From makino ＠ giga.it.okayama-u.ac.jp  Mon Jun  2 15:10:59 2003
From: makino ＠ giga.it.okayama-u.ac.jp (Tetsuhisa MAKINO)
Date: Mon, 2 Jun 2003 15:10:59 +0900
Subject: [SCore-users-jp] SMP クラスタ上へのインストールについて
In-Reply-To: <20030530084839.B93C5128944@neal.il.is.s.u-tokyo.ac.jp>
References: <20030530171925.3c3fa418.makino@giga.it.okayama-u.ac.jp>
	<20030530084839.B93C5128944@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <20030602151059.2dd25831.makino@giga.it.okayama-u.ac.jp>

牧野です。

まず、問題と思われる /opt/score/etc/pm-ethernet.conf は以下のようでした。

[root ＠ anzu sbin]# cat ../etc/pm-ethernet.conf
unit 0
# maxnsend 0 - 32
maxnsend 16
# backoff 1000 - 20000 (usec)
backoff 4800
# checksum (0 if off, 1 is on)
checksum 0
# PE    MAC address             base hostname           # comment

どうやら、ノードの一覧がまったくないようなので、以下のように自分で記述しました。
MAC アドレスは /opt/score/etc/scorehosts.db が include していたファイルに
記述されていたものを使っています。(もちろん確認しました)

[root ＠ anzu sbin]# cat ../etc/pm-ethernet.conf
unit 0
# maxnsend 0 - 32
maxnsend 16
# backoff 1000 - 20000 (usec)
backoff 4800
# checksum (0 if off, 1 is on)
checksum 0
# PE    MAC address             base hostname           # comment
0       00:02:B3:40:C6:57       smp0.giga.it.okayama-u.ac.jp
1       00:02:B4:40:C2:AC       smp1.giga.it.okayama-u.ac.jp

このように変更した結果 cluster-network-check の結果が以下のようになりました。

[root ＠ anzu sbin]# cluster-network-check -v -g smpc
smp0.giga.it.okayama-u.ac.jp has 2 cpu, network:ethernet shmem0 shmem1
        scored use ethernet
smp1.giga.it.okayama-u.ac.jp has 2 cpu, network:ethernet shmem0 shmem1
        scored use ethernet
2 hosts has 2 cpu
all hosts has  ethernet
scored use ethernet

問題のない結果に満足し、続いて rpmtest をすると以下のようなメッセージが出力されました。

[root ＠ anzu sbin]# ./rpmtest smp1 ethernet -reply&
[1] 31872
[root ＠ anzu sbin]# Ethernet PM context #0 information (unit 0)
 channel 0 descripter information
  rx_p=00000000, rx_c=00000000, rx_bp=00000000, rx_bc=00000000
  tx_p=00000000, tx_c=00000000, tx_bp=00000000, tx_bc=00000000

 channel 0 statistics information
  st_txmit=00000000, st_rexmit=00000000, st_xmit_ctl=00000000
  st_xmit_ack=00000000 st_xmit_lost=00000000, st_xmit_stop=00000000
  st_xmit_err=00000000, st_xmit_received=00000000, st_rcv_valid=00000000
  st_rcv_ackonly=00000000, st_rcv_igonore=00000000, st_rcv_lose=00000000
  st_rcv_ov=00000000,st_rcv_ov=00000000
  st_rcv_stop=00000000, st_rcv_go=00000000
pmAssocateNodes: Invalid argument(22)

[1]+  Done                    ./rpmtest smp1 ethernet -reply

ほとんど全ての値が 0 である出力でした。引数が不正ということですが、どう不正なのかが分かりません。
そこで、以前の指摘のあった PM_DEBUG を 5 にして rpmtest を実行すると全ての出力値が

0 (nil), 0 00000000, 00000000, 00000000

のようになっていました。

どうかよろしくお願いします。


-- 
岡山大学大学院自然科学研究科博士前期課程2年
Name:	Tetsuhisa Makino(牧野 哲久)
E-MAIL: makino ＠ giga.it.okayama-u.ac.jp


From kameyama ＠ pccluster.org  Mon Jun  2 21:09:21 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Mon, 02 Jun 2003 21:09:21 +0900
Subject: [SCore-users-jp] SMP クラスタ上へのインストールについて
In-Reply-To: Your message of "Mon, 02 Jun 2003 15:10:59 JST."
             <20030602151059.2dd25831.makino@giga.it.okayama-u.ac.jp>
Message-ID: <20030602120738.7A0A8128944@neal.il.is.s.u-tokyo.ac.jp>

亀山です.

In article <20030602151059.2dd25831.makino ＠ giga.it.okayama-u.ac.jp> Tetsuhisa MAKINO <makino ＠ giga.it.okayama-u.ac.jp> wrotes:
> そこで、以前の指摘のあった PM_DEBUG を 5 にして rpmtest を実行すると全ての出
> 力値が
> 
> 0 (nil), 0 00000000, 00000000, 00000000
> 
> のようになっていました。

すみません.
PM_DEBUG を 5 (この場合 1 でもよいかも...) での出力を全部送ってください.
(多分省略した message の中に知りたい情報が入っていると思います.)

                       from Kameyama Toyohisa


From a-hasega ＠ ats.nis.nec.co.jp  Wed Jun  4 23:21:18 2003
From: a-hasega ＠ ats.nis.nec.co.jp (長谷川　篤史)
Date: Wed, 04 Jun 2003 23:21:18 +0900
Subject: [SCore-users-jp] fork(),execl()を使用したプログラムの並列化(2)
References: <002501c32589$5d5ed870$6400a8c0@sein>
Message-ID: <3EDE005E.9040008@ats.nis.nec.co.jp>

増田様、

長谷川＠NEC情報システムズです。

> 
> わかりづらい説明で申し訳ありません。
> ./main はコマンドラインからの実行です。
> 例えば、実行モジュールが
> 　　main
>     subpro1
>     subpro2
> とあった場合に、subpro1,subpro2をmainから
> fork,execlして起動させようとしています。

fork してから、execlで、
  scrun -scored=XXX subpro1
とされたわけですね？

>MPI,OpenMP共に評価した事がないので保証できませんが、fork, execl自体は
>>動作すると思います。ただ、execlで、SCoreのプログラム(OpenMPやMPIで書
かれたプログラム)を実行した場合、動かないはずです。

これは、私の勘違いですので忘れてください。

確認しなおしましたが、Omni/SCASH(SCore版のOmni)では、forkは使えないと
思っておいてください。

Omni/SCASHで生成したプログラムは、実行時に、ユーザが認識していないとこ
ろで通信をして、データの同期、コヒーレンシを確保しています。このため、
forkした後に、子プロセス側で通信を発生させた場合、同期が破綻し、正常に
動作しません。Omni/SCASHのコード生成を理解した上で、明示的に、通信を発
生させないようなプログラムを書くという手はありますが、一般的には、動か
ないと理解しておいてもらった方がよろしいと思います。

> execlのように一般的に使用しているようなシステム関数が
> SCore環境下では動作保証されていないというような情報は
> どこからか入手できるものでしょうか？
> もし、ご存知であれば教えていただければありがたいです。

Omni/SCASHに関しては詳細な記述はありませんので、SCoreの導入元にサポー
トを依頼されるか、ソースを読んで動作を理解して頂いた方が良いと思います。

> SCoreが、そもそも処理の分散化を目的としているわけなので
> 処理分散させる為に、子プロセスやスレッドを生成しているプログラムを
> 流用するという考え方に無理があるのでしょうか？

どういうプログラムなのかわからないのでなんともいえません。
ただ、通常、親プロセスはwaitで、子プロセスを待っていると思いますが、
SCoreのmulti user modeでは、jobをgang schedulingしているため、waitして
いるjobにも一定時間、CPUが割り当てられます。
mainから、subjob1をscrunで実行した場合、それぞれに、50%づつCPUリソース
が割り当てられることになるため、効率が悪くなる可能性があります。

> 一般的には、実行モジュールはひとつにするものなのでしょうか？
> つまり、子プロセスやスレッドを生成する必要性はないものなのでしょうか？

一般的には、増田さんの言われるような方法や、batシステムを使用する方法
をとると思いますが、SCoreのmulti user modeを使われるのであれば、効率的
にも、実行モジュールを1つにするか、jobを実行するだけのプログラムを書い
て、そこから、scrunでjobを実行するようにされた方ががいいのではないで
しょうか？

-- 
長谷川 篤史  E-Mail:a-hasega ＠ ats.nis.nec.co.jp
株式会社NEC情報システムズ 基盤ソフトウェア事業部 サイエンス基盤部
外線:03-3798-9991(Fax.03-3798-9198) / 内線:8-115-2410(Fax.8-115-2419)



From M.Newiger ＠ deltacomputer.de  Fri Jun  6 17:52:39 2003
From: M.Newiger ＠ deltacomputer.de (Martin Newiger)
Date: Fri, 6 Jun 2003 10:52:39 +0200
Subject: [SCore-users-jp] [SCore-users] SCore and Lanai XP
Message-ID: <c=DE%a=_%p=Delta_GmbH%l=ABAKUS-030606085239Z-1178@abakus.delnet>

Hi,

is the new Myricom Lanai XP-Card supported by SCore?

>Regards
>Martin Newiger
>
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Fri Jun  6 18:22:25 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Fri, 06 Jun 2003 18:22:25 +0900
Subject: [SCore-users-jp] Re: [SCore-users] SCore and Lanai XP
In-Reply-To: Your message of "Fri, 06 Jun 2003 10:52:39 JST."
             <c=DE%a=_%p=Delta_GmbH%l=ABAKUS-030606085239Z-1178@abakus.delnet>
Message-ID: <20030606092030.98A20128944@neal.il.is.s.u-tokyo.ac.jp>

In article <c=DE%a=_%p=Delta_GmbH%l=ABAKUS-030606085239Z-1178 ＠ abakus.delnet> Martin Newiger <M.Newiger ＠ deltacomputer.de> wrotes:
> is the new Myricom Lanai XP-Card supported by SCore?

SCore 5.4.0 is not support Myrinet XP-Card.
We will support Myrinet XP card in the next release.

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From n-masuda ＠ sp.nas.nec.co.jp  Fri Jun  6 19:53:52 2003
From: n-masuda ＠ sp.nas.nec.co.jp (増田　尚美)
Date: Fri, 6 Jun 2003 19:53:52 +0900
Subject: [SCore-users-jp] fork(),execl()を使用したプログラムの並列化(2)
References: <002501c32589$5d5ed870$6400a8c0@sein> <3EDE005E.9040008@ats.nis.nec.co.jp>
Message-ID: <000a01c32c19$edae3f10$6400a8c0@sein>

長谷川様

増田です。
お世話になっております。

参考になる情報をいただきありがとうございます。
（探しても探しても得られなかった情報が入手できたという感じです。）
理由付けもはっきりでき、大変助かりました。


> Omni/SCASHで生成したプログラムは、実行時に、ユーザが認識していないとこ
> ろで通信をして、データの同期、コヒーレンシを確保しています。このため、
> forkした後に、子プロセス側で通信を発生させた場合、同期が破綻し、正常に
> 動作しません。Omni/SCASHのコード生成を理解した上で、明示的に、通信を発
> 生させないようなプログラムを書くという手はありますが、一般的には、動か
> ないと理解しておいてもらった方がよろしいと思います。

了解しました。
サンプルプログラムを使用して、実際の動きの確認はできても
内部でどのように動いているのかわからず、釈然としない部分が
多かったのですが、これですっきりしました。


> ただ、通常、親プロセスはwaitで、子プロセスを待っていると思いますが、
> SCoreのmulti user modeでは、jobをgang schedulingしているため、waitして
> いるjobにも一定時間、CPUが割り当てられます。
> mainから、subjob1をscrunで実行した場合、それぞれに、50%づつCPUリソース
> が割り当てられることになるため、効率が悪くなる可能性があります。

アイドル状態であるのに、CPU使用率が高い事が気になっていました。
一定時間ということは、いづれは低くなるのでしょうか？



> 一般的には、増田さんの言われるような方法や、batシステムを使用する方法
> をとると思いますが、SCoreのmulti user modeを使われるのであれば、効率的
> にも、実行モジュールを1つにするか、jobを実行するだけのプログラムを書い
> て、そこから、scrunでjobを実行するようにされた方ががいいのではないで
> しょうか？

了解いたしました。
おそらく、今回のシステムでは、
Omni/SCASH(SCore版のOmni)を採用することになると思います。

この後も、つまずくことが山ほどあると思われます。
どうぞ、今後ともよろしくお願いいたします。

以上です。



From M.Newiger ＠ deltacomputer.de  Thu Jun 12 01:12:29 2003
From: M.Newiger ＠ deltacomputer.de (Martin Newiger)
Date: Wed, 11 Jun 2003 18:12:29 +0200
Subject: [SCore-users-jp] [SCore-users] Installation Problem
Message-ID: <c=DE%a=_%p=Delta_GmbH%l=ABAKUS-030611161229Z-1467@abakus.delnet>

Hi,

I have a 16 nodes cluster and the installation on all nodes except one
works fine. Only one node "hangs" at the welcome screen. What's wrong?

>Regards
>
>Martin Newiger
>
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Thu Jun 12 08:59:21 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Thu, 12 Jun 2003 08:59:21 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Installation Problem
In-Reply-To: Your message of "Wed, 11 Jun 2003 18:12:29 JST."
             <c=DE%a=_%p=Delta_GmbH%l=ABAKUS-030611161229Z-1467@abakus.delnet>
Message-ID: <20030611235709.5764E128944@neal.il.is.s.u-tokyo.ac.jp>

In article <c=DE%a=_%p=Delta_GmbH%l=ABAKUS-030611161229Z-1467 ＠ abakus.delnet> Martin Newiger <M.Newiger ＠ deltacomputer.de> wrotes:
> I have a 16 nodes cluster and the installation on all nodes except one
> works fine. Only one node "hangs" at the welcome screen. What's wrong?

Please look at console 1 and console 3 (ALT-Shift-F3) on the node
to idetify the probrem.

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From Yamamoto.Takaya ＠ wrc.melco.co.jp  Mon Jun 16 17:20:49 2003
From: Yamamoto.Takaya ＠ wrc.melco.co.jp (Takaya Yamamoto)
Date: Mon, 16 Jun 2003 17:20:49 +0900
Subject: [SCore-users-jp] シングル CPUとデュアルCPU
In-Reply-To: <5.0.2.5.2.20030328185453.0333ed58@133.141.16.40>
References: <20030328072750.CA5202005C@neal.il.is.s.u-tokyo.ac.jp>
 <"Your message of Fri, 28 Mar 2003 15:48:25 JST."<5.0.2.5.2.20030328153729.033c49f0@133.141.16.40>
Message-ID: <5.0.2.5.2.20030616171047.03597220@133.141.16.40>

亀山様

山本です。
古いメールですが。。。

教えていただいた方法で、SCore5.0.0をインストールしたのですが、
どうも計算ホストの2x2=4CPUしか認識していないようです。

例えば、/opt/score/exsample/mpi/cpi.cを実行したとき、
　　% scrun cpi
計算ホストの4CPU分しかメッセージが表示されません。

また、/opt/score/exsample/mttl/hello.ccを5ノードで実行しようとしたとき、
　　% scrun -nodes=5 hello
FEP: ERROR SCore-D Login failed: Resource unavailable.
のメッセージが出ます。


% scorehosts -l -g pcc
% sceptic -v -g pcc
% msgb -group pcc
などは、問題なく3ホストが認識されています。

どこを調べればいいか教えていただけませんか？

以上

At 18:55 03/03/28 +0900, Takaya Yamamoto wrote:
>山本です。
>ありがとうございました。
>
>At 16:27 03/03/28 +0900, kameyama ＠ pccluster.org wrote:
>>亀山です.
>>
>>In article <5.0.2.5.2.20030328153729.033c49f0 ＠ 133.141.16.40> Takaya 
>>Yamamoto <Yamamoto.Takaya ＠ wrc.melco.co.jp> wrotes:
>> > 今、
>> > 　サーバー兼計算ホスト：シングルCPU
>> > 　計算ホスト2台：共にデュアルCPU
>> > の3PC（5CPU）の構成にしようとしています。
>> >
>> > EITでインストールしようとしているのですが、
>> > Group Creationのときに、シングルCPUのPCとデュアルCPUのPCを
>> > 同じグループに混在させる方法がわかりません。
>> > どのようにすればいいでしょうか？
>>
>>(直接 scorehosts.db を編集したほうが早いかも知れませんが...)
>>group を 2 つ作成します.
>>まず, SMP だけのグループを作成して. ここには shmem を入れます.
>>次に全部のホストを含む別の group を作成して, そちらには shmem を
>>入れないようにします.
>>
>>最終的な scorehosts.db は network は host ごとに指定されますので,
>>後者のグループを使用すれば, 5 CPU 使用することができると思います.
>>
>>                        from Kameyama Toyohisa
>>_______________________________________________
>>SCore-users-jp mailing list
>>SCore-users-jp ＠ pccluster.org
>>http://www.pccluster.org/mailman/listinfo/score-users-jp



From kameyama ＠ pccluster.org  Mon Jun 16 17:32:32 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Mon, 16 Jun 2003 17:32:32 +0900
Subject: [SCore-users-jp] シングル CPUとデュアルCPU
In-Reply-To: Your message of "Mon, 16 Jun 2003 17:20:49 JST."
             <5.0.2.5.2.20030616171047.03597220@133.141.16.40>
Message-ID: <20030616083006.130C3128944@neal.il.is.s.u-tokyo.ac.jp>

亀山です.

In article <5.0.2.5.2.20030616171047.03597220 ＠ 133.141.16.40> Takaya Yamamoto <Yamamoto.Takaya ＠ wrc.melco.co.jp> wrotes:
> 教えていただいた方法で、SCore5.0.0をインストールしたのですが、
> どうも計算ホストの2x2=4CPUしか認識していないようです。

試しに mul.ti user mode で立ち上げてください.
そうすると, SCore-D がどのように認識しているかがわかります.

> 例えば、/opt/score/exsample/mpi/cpi.cを実行したとき、
> 　　% scrun cpi
> 計算ホストの4CPU分しかメッセージが表示されません。

dual と single が混ざっている cluster は sub cluster に別れていて,
引数なしだと, 一方の sub cluster しか割り当ててくれないと思いました.


> 
> また、/opt/score/exsample/mttl/hello.ccを5ノードで実行しようとしたとき、
> 　　% scrun -nodes=5 hello
> FEP: ERROR SCore-D Login failed: Resource unavailable.
> のメッセージが出ます。

    % scrun -nodes=2x2+1 hello
ではどうでしょうか?

                       from Kameyama Toyohisa


From Yamamoto.Takaya ＠ wrc.melco.co.jp  Mon Jun 16 17:53:59 2003
From: Yamamoto.Takaya ＠ wrc.melco.co.jp (Takaya Yamamoto)
Date: Mon, 16 Jun 2003 17:53:59 +0900
Subject: [SCore-users-jp] シングル  CPUとデュアルCPU
In-Reply-To: <20030616083006.130C3128944@neal.il.is.s.u-tokyo.ac.jp>
References: <"Your message of Mon, 16 Jun 2003 17:20:49 JST."<5.0.2.5.2.20030616171047.03597220@133.141.16.40>
Message-ID: <5.0.2.5.2.20030616174851.0359ae18@133.141.16.40>

亀山様

山本です。

ありがとうございました。
     % scrun -nodes=2x2+1 hello
で5CPUが認識されました。

昨日はエラーをはいて動かなかった（メモしてなかった）のですが、
今は動いています。

以上

At 17:32 03/06/16 +0900, kameyama ＠ pccluster.org wrote:
>亀山です.
>
>In article <5.0.2.5.2.20030616171047.03597220 ＠ 133.141.16.40> Takaya 
>Yamamoto <Yamamoto.Takaya ＠ wrc.melco.co.jp> wrotes:
> > 教えていただいた方法で、SCore5.0.0をインストールしたのですが、
> > どうも計算ホストの2x2=4CPUしか認識していないようです。
>
>試しに mul.ti user mode で立ち上げてください.
>そうすると, SCore-D がどのように認識しているかがわかります.
>
> > 例えば、/opt/score/exsample/mpi/cpi.cを実行したとき、
> > 　　% scrun cpi
> > 計算ホストの4CPU分しかメッセージが表示されません。
>
>dual と single が混ざっている cluster は sub cluster に別れていて,
>引数なしだと, 一方の sub cluster しか割り当ててくれないと思いました.
>
>
> >
> > また、/opt/score/exsample/mttl/hello.ccを5ノードで実行しようとしたとき、
> > 　　% scrun -nodes=5 hello
> > FEP: ERROR SCore-D Login failed: Resource unavailable.
> > のメッセージが出ます。
>
>     % scrun -nodes=2x2+1 hello
>ではどうでしょうか?
>
>                        from Kameyama Toyohisa



From smramesh3 ＠ yahoo.com  Tue Jun 17 16:45:04 2003
From: smramesh3 ＠ yahoo.com (M.Ramesh)
Date: Tue, 17 Jun 2003 00:45:04 -0700 (PDT)
Subject: [SCore-users-jp] [SCore-users] Problem with : Compute hosts boot up IN SCore
Message-ID: <20030617074504.80303.qmail@web20101.mail.yahoo.com>

Hi all..
i am having trouble with Compute host installation after creating a boot
up floppy from server,
The message im getting from compute host when its getting booted form
floppy:

***********************************************************
....
...
...
VFS: Mounted root       (ext2 filesystem)
Using EIT5 feature
mounting /proc filesystem... done
Testing............
No dhcp_server specified. Used Broadcast
setupNetwork cannot set the gateway address
done
NFS mount 192.168.16.10: /mnt/runtime
mount: cannot access portmapper on 192.168.16.10: Connection refused
cannot mount
exiting
See the documentation for this trouble
***********************************************************

Here 192.168.16.10 is my server's ip address;
these server and host r connected thru ethernet switch. <eth0>

Please help me thru this: it would be really great if anyone of u can
suggest me wat may b the reason ...!

Regards and thanks from ,
Ramesh- 


__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Tue Jun 17 17:06:52 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Tue, 17 Jun 2003 17:06:52 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
In-Reply-To: Your message of "Tue, 17 Jun 2003 00:45:04 JST."
             <20030617074504.80303.qmail@web20101.mail.yahoo.com>
Message-ID: <20030617080423.1B780128944@neal.il.is.s.u-tokyo.ac.jp>

In article <20030617074504.80303.qmail ＠ web20101.mail.yahoo.com> "M.Ramesh" <smramesh3 ＠ yahoo.com> wrotes:
> Testing............
> No dhcp_server specified. Used Broadcast
> setupNetwork cannot set the gateway address
> done
> NFS mount 192.168.16.10: /mnt/runtime
> mount: cannot access portmapper on 192.168.16.10: Connection refused
> cannot mount
> exiting
> See the documentation for this trouble
> ***********************************************************
> 
> Here 192.168.16.10 is my server's ip address;
> these server and host r connected thru ethernet switch. <eth0>

Please check on your server:
1. rpc.mountd and nfd is running.
2. No firewall settings in kernel.

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From smramesh3 ＠ yahoo.com  Tue Jun 17 18:09:36 2003
From: smramesh3 ＠ yahoo.com (M.Ramesh)
Date: Tue, 17 Jun 2003 02:09:36 -0700 (PDT)
Subject: [SCore-users-jp] Fwd: Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
Message-ID: <20030617090936.48378.qmail@web20104.mail.yahoo.com>


Note: forwarded message attached.


__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com
-------------- next part --------------
添付メールを保管しました...
送信者: "M.Ramesh" <smramesh3 ＠ yahoo.com>
件名:   Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
日付:   Tue, 17 Jun 2003 02:04:57 -0700 (PDT)
サイズ: 1646 バイト
URL:    <http://new1.pccluster.org/pipermail/score-users-jp/attachments/20030617/ecc151e0/attachment.mht>

From smramesh3 ＠ yahoo.com  Tue Jun 17 18:56:14 2003
From: smramesh3 ＠ yahoo.com (M.Ramesh)
Date: Tue, 17 Jun 2003 02:56:14 -0700 (PDT)
Subject: [SCore-users-jp] Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
In-Reply-To: <20030617092319.78709128944@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <20030617095614.20547.qmail@web20109.mail.yahoo.com>

hi...

both r running...! nfsd, and rpc.mountd.

And, i dont know wats the problem then ?
One mroe thing is when i try with ... /ypinit -m in server,
i get errors like

failed to send 'clear' to local ypserv: RPC: Unknown hostUpdating
passwd.byuid...
and the same for group.byname..., group.bygid....and many!
is this may b the reason ??? wats the way to get rid oif this error?

Ramesh-


--- kameyama ＠ pccluster.org wrote:
> In article <20030617090457.15241.qmail ＠ web20109.mail.yahoo.com>
> "M.Ramesh" <smramesh3 ＠ yahoo.com> wrotes:
> > thanks for the reply...! at present im working on this...!
> > im sure theres no firewall settings...!
> > but i dont know how to check rpc.mountd and nfd is running or not!
> pls,
> > could u tell me how to do this ?
> 
> Please try as following:
>      % ps ax | grep  rpc.mountd
>      % ps ax | grep  nfsd
> 
> > > 1. rpc.mountd and nfd is running.
> 
> Sprry, I'm typo.
>     nfd -> nfsd.
> 
>                        from Kameyama Toyohisa


__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Tue Jun 17 19:00:42 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Tue, 17 Jun 2003 19:00:42 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
In-Reply-To: Your message of "Tue, 17 Jun 2003 02:56:14 JST."
             <20030617095614.20547.qmail@web20109.mail.yahoo.com>
Message-ID: <20030617095813.01D50128944@neal.il.is.s.u-tokyo.ac.jp>

In article <20030617095614.20547.qmail ＠ web20109.mail.yahoo.com> "M.Ramesh" <smramesh3 ＠ yahoo.com> wrotes:
> hi...
> 
> both r running...! nfsd, and rpc.mountd.
> 
> And, i dont know wats the problem then ?
> One mroe thing is when i try with ... /ypinit -m in server,
> i get errors like
> 
> failed to send 'clear' to local ypserv: RPC: Unknown hostUpdating

Please try
    % ps axwww | grep portmap

If portmap us not runnning, please execute following command:
    # /etc/rc.d/init.d/portmap start

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From smramesh3 ＠ yahoo.com  Tue Jun 17 19:13:26 2003
From: smramesh3 ＠ yahoo.com (M.Ramesh)
Date: Tue, 17 Jun 2003 03:13:26 -0700 (PDT)
Subject: [SCore-users-jp] Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
In-Reply-To: <20030617095813.01D50128944@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <20030617101326.36234.qmail@web20110.mail.yahoo.com>

hi...

yea! tis running...! :-( but still im getting same RPC error
messages...!!!

Shall i restart installing from Score5.4.0 - CD once again into /opt
...???

Or before doing that, i have uninstall this somehow ? im really
worried...!

Regards 
Ramesh -

--- kameyama ＠ pccluster.org wrote:
> In article <20030617095614.20547.qmail ＠ web20109.mail.yahoo.com>
> "M.Ramesh" <smramesh3 ＠ yahoo.com> wrotes:
> > hi...
> > 
> > both r running...! nfsd, and rpc.mountd.
> > 
> > And, i dont know wats the problem then ?
> > One mroe thing is when i try with ... /ypinit -m in server,
> > i get errors like
> > 
> > failed to send 'clear' to local ypserv: RPC: Unknown hostUpdating
> 
> Please try
>     % ps axwww | grep portmap
> 
> If portmap us not runnning, please execute following command:
>     # /etc/rc.d/init.d/portmap start
> 
>                        from Kameyama Toyohisa


__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Tue Jun 17 19:18:28 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Tue, 17 Jun 2003 19:18:28 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
In-Reply-To: Your message of "Tue, 17 Jun 2003 03:13:26 JST."
             <20030617101326.36234.qmail@web20110.mail.yahoo.com>
Message-ID: <20030617101558.8761F128944@neal.il.is.s.u-tokyo.ac.jp>

In article <20030617101326.36234.qmail ＠ web20110.mail.yahoo.com> "M.Ramesh" <smramesh3 ＠ yahoo.com> wrotes:
> hi...
> yea! tis running...! :-( but still im getting same RPC error
> messages...!!!
> 
> Shall i restart installing from Score5.4.0 - CD once again into /opt
> ...???

Probably, yes...
Please reboot server to initializ portmapper, and please try again.

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From smramesh3 ＠ yahoo.com  Tue Jun 17 19:30:18 2003
From: smramesh3 ＠ yahoo.com (M.Ramesh)
Date: Tue, 17 Jun 2003 03:30:18 -0700 (PDT)
Subject: [SCore-users-jp] Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
In-Reply-To: <20030617101558.8761F128944@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <20030617103018.64694.qmail@web20106.mail.yahoo.com>

IN my server,
eth0 is connected to computing host (local cluster through switch), eth1
is connected to internet... i mean DNS, (i guess, it is DNS, right ?).

Is it alright ?
if its alrite, i ll try from the scratch once again...!!! But i dont want
to reinstall Linux redhat 7.3 oncer again since i have some softwares
running already...! will it b  alrite?

regards from
Ramesh-


--- kameyama ＠ pccluster.org wrote:
> In article <20030617101326.36234.qmail ＠ web20110.mail.yahoo.com>
> "M.Ramesh" <smramesh3 ＠ yahoo.com> wrotes:
> > hi...
> > yea! tis running...! :-( but still im getting same RPC error
> > messages...!!!
> > 
> > Shall i restart installing from Score5.4.0 - CD once again into /opt
> > ...???
> 
> Probably, yes...
> Please reboot server to initializ portmapper, and please try again.
> 
>                        from Kameyama Toyohisa


__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Tue Jun 17 19:39:47 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Tue, 17 Jun 2003 19:39:47 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
In-Reply-To: Your message of "Tue, 17 Jun 2003 03:30:18 JST."
             <20030617103018.64694.qmail@web20106.mail.yahoo.com>
Message-ID: <20030617103717.D1E32128944@neal.il.is.s.u-tokyo.ac.jp>

In article <20030617103018.64694.qmail ＠ web20106.mail.yahoo.com> "M.Ramesh" <smramesh3 ＠ yahoo.com> wrotes:
> IN my server,
> eth0 is connected to computing host (local cluster through switch), eth1
> is connected to internet... i mean DNS, (i guess, it is DNS, right ?).

It meen that if you want to access other host to compute hosts,
the server routes packet from otherhost to compute hosts.
And If you want to access DNS on compute hosts,
you must set gateway (Please look at Figure 3 in
    http://www.pccluster.org/score/dist/score/html/en/installation/eit.html
) to server's eth0 IP address.

> if its alrite, i ll try from the scratch once again...!!! But i dont want
> to reinstall Linux redhat 7.3 oncer again since i have some softwares
> running already...! will it b  alrite?

You don't need to reinstall redhat.

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From smramesh3 ＠ yahoo.com  Tue Jun 17 20:11:05 2003
From: smramesh3 ＠ yahoo.com (M.Ramesh)
Date: Tue, 17 Jun 2003 04:11:05 -0700 (PDT)
Subject: [SCore-users-jp] Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
In-Reply-To: <20030617103717.D1E32128944@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <20030617111105.29812.qmail@web20109.mail.yahoo.com>

hi...

so for Name Server , which address i shud give ?

eth1 is connected to internet.
eth0 is connected to the computing host. 
eth0 details :
(140.168.46.10,192.168.46.255,255.255.255.0)
eth1 details:
(140.158.46.177,140.158.46.177,140.158.46.255,255.255.254.0)

in this case, for Name Server , which address i shud give ?
By default , it appears like, 
*************************************
server name : cosmos.lamar.edu
domain name: lamar.edu
Netmask : 255.255.255.0
Gateway: 192.168.16.255
NIS : lamar.edu
Name server : 140.158.208.164 ( - Note:which is the add of Primary DNS as
it appears in Network Configuration wondow. )
mount point : /mnt/cdrom
Display : cosmos.lamar.edu
*************************************
And local host address 127.0.01 shud be mentioned anywhere in the files in
server ???

and , compute host shud be connected to server via eth0 or eth1 ???      
(thru 'switch to server).

regards -
Ramesh -

--- kameyama ＠ pccluster.org wrote:
> In article <20030617103018.64694.qmail ＠ web20106.mail.yahoo.com>
> "M.Ramesh" <smramesh3 ＠ yahoo.com> wrotes:
> > IN my server,
> > eth0 is connected to computing host (local cluster through switch),
> eth1
> > is connected to internet... i mean DNS, (i guess, it is DNS, right ?).
> 
> It meen that if you want to access other host to compute hosts,
> the server routes packet from otherhost to compute hosts.
> And If you want to access DNS on compute hosts,
> you must set gateway (Please look at Figure 3 in
>    
> http://www.pccluster.org/score/dist/score/html/en/installation/eit.html
> ) to server's eth0 IP address.
> 
> > if its alrite, i ll try from the scratch once again...!!! But i dont
> want
> > to reinstall Linux redhat 7.3 oncer again since i have some softwares
> > running already...! will it b  alrite?
> 
> You don't need to reinstall redhat.
> 
>                        from Kameyama Toyohisa


__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Tue Jun 17 20:21:24 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Tue, 17 Jun 2003 20:21:24 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
In-Reply-To: Your message of "Tue, 17 Jun 2003 04:11:05 JST."
             <20030617111105.29812.qmail@web20109.mail.yahoo.com>
Message-ID: <20030617111854.C450A128944@neal.il.is.s.u-tokyo.ac.jp>

In article <20030617111105.29812.qmail ＠ web20109.mail.yahoo.com> "M.Ramesh" <smramesh3 ＠ yahoo.com> wrotes:
> hi...
> 
> so for Name Server , which address i shud give ?
> 
> eth1 is connected to internet.
> eth0 is connected to the computing host. 
> eth0 details :
> (140.168.46.10,192.168.46.255,255.255.255.0)
> eth1 details:
> (140.158.46.177,140.158.46.177,140.158.46.255,255.255.254.0)

In first message you wrote:
> Here 192.168.16.10 is my server's ip address;

eth0 is (140.168.16.0,192.168.16.255, 255.255.255.0)?


> in this case, for Name Server , which address i shud give ?
> By default , it appears like, 
> *************************************
> server name : cosmos.lamar.edu
> domain name: lamar.edu
> Netmask : 255.255.255.0
> Gateway: 192.168.16.255

If you want to access DNS, and the server forwarding to other networks,
Gateway must be 192.168.16.0 (server's IP address).

> and , compute host shud be connected to server via eth0 or eth1 ???      

If you want to use EIT, compute hosts connected to server via eth0.

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From smramesh3 ＠ yahoo.com  Tue Jun 17 20:28:52 2003
From: smramesh3 ＠ yahoo.com (M.Ramesh)
Date: Tue, 17 Jun 2003 04:28:52 -0700 (PDT)
Subject: [SCore-users-jp] Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
In-Reply-To: <20030617111854.C450A128944@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <20030617112852.44458.qmail@web20107.mail.yahoo.com>

hi

yea, my servers ip address is still the same! im sorry for the confusion.

eth0 details :(140.168.46.10,192.168.46.255,255.255.255.0, repectively, ip
address of server, Bcast, Mask).

Gateway thing is clear for me now!!! but im having doubt on Name serever
still ? pls, reply me ! :-(

For Name Server , which address i shud give in the figure-(3) in 
<http://www.pccluster.org/score/dist/score/html/en/installation/eit.html>?
Name Server is the server's ip address, simply ???

Regards from 
Ramesh -

--- kameyama ＠ pccluster.org wrote:
> In article <20030617111105.29812.qmail ＠ web20109.mail.yahoo.com>
> "M.Ramesh" <smramesh3 ＠ yahoo.com> wrotes:
> > hi...
> > 
> > so for Name Server , which address i shud give ?
> > 
> > eth1 is connected to internet.
> > eth0 is connected to the computing host. 
> > eth0 details :
> > (140.168.46.10,192.168.46.255,255.255.255.0)
> > eth1 details:
> > (140.158.46.177,140.158.46.177,140.158.46.255,255.255.254.0)
> 
> In first message you wrote:
> > Here 192.168.16.10 is my server's ip address;
> 
> eth0 is (140.168.16.0,192.168.16.255, 255.255.255.0)?
> 
> 
> > in this case, for Name Server , which address i shud give ?
> > By default , it appears like, 
> > *************************************
> > server name : cosmos.lamar.edu
> > domain name: lamar.edu
> > Netmask : 255.255.255.0
> > Gateway: 192.168.16.255
> 
> If you want to access DNS, and the server forwarding to other networks,
> Gateway must be 192.168.16.0 (server's IP address).
> 
> > and , compute host shud be connected to server via eth0 or eth1 ???   
>   
> 
> If you want to use EIT, compute hosts connected to server via eth0.
> 
>                        from Kameyama Toyohisa


__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Tue Jun 17 20:34:32 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Tue, 17 Jun 2003 20:34:32 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
In-Reply-To: Your message of "Tue, 17 Jun 2003 04:28:52 JST."
             <20030617112852.44458.qmail@web20107.mail.yahoo.com>
Message-ID: <20030617113202.8B6C9128944@neal.il.is.s.u-tokyo.ac.jp>

In article <20030617112852.44458.qmail ＠ web20107.mail.yahoo.com> "M.Ramesh" <smramesh3 ＠ yahoo.com> wrotes:
> For Name Server , which address i shud give in the figure-(3) in 
> <http://www.pccluster.org/score/dist/score/html/en/installation/eit.html>?
> Name Server is the server's ip address, simply ???

If name server is running on the server, this is true.
If you don't use name server on compute hosts,
please execute eit on -nisonly option:
    #  /opt/score/bin/eit -nisonly
and please don't wite name server.

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From smramesh3 ＠ yahoo.com  Tue Jun 17 20:53:01 2003
From: smramesh3 ＠ yahoo.com (M.Ramesh)
Date: Tue, 17 Jun 2003 04:53:01 -0700 (PDT)
Subject: [SCore-users-jp] Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
In-Reply-To: <20030617113202.8B6C9128944@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <20030617115301.24328.qmail@web20102.mail.yahoo.com>

hi

im sorry ! now im getting mesages after i booted the compute host from
floppy,  (after many messages...)

**********************************
...
...
...
mounting /proc filesystem...done
testing.......
No dhcp_server specified. Used Broadcast
No DHCP reply received
Try it again
No DHCP reply received
Try it again

Configure Networks fails
done
NFS mount : /mnt/runtime
cannot mount
exiting

**********************************

PLs, suggest me wat to do now ?
regards
Ramesh -


--- kameyama ＠ pccluster.org wrote:
> In article <20030617112852.44458.qmail ＠ web20107.mail.yahoo.com>
> "M.Ramesh" <smramesh3 ＠ yahoo.com> wrotes:
> > For Name Server , which address i shud give in the figure-(3) in 
> >
>
<http://www.pccluster.org/score/dist/score/html/en/installation/eit.html>?
> > Name Server is the server's ip address, simply ???
> 
> If name server is running on the server, this is true.
> If you don't use name server on compute hosts,
> please execute eit on -nisonly option:
>     #  /opt/score/bin/eit -nisonly
> and please don't wite name server.
> 
>                        from Kameyama Toyohisa



_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Tue Jun 17 21:11:46 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Tue, 17 Jun 2003 21:11:46 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
In-Reply-To: Your message of "Tue, 17 Jun 2003 04:53:01 JST."
             <20030617115301.24328.qmail@web20102.mail.yahoo.com>
Message-ID: <20030617120916.E22E0128944@neal.il.is.s.u-tokyo.ac.jp>

In article <20030617115301.24328.qmail ＠ web20102.mail.yahoo.com> "M.Ramesh" <smramesh3 ＠ yahoo.com> wrotes:
> im sorry ! now im getting mesages after i booted the compute host from
> floppy,  (after many messages...)
> 
> **********************************
> ...
> ...
> ...
> mounting /proc filesystem...done
> testing.......
> No dhcp_server specified. Used Broadcast
> No DHCP reply received
> Try it again
> No DHCP reply received
> Try it again
> 
> Configure Networks fails

It meens EIT's dhcp server is not responce anything.
Probably cluster setup0 window
(http://www.pccluster.org/score/dist/score/html/en/installation/eit.html
Figure 17) is blank.

Please check network settings.
(ip address and netmask for compute hosts.)

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From smramesh3 ＠ yahoo.com  Tue Jun 17 21:19:04 2003
From: smramesh3 ＠ yahoo.com (M.Ramesh)
Date: Tue, 17 Jun 2003 05:19:04 -0700 (PDT)
Subject: [SCore-users-jp] Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
In-Reply-To: <20030617120916.E22E0128944@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <20030617121904.27410.qmail@web20102.mail.yahoo.com>

hi

what do u mean by setup0 is blank ???
and where to check the compute host's ip address and net mask ?? is that
in figure (3) ???
pls, reply me, sir! waiting for ur reply !
regards,
Ramesh -


--- kameyama ＠ pccluster.org wrote:
> In article <20030617115301.24328.qmail ＠ web20102.mail.yahoo.com>
> "M.Ramesh" <smramesh3 ＠ yahoo.com> wrotes:
> > im sorry ! now im getting mesages after i booted the compute host from
> > floppy,  (after many messages...)
> > 
> > **********************************
> > ...
> > ...
> > ...
> > mounting /proc filesystem...done
> > testing.......
> > No dhcp_server specified. Used Broadcast
> > No DHCP reply received
> > Try it again
> > No DHCP reply received
> > Try it again
> > 
> > Configure Networks fails
> 
> It meens EIT's dhcp server is not responce anything.
> Probably cluster setup0 window
> (http://www.pccluster.org/score/dist/score/html/en/installation/eit.html
> Figure 17) is blank.
> 
> Please check network settings.
> (ip address and netmask for compute hosts.)
> 
>                        from Kameyama Toyohisa



_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From a-hasega ＠ ats.nis.nec.co.jp  Tue Jun 17 22:41:02 2003
From: a-hasega ＠ ats.nis.nec.co.jp (長谷川　篤史)
Date: Tue, 17 Jun 2003 22:41:02 +0900
Subject: [SCore-users-jp] fork(),execl()を使用したプログラムの並列化(2)
References: <002501c32589$5d5ed870$6400a8c0@sein> <3EDE005E.9040008@ats.nis.nec.co.jp> <000a01c32c19$edae3f10$6400a8c0@sein>
Message-ID: <3EEF1A6E.9060408@ats.nis.nec.co.jp>

長谷川＠NEC情報システムズです。

> アイドル状態であるのに、CPU使用率が高い事が気になっていました。
> 一定時間ということは、いづれは低くなるのでしょうか？

いえ、そうではなく、SCoreとして、リソースは割り当てますが、
waitしているのなら、その間、CPUリソースを使用しないため、
  job 1 : 50%
  job 2 :  0%
のように。なるはずです
CPU使用率が高いということは、waitせずに別の処理をしているのではないで
しょうか？
  % gdb PROGRAM PID
として、debuggerで動作を確認されたほうがよいと思います。

SCASHのドキュメントを読んでもらった方がいいと思いますが、
SCASH側の制限により、Omni/SCASHにも、いくつか制限事項が存在します。

-- 
長谷川 篤史  E-Mail:a-hasega ＠ ats.nis.nec.co.jp
株式会社NEC情報システムズ 基盤ソフトウェア事業部 サイエンス基盤部
外線:03-3798-9991(Fax.03-3798-9198) / 内線:8-115-2410(Fax.8-115-2419)



From nick ＠ streamline-computing.com  Wed Jun 18 02:39:30 2003
From: nick ＠ streamline-computing.com (Nick Birkett)
Date: Tue, 17 Jun 2003 18:39:30 +0100
Subject: [SCore-users-jp] [SCore-users] ERROR dup2(null,0)
Message-ID: <200306171839.30479.nick@streamline-computing.com>

I get this error on one of our new clusters (myrinet 2k C pards).

<6> SCore-D:ERROR dup2(null,0)

Any ideas what this mighr be ? Hardware problem ?

Cheers,

Nick

_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Wed Jun 18 09:42:16 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Wed, 18 Jun 2003 09:42:16 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
In-Reply-To: Your message of "Tue, 17 Jun 2003 05:19:04 JST."
             <20030617121904.27410.qmail@web20102.mail.yahoo.com>
Message-ID: <20030618003945.63D85128944@neal.il.is.s.u-tokyo.ac.jp>

In article <20030617121904.27410.qmail ＠ web20102.mail.yahoo.com> "M.Ramesh" <smramesh3 ＠ yahoo.com> wrotes:
> what do u mean by setup0 is blank ???

When you begin to install copiute hosts,
EIT displays Cluster setup window.
(In
    http://www.pccluster.org/score/dist/score/html/en/installation/eit.html
Figure 17)

If the compute host send DHCP request packet for search server hosts,
and that packet recieve the server, EIT displays MAC address and
allocated MAC address in the setup window.
(like Figure 18.)

If the setup window dose not display MAC address,
the compute host cannot send DHCP request packet or the packet
dose not reach from server.
But this step is successful in previous...


Sorry, please check this check list.
1. Compute hosts connected correct network?
   If you want to use EIT, compute hosts connected the network
   that server host connected by eth0.
2. Network onnection is correct?
   Figure 3 setting is compute hosts.
3. NIS setting is correct?
   If you want to -nisonly  setting, all compute hosts is registered
   to the NIS hosts database.
4. Select Boot Network Device (Figure 14) is correct?
   You select floppy image in this window.
   The floppy disk image dose not change when you change EIT setting.
   So if you keep privious floppy image, you can use the floppy.
5. Please check EIT found network card.
   Please look at console 3 (press ALT-SHIFT-F3) in the compute host.
   This console shows EIT output.
   If EIT found any network card, EIT load the network driver module
   for the NIC.
6. Please check kernel fond network card.
   Please look at console 4 (press ALT-SHIFT-F4) in the compute host.
   This console shows kernel messages.
   If kernel fails to load network driver, you can found this console.

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From hori ＠ swimmy-soft.com  Wed Jun 18 09:46:32 2003
From: hori ＠ swimmy-soft.com (Atsushi HORI)
Date: Wed, 18 Jun 2003 09:46:32 +0900
Subject: [SCore-users-jp] [SCore-users] Re: [score-info] score problem
In-Reply-To: <20030617183811.GA5490@titan.cse.psu.edu>
References: <20030617183811.GA5490@titan.cse.psu.edu>
Message-ID: <3138774392.hori0000@swimmy-soft.com>

Hi,

>We are developing several different coscheduling algorithms for 
>cluster using myrinet network interfaces. 

Sounds interesting.

Which Myrinet ? Myrinet 2K ? How did you specify the myrinet in the 
/opt/score/etc/scorehost.db file ?

----
Atsushi HORI
Swimmy Software, Inc.

_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From hori ＠ swimmy-soft.com  Wed Jun 18 09:58:39 2003
From: hori ＠ swimmy-soft.com (Atsushi HORI)
Date: Wed, 18 Jun 2003 09:58:39 +0900
Subject: [SCore-users-jp] Re: [SCore-users] ERROR dup2(null,0)
In-Reply-To: <200306171839.30479.nick@streamline-computing.com>
References: <200306171839.30479.nick@streamline-computing.com>
Message-ID: <3138775119.hori0002@swimmy-soft.com>

Hi, Nick,

>I get this error on one of our new clusters (myrinet 2k C pards).
>
><6> SCore-D:ERROR dup2(null,0)
>
>Any ideas what this mighr be ? Hardware problem ?

Sounds like a software problem. Can you re-produce the same error ? 
If so, would you like to tell me more detail configuration and 
situation ? 

----
Atsushi HORI
Swimmy Software, Inc.

_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From haddock ＠ webgroup.co.jp  Wed Jun 18 13:54:47 2003
From: haddock ＠ webgroup.co.jp (=?iso-2022-jp?b?aGFkZG9jayAbJEIhdxsoQiB3ZWJncm91cC5jby5qcA==?=)
Date: Wed, 18 Jun 2003 13:54:47 +0900
Subject: [SCore-users-jp] [SCore-users] 将来像
Message-ID: <6abd64d.64d6abd@webgroup.co.jp>

みなさんこんにちは、
　　
　　　　Scoreに関する100%の技術問題ではありませんが、
どなたかにScoreをはじめとするcluster wareの将来像について
語っていただければと思います。
　　・grid (or cluster)技術の寿命は？
　　
　　・grid , clusterについて、大学等での最先端の研究方向は？

　　・grid , cluster用のprotocol, OSの研究方向は？
　　　既存のprotocol , OSにmiddlewareを乗っけて走らせると　
　　　いうのは、既存リソースの有効利用に役立つと思うが、
　　　これ以上のメリットはあるのでしょうか。

　　

_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From smramesh3 ＠ yahoo.com  Thu Jun 19 08:57:23 2003
From: smramesh3 ＠ yahoo.com (M.Ramesh)
Date: Wed, 18 Jun 2003 16:57:23 -0700 (PDT)
Subject: [SCore-users-jp] Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
In-Reply-To: <20030618003945.63D85128944@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <20030618235723.30393.qmail@web20107.mail.yahoo.com>

hi..

Here now i could see the compute hosts mac address in the Figure 18 at
<http://www.pccluster.org/score/dist/score/html/en/installation/eit.html>.

But in the compute host, im getting message like 

***************************************
No dhcp_server specified. used Broadcast.
SetupNetwork cannot set the Gateway address
done
NFS mount ***.***.**.** : /opt/score/setup//Redhat/instimage/ /mnt/runtime
mount: cannot access the portmap on ***.***.**.** : Connection rRefused
cannot mount /opt/score/setup//Redhat/instimage/
exiting
See the documentation for this trouble
**********************************************************

Here ***.***.**.** is my server's eth0 IP address.

At this time, i pressed alt F4, then i see some messages, like...
******************************************************
probing buses.
.
.
.
found 3c59xdevice
found suggestion of usb-ohci
found devices justProbe is 0
going to insmod sc59x.o (path is NULL)
boot_file= /opt/score/setup//Redhat/instimage/:/mnt/cdrom/
sys_path= /opt/score/setup//Redhat/instimage/  source_path= /mnt/cdrom/
DEVICE =eth0 IPADDR = 192.168.16.11 NETMASK=  192.168.16.11
GATEWAY=0.0.0.0
HOSTNAME=0
DOMAIN:<my_right_domain_name>
writing /tmp/modules.conf
going to insmod sunrpc.o (path is NULL)
going to insmod lockd.o (path is NULL)
going to insmod nfs.o (path is NULL)
******************************************************

ALT F4 shows page of messages like

******************************************************
hda: ....
hdc: ....
ide0 at ....
hda: ...
hdc: ....cd-rom drive...
partition check
hda: hda1 hda2 hda3
Floppy drive fd0 is 1.44
RAMDISK

autorun DONE...

IP protocols
TCP....

VFS: mounted root
PCI: Found IRQ 10 with ...

#com PCI 3c905C Tornado at 0x3000. Ver LK1.1.16
******************************************************
I dont understand wat it means ? pls have a look at it. and suggest me
somehting...! pls!


please, let me know wat to do now to get rid of this.
Regards,
Ramesh-


******************************************************************



--- kameyama ＠ pccluster.org wrote:
> In article <20030617121904.27410.qmail ＠ web20102.mail.yahoo.com>
> "M.Ramesh" <smramesh3 ＠ yahoo.com> wrotes:
> > what do u mean by setup0 is blank ???
> 
> When you begin to install copiute hosts,
> EIT displays Cluster setup window.
> (In
>    
> http://www.pccluster.org/score/dist/score/html/en/installation/eit.html
> Figure 17)
> 
> If the compute host send DHCP request packet for search server hosts,
> and that packet recieve the server, EIT displays MAC address and
> allocated MAC address in the setup window.
> (like Figure 18.)
> 
> If the setup window dose not display MAC address,
> the compute host cannot send DHCP request packet or the packet
> dose not reach from server.
> But this step is successful in previous...
> 
> 
> Sorry, please check this check list.
> 1. Compute hosts connected correct network?
>    If you want to use EIT, compute hosts connected the network
>    that server host connected by eth0.
> 2. Network onnection is correct?
>    Figure 3 setting is compute hosts.
> 3. NIS setting is correct?
>    If you want to -nisonly  setting, all compute hosts is registered
>    to the NIS hosts database.
> 4. Select Boot Network Device (Figure 14) is correct?
>    You select floppy image in this window.
>    The floppy disk image dose not change when you change EIT setting.
>    So if you keep privious floppy image, you can use the floppy.
> 5. Please check EIT found network card.
>    Please look at console 3 (press ALT-SHIFT-F3) in the compute host.
>    This console shows EIT output.
>    If EIT found any network card, EIT load the network driver module
>    for the NIC.
> 6. Please check kernel fond network card.
>    Please look at console 4 (press ALT-SHIFT-F4) in the compute host.
>    This console shows kernel messages.
>    If kernel fails to load network driver, you can found this console.
> 
>                        from Kameyama Toyohisa


__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Thu Jun 19 10:35:54 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Thu, 19 Jun 2003 10:35:54 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
In-Reply-To: Your message of "Wed, 18 Jun 2003 16:57:23 JST."
             <20030618235723.30393.qmail@web20107.mail.yahoo.com>
Message-ID: <20030619013319.8C93C128944@neal.il.is.s.u-tokyo.ac.jp>

In article <20030618235723.30393.qmail ＠ web20107.mail.yahoo.com> "M.Ramesh" <smramesh3 ＠ yahoo.com> wrotes:
> ***************************************
> No dhcp_server specified. used Broadcast.
> SetupNetwork cannot set the Gateway address
> done
> NFS mount ***.***.**.** : /opt/score/setup//Redhat/instimage/ /mnt/runtime
> mount: cannot access the portmap on ***.***.**.** : Connection rRefused
> cannot mount /opt/score/setup//Redhat/instimage/

Please check server:
1. Please execute following command:
      % ypcat hosts
   This command is tested NIS client is running.
   But NIS and NFS use SUN RPC (and SUN RPC use portmap daemon).
2. Please execute following command by root:
      # mount localhost:/opt/score5.4.0 /mnt/floppy
   This command is tested NFS server is running.
3. Please execute following command by root:
      # /etc/rc.d/init.d/ipchains status
      # /etc/rc.d/init.d/iptables status
   This command checks firewall settings.

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From smramesh3 ＠ yahoo.com  Thu Jun 19 10:51:34 2003
From: smramesh3 ＠ yahoo.com (M.Ramesh)
Date: Wed, 18 Jun 2003 18:51:34 -0700 (PDT)
Subject: [SCore-users-jp] Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
In-Reply-To: <20030619013319.8C93C128944@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <20030619015134.68124.qmail@web20106.mail.yahoo.com>

hi...

*********************************************
 # /etc/rc.d/init.d/ipchains status
gives the followong messages...!

ipchains: Incompatible with this kernel
*********************************************

Is this a problem ?

And,

# /etc/rc.d/init.d/iptables status
gives long output...! i dont think other stuffs r problems!!! wat to check
on this firewall ?



Ramesh-


--- kameyama ＠ pccluster.org wrote:
> In article <20030618235723.30393.qmail ＠ web20107.mail.yahoo.com>
> "M.Ramesh" <smramesh3 ＠ yahoo.com> wrotes:
> > ***************************************
> > No dhcp_server specified. used Broadcast.
> > SetupNetwork cannot set the Gateway address
> > done
> > NFS mount ***.***.**.** : /opt/score/setup//Redhat/instimage/
> /mnt/runtime
> > mount: cannot access the portmap on ***.***.**.** : Connection
> rRefused
> > cannot mount /opt/score/setup//Redhat/instimage/
> 
> Please check server:
> 1. Please execute following command:
>       % ypcat hosts
>    This command is tested NIS client is running.
>    But NIS and NFS use SUN RPC (and SUN RPC use portmap daemon).
> 2. Please execute following command by root:
>       # mount localhost:/opt/score5.4.0 /mnt/floppy
>    This command is tested NFS server is running.
> 3. Please execute following command by root:
>       # /etc/rc.d/init.d/ipchains status
>       # /etc/rc.d/init.d/iptables status
>    This command checks firewall settings.
> 
>                        from Kameyama Toyohisa


__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Thu Jun 19 11:05:37 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Thu, 19 Jun 2003 11:05:37 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Problem with : Compute hosts boot up IN SCore
In-Reply-To: Your message of "Wed, 18 Jun 2003 18:51:34 JST."
             <20030619015134.68124.qmail@web20106.mail.yahoo.com>
Message-ID: <20030619020302.73D83128944@neal.il.is.s.u-tokyo.ac.jp>

In article <20030619015134.68124.qmail ＠ web20106.mail.yahoo.com> "M.Ramesh" <smramesh3 ＠ yahoo.com> wrotes:
> *********************************************
>  # /etc/rc.d/init.d/ipchains status
> gives the followong messages...!
> 
> ipchains: Incompatible with this kernel
> *********************************************
> 
> Is this a problem ?

It is not a problem.

> # /etc/rc.d/init.d/iptables status
> gives long output...! i dont think other stuffs r problems!!! wat to check
> on this firewall ?

Probably, this is the probrem.
Probably, server's kernel ddose not accept NFS request by iptable setting.
Please accept all ports at lease eth0 for compute hosts.

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From hqzhou ＠ nju.edu.cn  Thu Jun 19 12:17:34 2003
From: hqzhou ＠ nju.edu.cn (Huiqun Zhou)
Date: Thu, 19 Jun 2003 11:17:34 +0800
Subject: [SCore-users-jp] [SCore-users] Help: using D-Link Gigabit NIC DGE-550T
Message-ID: <005a01c33611$53c1ce60$1a00a8c0@goofy>

Hi, Score users,

Does anyone have experience of using D-Link's Gigabit NIC DGE-550T with RedHat 7.3 and Score 5.4.0?
I've grabed source code of the driver from D-Link's website, but I got errors when I try to insmod the
compiled module into the kernel. 

Regards,

----------------------------------
Huiqun Zhou, Dr. of Science
Department of Earth Sciences
Nanjing University
China

e-mail: hqzhou ＠ nju.edu.cn
Tel: 86(25)359-4664
FAX: 86(25)368-6016
----------------------------------
-------------- next part --------------
HTMLの添付ファイルを保管しました...
URL: <http://new1.pccluster.org/pipermail/score-users-jp/attachments/20030619/fe046a8a/attachment.html>

From kameyama ＠ pccluster.org  Thu Jun 19 14:06:28 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Thu, 19 Jun 2003 14:06:28 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Help: using D-Link Gigabit NIC DGE-550T
In-Reply-To: Your message of "Thu, 19 Jun 2003 11:17:34 JST."
             <005a01c33611$53c1ce60$1a00a8c0@goofy>
Message-ID: <20030619050353.01A5B128944@neal.il.is.s.u-tokyo.ac.jp>

In article <005a01c33611$53c1ce60$1a00a8c0 ＠ goofy> "Huiqun Zhou" <hqzhou ＠ nju.edu.cn> wrotes:
> Does anyone have experience of using D-Link's Gigabit NIC DGE-550T with RedHat
> 7.3 and Score 5.4.0?
> I've grabed source code of the driver from D-Link's website, but I got errors
> when I try to insmod the compiled module into the kernel. 

If yoy want to compile kernel modules for compute hosts, you must install
kernel source and compile the modules with that kernel source.
Please install kernel-source-score-2.4.19-1SCORE.i386.rpm
(This file located /mnt/cdrom/score.rpm in SCore 5.4.0 CD-ROM or
following URL.
    http://www.pccluster.org//score/dist/pub/score-5.4.0/rpm.redhat7.3.i386
)
on server host.
When you install that rpm, kernel source is installed in
/usr/src/linux-2.4.19score.
So you can compile that module with following option for compute hosts:
   -I/usr/src/linux-2.4.19score/include

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From hqzhou ＠ nju.edu.cn  Thu Jun 19 16:44:29 2003
From: hqzhou ＠ nju.edu.cn (Huiqun Zhou)
Date: Thu, 19 Jun 2003 15:44:29 +0800
Subject: [SCore-users-jp] Re: [SCore-users] Help: using D-Link Gigabit NIC DGE-550T
References: <20030619050353.01A5B128944@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <00a601c33636$9dea7cb0$1a00a8c0@goofy>

Kameyama-san,

Thank you for your reply.

Just want to clarify one thing further. Do you mean that if I want to modify or add a new NIC driver
into Linux kernel, I must compile it together with the relevant Score source code as you indicated?
Put question in another way, does Score have its own changed NIC drivers, or just talk to the NIC
drivers embedded in Linux?

Best Regards,

Huiqun Zhou


----- Original Message ----- 
From: <kameyama ＠ pccluster.org>
To: "Huiqun Zhou" <hqzhou ＠ nju.edu.cn>
Cc: <SCore-users ＠ pccluster.org>
Sent: Thursday, June 19, 2003 1:06 PM
Subject: Re: [SCore-users] Help: using D-Link Gigabit NIC DGE-550T


> In article <005a01c33611$53c1ce60$1a00a8c0 ＠ goofy> "Huiqun Zhou" <hqzhou ＠ nju.edu.cn> wrotes:
> > Does anyone have experience of using D-Link's Gigabit NIC DGE-550T with RedHat
> > 7.3 and Score 5.4.0?
> > I've grabed source code of the driver from D-Link's website, but I got errors
> > when I try to insmod the compiled module into the kernel. 
> 
> If yoy want to compile kernel modules for compute hosts, you must install
> kernel source and compile the modules with that kernel source.
> Please install kernel-source-score-2.4.19-1SCORE.i386.rpm
> (This file located /mnt/cdrom/score.rpm in SCore 5.4.0 CD-ROM or
> following URL.
>     http://www.pccluster.org//score/dist/pub/score-5.4.0/rpm.redhat7.3.i386
> )
> on server host.
> When you install that rpm, kernel source is installed in
> /usr/src/linux-2.4.19score.
> So you can compile that module with following option for compute hosts:
>    -I/usr/src/linux-2.4.19score/include
> 
>                        from Kameyama Toyohisa
> 
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Thu Jun 19 17:03:00 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Thu, 19 Jun 2003 17:03:00 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Help: using D-Link Gigabit NIC DGE-550T
In-Reply-To: Your message of "Thu, 19 Jun 2003 15:44:29 JST."
             <00a601c33636$9dea7cb0$1a00a8c0@goofy>
Message-ID: <20030619080024.B6631128944@neal.il.is.s.u-tokyo.ac.jp>

In article <00a601c33636$9dea7cb0$1a00a8c0 ＠ goofy> "Huiqun Zhou" <hqzhou ＠ nju.edu.cn> wrotes:
> Just want to clarify one thing further. Do you mean that if I want to modify or 
> add a new NIC driver
> into Linux kernel, I must compile it together with the relevant Score source cod
> e as you indicated?

Yes.
Linux kernel module binary includes kernel vereion number
(For example SCopre 5.4.0  2.4.19-1SCORE or 2.4.19-1SCOREsmp).
So you must compile kernel modules with kernel include files what you want.

> Put question in another way, does Score have its own changed NIC drivers, or just
> talk to the NIC drivers embedded in Linux?

Basically, SCore use original NIC driver.

But we change some NIC drivers because PM/Ethernet is not worked
or have performance problem with these drivers.

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From hqzhou ＠ nju.edu.cn  Thu Jun 19 17:39:14 2003
From: hqzhou ＠ nju.edu.cn (Huiqun Zhou)
Date: Thu, 19 Jun 2003 16:39:14 +0800
Subject: [SCore-users-jp] Re: [SCore-users] Help: using D-Link Gigabit NIC DGE-550T
References: <20030619080024.B6631128944@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <00c701c3363e$43683180$1a00a8c0@goofy>

Thanks again.

Two quick questions:
1. My current installation of RedHat is with a kernel version of 2.4.18-3 (original distribution).
  Should I upgrade my kernel to 2.4.19-1 first for coinciding with those from Score?
2. Where is the kernel source after I install kernel-score-2.4.19-1SCORE.i386.rpm? I can't
  find them under /usr/src.

Regards,

Huiqun Zhou


----- Original Message ----- 
From: <kameyama ＠ pccluster.org>
To: "Huiqun Zhou" <hqzhou ＠ nju.edu.cn>
Cc: <kameyama ＠ pccluster.org>; <SCore-users ＠ pccluster.org>
Sent: Thursday, June 19, 2003 4:03 PM
Subject: Re: [SCore-users] Help: using D-Link Gigabit NIC DGE-550T


> In article <00a601c33636$9dea7cb0$1a00a8c0 ＠ goofy> "Huiqun Zhou" <hqzhou ＠ nju.edu.cn> wrotes:
> > Just want to clarify one thing further. Do you mean that if I want to modify or 
> > add a new NIC driver
> > into Linux kernel, I must compile it together with the relevant Score source cod
> > e as you indicated?
> 
> Yes.
> Linux kernel module binary includes kernel vereion number
> (For example SCopre 5.4.0  2.4.19-1SCORE or 2.4.19-1SCOREsmp).
> So you must compile kernel modules with kernel include files what you want.
> 
> > Put question in another way, does Score have its own changed NIC drivers, or just
> > talk to the NIC drivers embedded in Linux?
> 
> Basically, SCore use original NIC driver.
> 
> But we change some NIC drivers because PM/Ethernet is not worked
> or have performance problem with these drivers.
> 
>                        from Kameyama Toyohisa
> 
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Thu Jun 19 17:41:30 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Thu, 19 Jun 2003 17:41:30 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Help: using D-Link Gigabit NIC DGE-550T
In-Reply-To: Your message of "Thu, 19 Jun 2003 16:39:14 JST."
             <00c701c3363e$43683180$1a00a8c0@goofy>
Message-ID: <20030619083854.C5E32128944@neal.il.is.s.u-tokyo.ac.jp>

In article <00c701c3363e$43683180$1a00a8c0 ＠ goofy> "Huiqun Zhou" <hqzhou ＠ nju.edu.cn> wrotes:
> 1. My current installation of RedHat is with a kernel version of 2.4.18-3
> (original distribution).
>  Should I upgrade my kernel to 2.4.19-1 first for coinciding with those
> from Score?

You don't need upgrade server's kernel.

> 2. Where is the kernel source after I install kernel-score-2.4.19-1SCORE.i386.rpm? I can't find them under /usr/src.

Please execute following command:
   # rpm -Uvh kernel-score-2.4.19-1SCORE.i386.rpm
The kernel source will be install under /usr/src/linux-2.4.19score

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From hqzhou ＠ nju.edu.cn  Thu Jun 19 17:53:59 2003
From: hqzhou ＠ nju.edu.cn (Huiqun Zhou)
Date: Thu, 19 Jun 2003 16:53:59 +0800
Subject: [SCore-users-jp] Re: [SCore-users] Help: using D-Link Gigabit NIC DGE-550T
References: <20030619080024.B6631128944@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <00ec01c33640$531cd070$1a00a8c0@goofy>

Sorry I got a wrong file so I failed to locate the kernel source folder.

Huiqun Zhou

----- Original Message ----- 
From: <kameyama ＠ pccluster.org>
To: "Huiqun Zhou" <hqzhou ＠ nju.edu.cn>
Cc: <kameyama ＠ pccluster.org>; <SCore-users ＠ pccluster.org>
Sent: Thursday, June 19, 2003 4:03 PM
Subject: Re: [SCore-users] Help: using D-Link Gigabit NIC DGE-550T


> In article <00a601c33636$9dea7cb0$1a00a8c0 ＠ goofy> "Huiqun Zhou" <hqzhou ＠ nju.edu.cn> wrotes:
> > Just want to clarify one thing further. Do you mean that if I want to modify or 
> > add a new NIC driver
> > into Linux kernel, I must compile it together with the relevant Score source cod
> > e as you indicated?
> 
> Yes.
> Linux kernel module binary includes kernel vereion number
> (For example SCopre 5.4.0  2.4.19-1SCORE or 2.4.19-1SCOREsmp).
> So you must compile kernel modules with kernel include files what you want.
> 
> > Put question in another way, does Score have its own changed NIC drivers, or just
> > talk to the NIC drivers embedded in Linux?
> 
> Basically, SCore use original NIC driver.
> 
> But we change some NIC drivers because PM/Ethernet is not worked
> or have performance problem with these drivers.
> 
>                        from Kameyama Toyohisa
> 
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From hqzhou ＠ nju.edu.cn  Thu Jun 19 18:21:52 2003
From: hqzhou ＠ nju.edu.cn (Huiqun Zhou)
Date: Thu, 19 Jun 2003 17:21:52 +0800
Subject: [SCore-users-jp] Re: [SCore-users] Help: using D-Link Gigabit NIC DGE-550T
References: <20030619083854.C5E32128944@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <00fa01c33644$384a1100$1a00a8c0@goofy>

Sorry to bother you again. This time I got a error message "No boot configuration file" 
when I run eit and after I created a cluster group I got an error of locating files under
opt/score/setup/.... What's wrong?

Huiqun Zhou

----- Original Message ----- 
From: <kameyama ＠ pccluster.org>
To: "Huiqun Zhou" <hqzhou ＠ nju.edu.cn>
Cc: <kameyama ＠ pccluster.org>; <SCore-users ＠ pccluster.org>
Sent: Thursday, June 19, 2003 4:41 PM
Subject: Re: [SCore-users] Help: using D-Link Gigabit NIC DGE-550T


> In article <00c701c3363e$43683180$1a00a8c0 ＠ goofy> "Huiqun Zhou" <hqzhou ＠ nju.edu.cn> wrotes:
> > 1. My current installation of RedHat is with a kernel version of 2.4.18-3
> > (original distribution).
> >  Should I upgrade my kernel to 2.4.19-1 first for coinciding with those
> > from Score?
> 
> You don't need upgrade server's kernel.
> 
> > 2. Where is the kernel source after I install kernel-score-2.4.19-1SCORE.i386.rpm? I can't find them under /usr/src.
> 
> Please execute following command:
>    # rpm -Uvh kernel-score-2.4.19-1SCORE.i386.rpm
> The kernel source will be install under /usr/src/linux-2.4.19score
> 
>                        from Kameyama Toyohisa
> _______________________________________________
> SCore-users mailing list
> SCore-users ＠ pccluster.org
> http://www.pccluster.org/mailman/listinfo/score-users
> 
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Thu Jun 19 20:02:00 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Thu, 19 Jun 2003 20:02:00 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Help: using D-Link Gigabit NIC DGE-550T
In-Reply-To: Your message of "Thu, 19 Jun 2003 17:21:52 JST."
             <00fa01c33644$384a1100$1a00a8c0@goofy>
Message-ID: <20030619105924.C0259128944@neal.il.is.s.u-tokyo.ac.jp>

In article <00fa01c33644$384a1100$1a00a8c0 ＠ goofy> "Huiqun Zhou" <hqzhou ＠ nju.edu.cn> wrotes:
> Sorry to bother you again. This time I got a error message "No boot configuration file" 
> when I run eit and after I created a cluster group I got an error of locating files under
> opt/score/setup/.... What's wrong?

The image files located aqs following:
    /opt/score/ndboot/images/

Note that If you want to install with DGE-550T NIC,
you must add kernel modules in the floppy disk.
And you must recompile NIC driver module again.

because the kernel in the floppy disk is different version (2.4.19-1SCOREBOOT).
I recommended you install EIT with other NIC.
And if you want to use DGE-550T with PPM/Ethernet,
please load the driver and rewrite PM/ethernet configuration file
with mkpmethernetconf.

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From smramesh3 ＠ yahoo.com  Fri Jun 20 05:51:59 2003
From: smramesh3 ＠ yahoo.com (M.Ramesh)
Date: Thu, 19 Jun 2003 13:51:59 -0700 (PDT)
Subject: [SCore-users-jp] [SCore-users] RE: Success in : Compute hosts boot up In SCore , but...  !!! -Ramesh !
Message-ID: <20030619205159.73254.qmail@web20102.mail.yahoo.com>

hi Kameyama,...

Now i did installed Score cluster system successfully with ur
guidance...!!! i would like to thank you very much....for the continuous
help...! 

1) Now, i added server also as a computing host, and then i restarted
server host, then it shows, during start-up messages,  like, 
...
...
...
Binding to the NIS Domain:                    [OK]
Listening for a NIS Domain Serve .......... [FAILED]
...
...

Why is that so ? wat shud i do now to get the NIS Server up successfully
on server host during boot-up ?

2)  And also every time i reboot my compute hosts, the same problem comes
up. 
If i type 
# ypwhich 
on compute host after reboot, it says, 
# ypwhich: Cant communicate with ypbind.
Then i typed 
# ypbind
then i checked ypwhich, it says my server hostname correctly. Why is it
not happening automatically during boot up. What shud i do for this?

3) And then in server host i stoped the iptables to make it default policy
... :

#  /etc/rc.d/init.d/iptables stop

Then in compute host, i had to mount /home of server hosts by using the
command in compute host as follows:
# mount cosmos.lamar.edu:/home /home
Then in compute hosts, its mounted successfully...!!!

wat shud i do for making it permenently mounted every time i reboot the
compute host ?

4) Is there anything i shud take care before start working on cluster for
my programming ? 

I am doing all-cluster testing now. once again thanks a lot Mr.Kameyama.
im really proud having ur help on time. :-)

Thanks and regards from
Ramesh - Lamar University, Beaumont, Texas -USA
***************************************************************************

__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From hqzhou ＠ nju.edu.cn  Fri Jun 20 12:05:07 2003
From: hqzhou ＠ nju.edu.cn (Huiqun Zhou)
Date: Fri, 20 Jun 2003 11:05:07 +0800
Subject: [SCore-users-jp] Re: [SCore-users] Help: using D-Link Gigabit NIC DGE-550T
References: <20030619105924.C0259128944@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <004801c336d8$c0cfb720$1a00a8c0@goofy>

Kameyama-san,

What's the reason of the error message "No boot configuration files" when I run EIT?
Right now, I have finished installing server node, and is about to create the floppy for 
installing compute nodes. 

Regards,

Huiqun Zhou

----- Original Message ----- 
From: <kameyama ＠ pccluster.org>
To: "Huiqun Zhou" <hqzhou ＠ nju.edu.cn>
Cc: <kameyama ＠ pccluster.org>; <SCore-users ＠ pccluster.org>
Sent: Thursday, June 19, 2003 7:02 PM
Subject: Re: [SCore-users] Help: using D-Link Gigabit NIC DGE-550T


> In article <00fa01c33644$384a1100$1a00a8c0 ＠ goofy> "Huiqun Zhou" <hqzhou ＠ nju.edu.cn> wrotes:
> > Sorry to bother you again. This time I got a error message "No boot configuration file" 
> > when I run eit and after I created a cluster group I got an error of locating files under
> > opt/score/setup/.... What's wrong?
> 
> The image files located aqs following:
>     /opt/score/ndboot/images/
> 
> Note that If you want to install with DGE-550T NIC,
> you must add kernel modules in the floppy disk.
> And you must recompile NIC driver module again.
> 
> because the kernel in the floppy disk is different version (2.4.19-1SCOREBOOT).
> I recommended you install EIT with other NIC.
> And if you want to use DGE-550T with PPM/Ethernet,
> please load the driver and rewrite PM/ethernet configuration file
> with mkpmethernetconf.
> 
>                        from Kameyama Toyohisa
> 
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From smramesh3 ＠ yahoo.com  Fri Jun 20 16:17:46 2003
From: smramesh3 ＠ yahoo.com (M.Ramesh)
Date: Fri, 20 Jun 2003 00:17:46 -0700 (PDT)
Subject: [SCore-users-jp] Re: AW: [SCore-users] RE: Success in : Compute hosts boot up In SCore , but...  !!! -Ramesh !
In-Reply-To: <c=DE%a=_%p=Delta_GmbH%l=ABAKUS-030619213648Z-103@abakus.delnet>
Message-ID: <20030620071746.48769.qmail@web20109.mail.yahoo.com>

its not working still, sir!

and moreover, when i tried to test the cluster, it gives the message like
# scorehosts -l -g om

SCBD: Retrying to connect to scoreboard server cosmos.lamar.edu:9950) ...
retrying...
retrying...
retrying...
Unable to connect to scoreboard server cosmos.lamar.edu:9950) error no=111
)...

pls suyggest me somehting...!
regards from Ramesh ...!







--- Martin Newiger <M.Newiger ＠ deltacomputer.de> wrote:
> hi,
> 
> try the command /opt/score/bin/configNIS server
> <whateveryourNIS-Domainis>
> 
> Regards
> M.Newiger
> 
> >-----Urspr?ngliche Nachricht-----
> >Von:	M.Ramesh [SMTP:smramesh3 ＠ yahoo.com]
> >Gesendet am:	Donnerstag, 19. Juni 2003 22:52
> >An:	kame yama; score users
> >Betreff:	[SCore-users] RE: Success in : Compute hosts boot up In SCore
> >, but...  !!! -Ramesh !
> >
> >hi Kameyama,...
> >
> >Now i did installed Score cluster system successfully with ur
> >guidance...!!! i would like to thank you very much....for the
> >continuous
> >help...! 
> >
> >1) Now, i added server also as a computing host, and then i restarted
> >server host, then it shows, during start-up messages,  like, 
> >...
> >...
> >...
> >Binding to the NIS Domain:                    [OK]
> >Listening for a NIS Domain Serve .......... [FAILED]
> >...
> >...
> >
> >Why is that so ? wat shud i do now to get the NIS Server up
> >successfully
> >on server host during boot-up ?
> >
> >2)  And also every time i reboot my compute hosts, the same problem
> >comes
> >up. 
> >If i type 
> ># ypwhich 
> >on compute host after reboot, it says, 
> ># ypwhich: Cant communicate with ypbind.
> >Then i typed 
> ># ypbind
> >then i checked ypwhich, it says my server hostname correctly. Why is it
> >not happening automatically during boot up. What shud i do for this?
> >
> >3) And then in server host i stoped the iptables to make it default
> >policy
> >... :
> >
> >#  /etc/rc.d/init.d/iptables stop
> >
> >Then in compute host, i had to mount /home of server hosts by using the
> >command in compute host as follows:
> ># mount cosmos.lamar.edu:/home /home
> >Then in compute hosts, its mounted successfully...!!!
> >
> >wat shud i do for making it permenently mounted every time i reboot the
> >compute host ?
> >
> >4) Is there anything i shud take care before start working on cluster
> >for
> >my programming ? 
> >
> >I am doing all-cluster testing now. once again thanks a lot
> >Mr.Kameyama.
> >im really proud having ur help on time. :-)
> >
> >Thanks and regards from
> >Ramesh - Lamar University, Beaumont, Texas -USA
>
>************************************************************************
> >***
> >
> >__________________________________
> >Do you Yahoo!?
> >SBC Yahoo! DSL - Now only $29.95 per month!
> >http://sbc.yahoo.com
> >_______________________________________________
> >SCore-users mailing list
> >SCore-users ＠ pccluster.org
> >http://www.pccluster.org/mailman/listinfo/score-users


__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From smramesh3 ＠ yahoo.com  Sun Jun 22 17:04:37 2003
From: smramesh3 ＠ yahoo.com (M.Ramesh)
Date: Sun, 22 Jun 2003 01:04:37 -0700 (PDT)
Subject: [SCore-users-jp] [SCore-users] Problem with SCASH OMP program - efficiency ...!!!
In-Reply-To: <20030619020302.73D83128944@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <20030622080437.49886.qmail@web20109.mail.yahoo.com>

hi...

(/opt/omni/lib/openmp/examples/scash-test/laplace-f77)

Here when i run the lapalace-f77 omp program on 2 dual processor linux
systems , i get the result as follows :


*********************************************
[root ＠ cosmos home]# scrun -nodes=1x2 ./lapsm
SCore-D 5.4.0 connected.
<0:0> SCORE: 2 nodes (1x2) ready.
 sum=  52.1801176
 time=  10.432076
 end..
[root ＠ cosmos home]# scrun -nodes=2x2 ./lapsm
SCore-D 5.4.0 connected.
<0:0> SCORE: 4 nodes (2x2) ready.
 sum=  52.1801176
 time=  10.829392
 end..
[root ＠ cosmos home]# 
*********************************************

And every time it varies...!!! why is that ? 

Is there any reason y im getting more time if i use more nodes...???

Is there any way to make it as faster as usually expected result ? 
pls reply me ...!!!
regards from 
Ramesh --



__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From ishikawa ＠ is.s.u-tokyo.ac.jp  Mon Jun 23 15:03:53 2003
From: ishikawa ＠ is.s.u-tokyo.ac.jp (Yutaka Ishikawa)
Date: Mon, 23 Jun 2003 15:03:53 +0900 (JST)
Subject: [SCore-users-jp] [SCore-users] PC cluster servers are now up
Message-ID: <20030623.150353.294709665.ishikawa@is.s.u-tokyo.ac.jp>

Dear users,

The PC Cluster Consortium server for WWW and Mail services has been
stopped because it was suspected that the server was intruded and was
a DDOS server.  Due to this suspicion, the server was isolated.

We have been checking what happened and concluded that a user's down
load stream, using TCP 12754 port, was suspected, but it was a normal
stream and our server has not been intruded.

The server is now restarted.  If you have sent emails to PC cluster
consortium mailing lists, and they were bounced, could you please resend
them ?

Sincerely,
	Yutaka Ishikawa, Chairman of PC Cluster Consortium
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Mon Jun 23 15:22:47 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Mon, 23 Jun 2003 15:22:47 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Help: using D-Link Gigabit NIC DGE-550T
In-Reply-To: Your message of "Fri, 20 Jun 2003 11:05:07 JST."
             <004801c336d8$c0cfb720$1a00a8c0@goofy>
Message-ID: <20030623061959.AB2AC128944@neal.il.is.s.u-tokyo.ac.jp>

In article <004801c336d8$c0cfb720$1a00a8c0 ＠ goofy> "Huiqun Zhou" <hqzhou ＠ nju.edu.cn> wrotes:
> What's the reason of the error message "No boot configuration files" when I run EIT?

The *.lst files is not found in /opt/score/ndboot/images/ directory.
It is list of network modules in *.img floppy disks.

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From terasawa ＠ soldec-solution.jp  Mon Jun 23 18:17:04 2003
From: terasawa ＠ soldec-solution.jp (Terasawa Hideki)
Date: Mon, 23 Jun 2003 18:17:04 +0900
Subject: [SCore-users-jp] ８７５Ｐチップセットへの対応について
Message-ID: <200306230937.h5N9bWAd024776@soldec-solution.jp>

皆様

寺沢＠ソルデックと申します。

現在、875Pチップセットを使ったＰ４のシステムを作りたいと考えています
が、現在のSCore5.4.0では残念ながら対応していないと聞いています。
私共でも出来そうな対応方法があるのでしょうか？

また、今後の予定としてはどう考えておけば良いか、判る範囲で教えていただ
ければ幸です。

よろしくお願いします。
以上

-- 
Terasawa Hideki mailto:terasawa ＠ soldec-solution.jp


From muna ＠ bioa.tokyo.jst.go.jp  Tue Jun 24 16:07:24 2003
From: muna ＠ bioa.tokyo.jst.go.jp (Munakata Yoshihisa)
Date: Tue, 24 Jun 2003 16:07:24 +0900
Subject: [SCore-users-jp] SCoreとGM
Message-ID: <001f01c33a1f$43b776c0$690b11ac@brdpc005>

宗像と申します。

お世話になります。

どなたか経験があればお教えいただきたいと思い、メールします。

現状、SCore5.0.1+Myrinet2000で、PCクラスタ（256台構成）を運用して
いるのですが、一部ノードを、MyricomのGMの運用に切り替えて確認を
したいと考えています。

このとき、Myrinet2000は、全二重で結合する為、128ポートのスイッチを
6台フル結合した例の構成になっているのですが、上で言っているような、
PMとGMの混在環境（混在といってももちろん、相互に通信させようなど
とは思っていません。別々に運用します。）にした時、


（１）まず、そのような混在をさせてMyrinet周りで、SCoreの動作におかしな
    影響を与えないか。

（２）混在可能の時、その設定としては、SCore側では

   管理ホスト上で、
   ・scorehosts.db、pm-myrinet.confで、SCore運用から抜くホストに対応
   する記述行を、コメントアウト。
   ・scoreboard、msgbservを再起動

   ・その後、抜いた計算ホスト側でscoreの自動起動を抑止して、再起動。

すれば、SCore自体は抜いた計算ホストには感知しなくなると思うのです
が、それだけでよいか。

（３）GMのMapper設定としては、大枠として、GMのマップ状態を静的
    に記述してしまうstatic mappingと、mapperを動かして動的にマップ
    生成を行うactive mappingがあると思っていますが、どちらでやった
    としても、SCore側に悪影響を与えないか。

という点について、どなたかコメントいただけないでしょうか。

SCore部分とGM部分をソフトウェア的に切り離してしまっていれば、
基本的には大丈夫ではないかとは思うのですが、Myrinetスイッチが
結線としては、全台またいでつながっており、認識していない範囲で
変な動作を引き起こさないか不安な為、SCoreに詳しい方か、どなた
か経験があればお話を伺えればと思い、メールさせて頂きました。

申し訳ありませんが、宜しくお願いします。









From smramesh3 ＠ yahoo.com  Wed Jun 25 17:31:38 2003
From: smramesh3 ＠ yahoo.com (M.Ramesh)
Date: Wed, 25 Jun 2003 01:31:38 -0700 (PDT)
Subject: [SCore-users-jp] [SCore-users] Error 139 in Fortran77 program when  i use omf77 -omniconfig=scash ...?
Message-ID: <20030625083138.47029.qmail@web20110.mail.yahoo.com>

Dear sir,
when i tried to run a programs (group of fortran programs ) using 

***************************************************************
# omf77 -omniconfig=scash -o output program1.f program2.f program3.f
Compiling 'program1.f'...
   pidate:
   prob:
Error 139
# 
***************************************************************

Please tell me what does it mean this error ? i cudnt get out of this
problem.
I have attached the source code of the program1.f with this mail. 

But i could run the same programs in PGI compilers without any
problem...using this command.

#pgf77 -mp -o output program1.f program2.f

Please reply me Sir...!!!

Ramesh.M 
Lamar University,
Texas, USA.

__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com
-------------- next part --------------
テキスト形式以外の添付ファイルを保管しました...
ファイル名: program1.f
型:         application/octet-stream
サイズ:     6611 バイト
説明:       program1.f
URL:        <http://new1.pccluster.org/pipermail/score-users-jp/attachments/20030625/1612383f/attachment.obj>

From NAIfle2FLE2 ＠ fle.fujitsu.com  Wed Jun 25 17:30:47 2003
From: NAIfle2FLE2 ＠ fle.fujitsu.com (=?iso-2022-jp?b?TkFJZmxlMkZMRTIgGyRCIXcbKEIgZmxlLmZ1aml0c3UuY29t?=)
Date: Wed, 25 Jun 2003 09:30:47 +0100
Subject: [SCore-users-jp] [SCore-users] ALERT -  GroupShield ticket number OA24_1056529844_FLE2_3 was gen
 erated
Message-ID: <E9978DD405A2D611913500047583E37A4C478E@fle2.fle.fujitsu.com>

Action Taken:
The attachment was quarantined from the message and replaced with a text
file informing the recipient of the action taken.

To:
pdp-openmp ＠ trc.rwcp.or.jp <pdp-openmp ＠ trc.rwcp.or.jp>; kame yama
<kameyama ＠ pccluster.org>; score users <score-users ＠ pccluster.org>

From:
M.Ramesh <smramesh3 ＠ yahoo.com>

Sent:
837062912,29571828

Subject:
[SCore-users] Error 139 in Fortran77 program when  i use omf77
-omniconfig=scash ...?

Attachment Details:-

Attachment Name: program1.f
File: program1.f
Infected? No
Repaired? No
Blocked? Yes
Deleted? No
Virus Name: 



	

-------------- next part --------------
テキスト形式以外の添付ファイルを保管しました...
ファイル名: 無し
型:         application/ms-tnef
サイズ:     1921 バイト
説明:       無し
URL:        <http://new1.pccluster.org/pipermail/score-users-jp/attachments/20030625/579ff29b/attachment.bin>

-------------- next part --------------
文字コード指定の無い添付文書を保管しました...
名前: InterScan_Disclaimer.txt
URL:  <http://new1.pccluster.org/pipermail/score-users-jp/attachments/20030625/579ff29b/attachment.txt>

From ulrich.oelmann ＠ iws.uni-stuttgart.de  Wed Jun 25 19:05:38 2003
From: ulrich.oelmann ＠ iws.uni-stuttgart.de (Ulrich Oelmann)
Date: Wed, 25 Jun 2003 12:05:38 +0200
Subject: [SCore-users-jp] [SCore-users] Question: Checkpointing and Sequential Programs
Message-ID: <20030625120537.Y23276@wum.bauingenieure.uni-stuttgart.de>

Dear list!

If I start a sequential program on our SCore_5.0 Cluster using SCores
system(6) command and try to generate a checkpoint later, the program
catches the sent SIGQUIT signal and immediately terminates instead of
producing a checkpoint.

As the program does not violate any of the limitations that are
necessary for checkpointing, I suspect the way system(6) works to make
checkpointing impossible.

Does anybody know a way around this problem? Can I still use Scores
system(6) command or do I have to change my sequential program into a
pseudo parallel program by artificially turning it into a MPI program?

Best regards
Ulrich Oelmann
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From hqzhou ＠ nju.edu.cn  Wed Jun 25 19:06:17 2003
From: hqzhou ＠ nju.edu.cn (Huiqun Zhou)
Date: Wed, 25 Jun 2003 18:06:17 +0800
Subject: [SCore-users-jp] [SCore-users] Help: Trouble in installing compute node
Message-ID: <002201c33b01$6affa210$1a00a8c0@goofy>

Kameyama-san and score users,

After tracing back the archive of the mailing list, we at last created the floppy
following the instructions for adding e1000 NIC driver into the boot image in 
the mail posted by Kameyama-san in last October. Currently, we are using D-Link
550T Gigabit NIC on all of our compute nodes. The driver for the card has been 
compiled and linked with Score kernel source.

The problemn is when we tried to use the floppy to boot the compute nodes, we got
following error messages:
    ...
    No DHCP Server specified. Use broadcast.
    SIOCSIFADDR No such device
    ...

I have noticed the posts from Mr. Ramesh, it seems the problem we are facing is a 
little bit similar to the one Ramesh addressed. I have checked the status of my server 
according to Kameyama-san's suggestions in the mail trails, and it seems no problem. 
What's wrong ?

Thanks in advance.

Huiqun Zhou  
-------------- next part --------------
HTMLの添付ファイルを保管しました...
URL: <http://new1.pccluster.org/pipermail/score-users-jp/attachments/20030625/cd8408fe/attachment.html>

From kameyama ＠ pccluster.org  Wed Jun 25 19:29:49 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Wed, 25 Jun 2003 19:29:49 +0900
Subject: [SCore-users-jp] [SCore-users] Re: Help: Trouble in installing compute node
In-Reply-To: Your message of "Wed, 25 Jun 2003 18:06:17 JST."
             <002201c33b01$6affa210$1a00a8c0@goofy>
Message-ID: <20030625102654.A3E06128944@neal.il.is.s.u-tokyo.ac.jp>

In article <002201c33b01$6affa210$1a00a8c0 ＠ goofy> "Huiqun Zhou" <hqzhou ＠ nju.edu.cn> wrotes:
> After tracing back the archive of the mailing list, we at last created the floppy
> following the instructions for adding e1000 NIC driver into the boot image in 
> the mail posted by Kameyama-san in last October. Currently, we are using D-Link
> 550T Gigabit NIC on all of our compute nodes. The driver for the card has been 
> compiled and linked with Score kernel source.

Probably, you cannot load or fail to load D-Link 550T driver
on the compute hosts.
Please look at console3 (ALT-shift-F3, installer log)
and console 4 (ALT-shift-F4, kernel log).

If D-Link 550T driver is not try to loaded (in consoile 3),
you must add the driver's information into module-info and pcitable.
(Note that redhat 7.x dose not support D-Link 550T driver.
So you must wriite driver's information byy yourself.)

If D-Link 550T driver loadded is failed, you may compile
the driver for the boot floppy.

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From hqzhou ＠ nju.edu.cn  Wed Jun 25 19:59:01 2003
From: hqzhou ＠ nju.edu.cn (Huiqun Zhou)
Date: Wed, 25 Jun 2003 18:59:01 +0800
Subject: [SCore-users-jp] [SCore-users] Re: Help: Trouble in installing compute node
References: <20030625102654.A3E06128944@neal.il.is.s.u-tokyo.ac.jp>
Message-ID: <000901c33b08$c95d54e0$1a00a8c0@goofy>

I have created module-info, as for pcitable, I don't know where I can get the
VendorID and DeviceID for the NIC. I don't know if I can get any useful information
from D-Link's tech support.

Thanks!

Huiqun Zhou


----- Original Message ----- 
From: <kameyama ＠ pccluster.org>
To: "Huiqun Zhou" <hqzhou ＠ nju.edu.cn>
Cc: <kameyama ＠ pccluster.org>; <SCore-users ＠ pccluster.org>
Sent: Wednesday, June 25, 2003 6:29 PM
Subject: Re: Help: Trouble in installing compute node


> In article <002201c33b01$6affa210$1a00a8c0 ＠ goofy> "Huiqun Zhou" <hqzhou ＠ nju.edu.cn> wrotes:
> > After tracing back the archive of the mailing list, we at last created the floppy
> > following the instructions for adding e1000 NIC driver into the boot image in 
> > the mail posted by Kameyama-san in last October. Currently, we are using D-Link
> > 550T Gigabit NIC on all of our compute nodes. The driver for the card has been 
> > compiled and linked with Score kernel source.
> 
> Probably, you cannot load or fail to load D-Link 550T driver
> on the compute hosts.
> Please look at console3 (ALT-shift-F3, installer log)
> and console 4 (ALT-shift-F4, kernel log).
> 
> If D-Link 550T driver is not try to loaded (in consoile 3),
> you must add the driver's information into module-info and pcitable.
> (Note that redhat 7.x dose not support D-Link 550T driver.
> So you must wriite driver's information byy yourself.)
> 
> If D-Link 550T driver loadded is failed, you may compile
> the driver for the boot floppy.
> 
>                        from Kameyama Toyohisa
> 
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From rene.storm ＠ emplics.com  Wed Jun 25 20:23:34 2003
From: rene.storm ＠ emplics.com (Rene Storm)
Date: Wed, 25 Jun 2003 13:23:34 +0200
Subject: [SCore-users-jp] AW: [SCore-users] Re: Help: Trouble in installing compute node
Message-ID: <29B376A04977B944A3D87D22C495FB23012758@vertrieb.emplics.com>

Hi Huiqun,

do a 

#lspci -n  

on a running (linux) system with your card inside.
You will see all VendorIDs and DeviceIDs.
If your Vendor isn't listed in /usr/src/linux/drivers/pci/pci.ids
you can see the numbers with 

#cat /proc/pci

D-Links Vendor ID should be 1186

Good Luck

Rene Storm
emplics AG



-----Ursprüngliche Nachricht-----
Von: Huiqun Zhou [mailto:hqzhou ＠ nju.edu.cn] 
Gesendet: Mittwoch, 25. Juni 2003 12:59
An: kameyama ＠ pccluster.org
Cc: kameyama ＠ pccluster.org; SCore-users ＠ pccluster.org
Betreff: [SCore-users] Re: Help: Trouble in installing compute node


I have created module-info, as for pcitable, I don't know where I can get the VendorID and DeviceID for the NIC. I don't know if I can get any useful information from D-Link's tech support.

Thanks!

Huiqun Zhou


----- Original Message ----- 
From: <kameyama ＠ pccluster.org>
To: "Huiqun Zhou" <hqzhou ＠ nju.edu.cn>
Cc: <kameyama ＠ pccluster.org>; <SCore-users ＠ pccluster.org>
Sent: Wednesday, June 25, 2003 6:29 PM
Subject: Re: Help: Trouble in installing compute node


> In article <002201c33b01$6affa210$1a00a8c0 ＠ goofy> "Huiqun Zhou" 
> <hqzhou ＠ nju.edu.cn> wrotes:
> > After tracing back the archive of the mailing list, we at last 
> > created the floppy following the instructions for adding e1000 NIC 
> > driver into the boot image in the mail posted by Kameyama-san in 
> > last October. Currently, we are using D-Link 550T Gigabit NIC on all 
> > of our compute nodes. The driver for the card has been compiled and 
> > linked with Score kernel source.
> 
> Probably, you cannot load or fail to load D-Link 550T driver on the 
> compute hosts. Please look at console3 (ALT-shift-F3, installer log)
> and console 4 (ALT-shift-F4, kernel log).
> 
> If D-Link 550T driver is not try to loaded (in consoile 3), you must 
> add the driver's information into module-info and pcitable. (Note that 
> redhat 7.x dose not support D-Link 550T driver. So you must wriite 
> driver's information byy yourself.)
> 
> If D-Link 550T driver loadded is failed, you may compile
> the driver for the boot floppy.
> 
>                        from Kameyama Toyohisa
> 
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org http://www.pccluster.org/mailman/listinfo/score-users
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From smramesh3 ＠ yahoo.com  Wed Jun 25 20:44:25 2003
From: smramesh3 ＠ yahoo.com (M.Ramesh)
Date: Wed, 25 Jun 2003 04:44:25 -0700 (PDT)
Subject: [SCore-users-jp] [SCore-users] Problem in SCore-D 5.4.0 :  VM Fault out of DSM area PC=[4b1dd640]:ADDR=[25280817]
Message-ID: <20030625114425.68106.qmail@web20110.mail.yahoo.com>

hi 
when i am trying to run one fortran 77 OMP program code in omni openMP in
scout environment (SCASH SYSTEM)..., im getting error messages, though i 
compiled it successfully using the following command:
# omf77 -omniconfig=scash -o out  program1.f program2.f 



********************************************************
[root ＠ cosmos ompside]# scrun -nodes=2x2 ./out

SCore-D 5.4.0 connected.
<1> SCORE-D:WARNING Unable to change directory (/root/ompside), and set to
/var/scored/singleuser/0/jobs/jid-1.
<0:0> SCORE: 4 nodes (2x2) ready.
<0> VM Fault out of DSM area PC=[4b1dd640]:ADDR=[25280817]
1Test Set for IVP Solvers (release 2.1)

 Solving Transform using PSIDE

[root ＠ cosmos ompside]#

*******************************************************

Please what is this VM Fault out of DSM area... problem ???
I installed openMP without-jvm option too !!! (OpemMP 1.4.) here.

Regards
Ramesh...!


__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From msato ＠ is.tsukuba.ac.jp  Thu Jun 26 14:46:11 2003
From: msato ＠ is.tsukuba.ac.jp (Mitsuhisa Sato)
Date: Thu, 26 Jun 2003 14:46:11 +0900 (JST)
Subject: [SCore-users-jp] [SCore-users] Re: Error 139 in Fortran77 program when i use omf77
 -omniconfig=scash ...?
In-Reply-To: <20030625083138.47029.qmail@web20110.mail.yahoo.com>
References: <20030625083138.47029.qmail@web20110.mail.yahoo.com>
Message-ID: <20030626.144611.424267513.msato@hpcs.is.tsukuba.ac.jp>

Dear Ramesh. M,

We have checked your program with the current release.
We have found a bug in the current release, but it is ok with the
recent version. 

We are going to upload the recent version soon. Then, let you know.

And, I am checking your problem with Omni/SCASH. Give me sometime.

-- M. Sato

From: "M.Ramesh" <smramesh3 ＠ yahoo.com>
Subject: Error 139 in Fortran77 program when i use omf77 -omniconfig=scash ...?
Date: Wed, 25 Jun 2003 01:31:38 -0700 (PDT)

> Dear sir,
> when i tried to run a programs (group of fortran programs ) using 
> 
> ***************************************************************
> # omf77 -omniconfig=scash -o output program1.f program2.f program3.f
> Compiling 'program1.f'...
>    pidate:
>    prob:
> Error 139
> # 
> ***************************************************************
> 
> Please tell me what does it mean this error ? i cudnt get out of this
> problem.
> I have attached the source code of the program1.f with this mail. 
> 
> But i could run the same programs in PGI compilers without any
> problem...using this command.
> 
> #pgf77 -mp -o output program1.f program2.f
> 
> Please reply me Sir...!!!
> 
> Ramesh.M 
> Lamar University,
> Texas, USA.
> 
> __________________________________
> Do you Yahoo!?
> SBC Yahoo! DSL - Now only $29.95 per month!
> http://sbc.yahoo.com
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From lunetix83 ＠ hotmail.com  Fri Jun 27 18:28:58 2003
From: lunetix83 ＠ hotmail.com (Michael N.)
Date: Fri, 27 Jun 2003 09:28:58 +0000
Subject: [SCore-users-jp] [SCore-users] Replacing Compute Node
Message-ID: <BAY1-F117CWO6hm8Ua800009bd9@hotmail.com>

Hi,

I'm running a bigger SCore-Cluster and now I have a node that's defect. How 
can I integrate it in the existing environment?

Regards
Michael

_________________________________________________________________
Fotos? - ?MSN Fotos das virtuelle Fotoalbum. Allen Freunden zeigen oder 
einfach online entwickeln lassen:  
http://photos.msn.de/support/worldwide.aspx

_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users


From kameyama ＠ pccluster.org  Fri Jun 27 20:20:48 2003
From: kameyama ＠ pccluster.org (=?iso-2022-jp?b?a2FtZXlhbWEgGyRCIXcbKEIgcGNjbHVzdGVyLm9yZw==?=)
Date: Fri, 27 Jun 2003 20:20:48 +0900
Subject: [SCore-users-jp] Re: [SCore-users] Replacing Compute Node
In-Reply-To: Your message of "Fri, 27 Jun 2003 09:28:58 JST."
             <BAY1-F117CWO6hm8Ua800009bd9@hotmail.com>
Message-ID: <20030627111747.1BEA4128944@neal.il.is.s.u-tokyo.ac.jp>

In article <BAY1-F117CWO6hm8Ua800009bd9 ＠ hotmail.com> "Michael N." <lunetix83 ＠ hotmail.com> wrotes:
> I'm running a bigger SCore-Cluster and now I have a node that's defect. How 
> can I integrate it in the existing environment?

Please read "Resuming SCore-D from an Unexpected Failure":
   http://www.pccluster.org/score/dist/score/html/en/reference/scored/resume.html
And "Automatic Operation and High Availability of SCore-D"
    http://www.pccluster.org/score/dist/score/html/en/reference/scored/auto.html

                       from Kameyama Toyohisa
_______________________________________________
SCore-users mailing list
SCore-users ＠ pccluster.org
http://www.pccluster.org/mailman/listinfo/score-users