute for user -
I have to run bpctl -S all -m 111 manually. It doesn't work to add this
to /etc/beowulf/node_up because the node status is still "boot" at this
point, I guess. I took a look at master.c and it seems to me that the
node_up_result routine, which sets node status to "up", is setting the
right permissions (S_IXUSR | S_IXGRP | S_IXOTH), so what is changing
them back?

the second problem is mounting node partitions: I can't get swap to work
but beoboot's setupfs.c says this is on the todo list. However, mounting
a ext2 or ext3 partition also fails, without explanation on the logs.
/home is nfs mounted from the server, so I just wanted to know if
setupfs is broken or I'm still doing something wrong.

Gotta take a look at bjs, now :)

Miguel

--=20
Miguel Costa <mc...@fc...>
Centro de F=EDsica do Porto

Re: [BProc] etherboot image

[Andrew S. <sh...@in...>]

On Thu, 27 Feb 2003 13:30:18 -0500 (EST)
steven james <py...@li...> wrote:

> Greetings,
>
> root=/dev/ram0 should do it.
> G'day,
> sjames

Thanks, I should have remembered that.

Andrew

--
Andrew Shewmaker
Associate Engineer
Phone:  208.526.1415
Fax:  208.526.4017

Idaho National Engineering and Environmental Laboratory
2525 Fremont Ave.
Idaho Falls, ID 83415-3605

Re: [BProc] etherboot image

[steven j. <py...@li...>]

Greetings,

root=/dev/ram0 should do it.
G'day,
sjames


On Thu, 27 Feb 2003, Andrew Shewmaker` wrote:

> Hi,
> 
> I'm building a cluster with NICs already set to use DHCP
> and tftp to boot up.  I told beoboot to create a separate
> phase 2 kernel and ramdisk instead of single image, and
> then I used etherboot's mknbi-linux to make the tagged
> image these cards expect.
> 
> I know that the root filesystem is a tmpfs, but when the
> slave boots it says it mounted root (romfs readonly) and
> then it says it cannot open the root device "nfs" or 00:ff
> and finally it tells me to I need to append a correct
> "root=" option.
> 
> What would the correct one be?  I couldn't tell from
> /etc/beowulf/fstab
> 
> Thanks for any help,
> 
> Andrew
> 
> --
> Andrew Shewmaker
> Associate Engineer
> Phone:  208.526.1415
> Fax:  208.526.4017
> 
> Idaho National Engineering and Environmental Laboratory
> 2525 Fremont Ave.
> Idaho Falls, ID 83415-3605
> 
> 
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> BProc-users mailing list
> BPr...@li...
> https://lists.sourceforge.net/lists/listinfo/bproc-users
> 

-- 
-------------------------steven james, director of research, linux labs
... ........ ..... ....                     230 peachtree st nw ste 701
the original linux labs                             atlanta.ga.us 30303
      -since 1995                              http://www.linuxlabs.com
                                   office 404.577.7747 fax 404.577.7743
-----------------------------------------------------------------------

[BProc] etherboot image

[Andrew S. <sh...@in...>]

Hi,

I'm building a cluster with NICs already set to use DHCP
and tftp to boot up.  I told beoboot to create a separate
phase 2 kernel and ramdisk instead of single image, and
then I used etherboot's mknbi-linux to make the tagged
image these cards expect.

I know that the root filesystem is a tmpfs, but when the
slave boots it says it mounted root (romfs readonly) and
then it says it cannot open the root device "nfs" or 00:ff
and finally it tells me to I need to append a correct
"root=" option.

What would the correct one be?  I couldn't tell from
/etc/beowulf/fstab

Thanks for any help,

Andrew

--
Andrew Shewmaker
Associate Engineer
Phone:  208.526.1415
Fax:  208.526.4017

Idaho National Engineering and Environmental Laboratory
2525 Fremont Ave.
Idaho Falls, ID 83415-3605

Re: [BProc] Multiple Node Masters?

[<er...@he...>]

On Sat, Feb 22, 2003 at 02:15:32PM -0800, Chris Austin wrote:
> 
> Is it possible to configure multiple master nodes?  How are people
> currently handling a beowulf master node failure?

In theory yes but I don't think anybody has written the glue to do
that.

BTW, when I say multiple masters, I mean masters sharing slave nodes,
not two master nodes that can see each other's processes.  A slave
node can run as many copies of the slave daemon as it wants.  The
processes belonging to one master won't be able to interact with the
processes belonging to another master.

Cooking up a beoboot-like boot image to do something like this is more
tricky than it sounds.

- Erik

Re: [BProc] bproc node down detection

[<er...@he...>]

On Fri, Feb 21, 2003 at 04:17:04PM -0800, Chris Austin wrote:
> Hi,
> 
> I'm looking for information on how bproc detects that a node is down and
> how does the bproc API inform the processes that have fork children that
> are on the down node?

Basically, if a slave disconnects or if there's a ping timeout to the
slave, the master daemon on the front end declares it down.  The ping
is inside the TCP connection that BProc uses.

Once a slave is down it notifies other processes that have a
bproc_notifier open (man bproc_notifier or look at bpstat source.)  If
you have a child process on a slave node that goes away for any
reason, the process will appear to be killed by SIGKILL.

- Erik

Re: [BProc] RedHat 8.0 bproc

[<er...@he...>]

On Fri, Feb 21, 2003 at 04:02:44PM -0800, Chris Austin wrote:
> Are there any distributions that include bproc?

Scyld is the only one I'm aware of but I don't know if there's a free
version anymore and the last one I saw was hopelessly out of date.

- Erik

Re: [BProc] RedHat 8.0 bproc

[<ha...@no...>]

>> I'm interested in running an 8.0 box with bproc support.  Is there any
>> howto/postings/rpms for Redhat 8.0?
>> 
>> Or should I be downloading stock 2.4.20 kernel and apply the bproc patch
>> from SourceForge?
>
>Try clustermatic 3 (http://www.clustermatic.org)
>
>It has a bunch of RPMs designed to be dropped on top of a clean RH 8.0
>installation.

For sure Clustermatic is an excellent choice.

Another one is HDDCS: ftp://ftp.harddata.com/pub/hddcs/3.2/2002-12-09/
(or look around at http://www.harddata.com/hddcs/downloads/)

If you happen to have hardware supported by RH kernel (RH call it 2.4.18
but includes tons of patches) and not supported even by 2.4.20, you
might be inclined towards HDDCS which adds bproc to RH kernel (while
Clustermatic adds bproc to vanilla stock kernels).

HDDCS also seems to include some wake-on-lan scripts and PXE boot
support (I did not try these).

I also have an unconfirmed feeling that MPICH via fast ethernet is
different (maybe free of some problems) in HDDCS while Clustermatic is
biased towards troublefree run on Myrinet.


And yet another source of some additional functionality might be to
recompile some source rpms from Scyld. They tend to be ancien and
there might be some work involved making them to work but there is a
lot of them. (I did not try this really long time so I may be totally
wrong here - maybe it is much better than I thing, maybe much worse.)


By the way, it would be nice to have something like "Bproc-based
distributions mixing HOWTO" concentrating this type of information.

Best Regards

Vaclav Hanzl

Re: [BProc] Multiple Node Masters?

[steven j. <py...@li...>]

Greetings,

Generally, I configure node 0 so that it can be easily reconfigured to be
the master. Naturally, any job that was running whene the master failed is
gone.

G'day,
sjames



On 22 Feb 2003, Chris Austin wrote:

> 
> Is it possible to configure multiple master nodes?  How are people
> currently handling a beowulf master node failure?
> 
> Thanks for any insight,
> Chris
> 
> 

-- 
-------------------------steven james, director of research, linux labs
... ........ ..... ....                     230 peachtree st nw ste 701
the original linux labs                             atlanta.ga.us 30303
      -since 1995                              http://www.linuxlabs.com
                                   office 404.577.7747 fax 404.577.7743
-----------------------------------------------------------------------

[BProc] Multiple Node Masters?

[Chris A. <Chr...@dc...>]

Is it possible to configure multiple master nodes?  How are people
currently handling a beowulf master node failure?

Thanks for any insight,
Chris

-- 
Chris Austin <Chr...@dc...>

[BProc] bproc node down detection

[Chris A. <Chr...@dc...>]

Hi,

I'm looking for information on how bproc detects that a node is down and
how does the bproc API inform the processes that have fork children that
are on the down node?

Thanks,
Chris
-- 
Chris Austin <Chr...@dc...>

Re: [BProc] RedHat 8.0 bproc

[Chris A. <Chr...@dc...>]

Are there any distributions that include bproc?

Chris

On Fri, 2003-02-21 at 15:35, Ali Naqi wrote:
> Generally speaking BProc has/had problems with Redhat
> (7.0/7.1/7.2 etc) and i dunno if it's been fixed or
> not. I fixed my own version of bproc and made it run
> (somehow) but if anyone has any idea, if there is an
> official release that's been tested with redhat, it'd
> be helpful.
> 
> --- Chris Austin <Chr...@dc...> wrote:
> > 
> > Hi,
> > 
> > I'm interested in running an 8.0 box with bproc
> > support.  Is there any
> > howto/postings/rpms for Redhat 8.0?
> > 
> > Or should I be downloading stock 2.4.20 kernel and
> > apply the bproc patch
> > from SourceForge?
> > 
> > Thanks in advance for any info,
> > Chris
> > 
> > -- 
> > Chris Austin <Chr...@dc...>
> > 
> > 
> > 
> >
> -------------------------------------------------------
> > This SF.net email is sponsored by: SlickEdit Inc.
> > Develop an edge.
> > The most comprehensive and flexible code editor you
> > can use.
> > Code faster. C/C++, C#, Java, HTML, XML, many more.
> > FREE 30-Day Trial.
> > www.slickedit.com/sourceforge
> > _______________________________________________
> > BProc-users mailing list
> > BPr...@li...
> >
> https://lists.sourceforge.net/lists/listinfo/bproc-users
> 
> 
> __________________________________________________
> Do you Yahoo!?
> Yahoo! Tax Center - forms, calculators, tips, more
> http://taxes.yahoo.com/
> 
> 
> -------------------------------------------------------
> This SF.net email is sponsored by: SlickEdit Inc. Develop an edge.
> The most comprehensive and flexible code editor you can use.
> Code faster. C/C++, C#, Java, HTML, XML, many more. FREE 30-Day Trial.
> www.slickedit.com/sourceforge
> _______________________________________________
> BProc-users mailing list
> BPr...@li...
> https://lists.sourceforge.net/lists/listinfo/bproc-users
-- 
Chris Austin <Chr...@dc...>

Re: [BProc] RedHat 8.0 bproc

[Ali N. <al...@ya...>]

Generally speaking BProc has/had problems with Redhat
(7.0/7.1/7.2 etc) and i dunno if it's been fixed or
not. I fixed my own version of bproc and made it run
(somehow) but if anyone has any idea, if there is an
official release that's been tested with redhat, it'd
be helpful.

--- Chris Austin <Chr...@dc...> wrote:
> 
> Hi,
> 
> I'm interested in running an 8.0 box with bproc
> support.  Is there any
> howto/postings/rpms for Redhat 8.0?
> 
> Or should I be downloading stock 2.4.20 kernel and
> apply the bproc patch
> from SourceForge?
> 
> Thanks in advance for any info,
> Chris
> 
> -- 
> Chris Austin <Chr...@dc...>
> 
> 
> 
>
-------------------------------------------------------
> This SF.net email is sponsored by: SlickEdit Inc.
> Develop an edge.
> The most comprehensive and flexible code editor you
> can use.
> Code faster. C/C++, C#, Java, HTML, XML, many more.
> FREE 30-Day Trial.
> www.slickedit.com/sourceforge
> _______________________________________________
> BProc-users mailing list
> BPr...@li...
>
https://lists.sourceforge.net/lists/listinfo/bproc-users


__________________________________________________
Do you Yahoo!?
Yahoo! Tax Center - forms, calculators, tips, more
http://taxes.yahoo.com/

Re: [BProc] RedHat 8.0 bproc

[<er...@he...>]

On Fri, Feb 21, 2003 at 03:25:18PM -0800, Chris Austin wrote:
> 
> Hi,
> 
> I'm interested in running an 8.0 box with bproc support.  Is there any
> howto/postings/rpms for Redhat 8.0?
> 
> Or should I be downloading stock 2.4.20 kernel and apply the bproc patch
> from SourceForge?

Try clustermatic 3 (http://www.clustermatic.org)

It has a bunch of RPMs designed to be dropped on top of a clean RH 8.0
installation.

- Erik