Re: [gridengine users] install_qmaster issue

[Jerome]

Hi Arnau

Le 19/04/2018 à 02:50, Arnau a écrit :
> Hi,
> 
> why don't you first install all dependecies and the sge?
> I don't know what's wrong with your apt, but in mycase apt install -f
> does not want to remove sge.

That the principle of Debian, all deb should install with all
dependencies. That's why i'm a user of debian from more that 20 years ;-)

What i can't realize is that you could install sge in debian 9.4, as the
dependencie od sge is that:

 dpkg -I sge_8.1.9_amd64.deb
 new debian package, version 2.0.
 size 8780400 bytes: control archive=3852 bytes.
1431 bytes,31 lines  control
5751 bytes,86 lines  md5sums
 135 bytes, 7 lines   *  postinst #!/bin/sh
 132 bytes, 7 lines   *  postrm   #!/bin/sh
  17 bytes, 1 lines  shlibs
 Package: sge
 Version: 8.1.9
 Architecture: amd64
 Maintainer: Dave Love 
 Installed-Size: 56370
 Depends: libc6 (>= 2.17), libdb5.3, db5.3-util, libhwloc5 (>= 1.10.0),
libice6 (>= 1:1.0.0), libjemalloc1 (>= 2.1.1), libmunge2 (>= 0.5.8),
libpam0g (>= 0.99.7.1), libsm6, libssl1.0.0 (>= 1.0.0), libtinfo5,
libx11-6, libxext6, libxm4 (>= 2.3.4), libxt6, sge-common, procps, xterm
 Recommends: sge-doc

And the problem is with "libssl1.0.0", There is uniquely libssl1.0.2, as
you have istalled on your own server.
If i force the instalation, i've got a this warning:


dpkg: sge: dependency problems, but configuring anyway as you requested:
 sge depends on libssl1.0.0 (>= 1.0.0); however:
  Package libssl1.0.0 is not installed.

That's normal, the library libssl1.0.0 not existe yet on debian 9.4.

the binaries needs "libssl.so.1.0.0" :

$ ldd /opt/sge/utilbin/lx-amd64/gethostname
linux-vdso.so.1 (0x7ffd94446000)
libmunge.so.2 => /usr/lib/x86_64-linux-gnu/libmunge.so.2
(0x7f72b8093000)
libssl.so.1.0.0 => not found
libcrypto.so.1.0.0 => not found
libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2 (0x7f72b7e8f000)
libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x7f72b7b8b000)
libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0
(0x7f72b796e000)
libjemalloc.so.1 => /usr/lib/x86_64-linux-gnu/libjemalloc.so.1
(0x7f72b7737000)
libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x7f72b7398000)
/lib64/ld-linux-x86-64.so.2 (0x7f72b8511000)

I've change of repository, using the debian oficial one. No matter, the
same error occurs.


> 
> cloud or not cloud should not make any difference.
> I'm attaching the list of installed packages in my system.

Yes, that correct, that's just a joke because i can't install SGE :-(


No problem, thank's for your help. I think i will let it because the sge
will to be unused in a short time, nobody seems to be on charge of it .
Regards!

-- 
-- Jérôme
La crise d'hier est la blague de demain.
(Herbert George Wells)
___
users mailing list
users@gridengine.org
https://gridengine.org/mailman/listinfo/users

Re: [gridengine users] install_qmaster issue

[Jerome]

Le 13/04/2018 à 03:09, Arnau a écrit :
> Hi,
> 
> $ cat /etc/debian_version
> 9.4
> 
> $ dpkg -l|grep  ssl
> ii  libflac8:amd64                   1.3.2-1                       
> amd64        Free Lossless Audio Codec - runtime C library
> ii  libssl-doc                       1.1.0f-3+deb9u2                all 
>         Secure Sockets Layer toolkit - development documentation
> ii  libssl1.0-dev:amd64              1.0.2l-2+deb9u3               
> amd64        Secure Sockets Layer toolkit - development files
> ii  libssl1.0.2:amd64                1.0.2l-2+deb9u3               
> amd64        Secure Sockets Layer toolkit - shared libraries
> ii  libssl1.1:amd64                  1.1.0f-3+deb9u2               
> amd64        Secure Sockets Layer toolkit - shared libraries
> ii  openssl                          1.1.0f-3+deb9u2               
> amd64        Secure Sockets Layer toolkit - cryptographic utility
> ii  python-openssl                   16.2.0-1                       all 
>         Python 2 wrapper around the OpenSSL library
> 
> I have both. I could try with 1.1
> 
> Arnau
> 

Hi Arnau.

Do you upgrade your system with sge first install? Or do you install sge
on a fresh debian 9.4 ? If you do the last, how do you resolve the unmet
conflict?

Regards!



-- 
-- Jérôme
Fuir: prendre son courage a deux pieds.
(Alexandre Breffort)
___
users mailing list
users@gridengine.org
https://gridengine.org/mailman/listinfo/users

Re: [gridengine users] install_qmaster issue

[Jerome]

Le 12/04/2018 à 03:11, Arnau a écrit :
> Hi Jerome,
> 
> the deb packages are in this link:
> https://drive.google.com/open?id=1tjEh1ygOxAPigoDWEy8IXi0o_nkH--NM
> 
> 
> HTH,
> Arnau
> 
>

Dear Arnau

Thank's foe get me this deb file. The problem is that they can't
install, because of change in debian 9.x :

Preparing to unpack sge_8.1.9_amd64.deb ...
Unpacking sge (8.1.9) ...
dpkg: dependency problems prevent configuration of sge:
 sge depends on db5.3-util; however:
  Package db5.3-util is not installed.
 sge depends on libhwloc5 (>= 1.10.0); however:
  Package libhwloc5 is not installed.
 sge depends on libjemalloc1 (>= 2.1.1); however:
  Package libjemalloc1 is not installed.
 sge depends on libmunge2 (>= 0.5.8); however:
  Package libmunge2 is not installed.
 sge depends on libssl1.0.0 (>= 1.0.0); however:
  Package libssl1.0.0 is not installed.
 sge depends on libxm4 (>= 2.3.4); however:
  Package libxm4 is not installed.

dpkg: error processing package sge (--install):
 dependency problems - leaving unconfigured
Errors were encountered while processing:
 sge


All of this library changed, and that what have changed in the 8.9.1_1
version, that i have had to compile, and that don't install well...

This deb file install well on a debian 8.x version. But this version is
old now..

Regards


-- 
-- Jérôme
Une vie sans chocolat est une vie à laquelle manque l'essentiel.
(Marcia Colman et Frederic Morton)
___
users mailing list
users@gridengine.org
https://gridengine.org/mailman/listinfo/users

Re: [gridengine users] install_qmaster issue

[Jerome]

Le 21/03/2018 à 06:39, Arnau a écrit :
> Hi,
> 
> I've tried to reproduce your issue in AWS. I did not build the packages
> but installed from github as you did.
> I have to say that everything work perfectly:
> 
> sh scripts/bootstrap.sh
> ./aimk -no-java -no-jni -only-core
> setenv SGE_ROOT /opt/sge/
> scripts/distinst -local -allall
> chown -R sgeadmin:sgeadmin /opt/sge/
> ./install_qmaster
> 
> I used soem random ports:
> 

Dear Arnau

As I can't configure SoGE on my first machine, i try to install it on
another one, that have a public ip . The firt machine is on a private
network. But the problem is the same, during the ./inst_sge -m:

unable to contact qmaster using port 6444 on host "cluster.ibt.unam.mx"

But the deamon is running on this machine:

nc -vz cluster 20-7000

 cluster: Thu Apr  5 12:53:28 2018

cluster.ibt.unam.mx [132.248.32.144] 6444 (sge-qmaster) open
cluster.ibt.unam.mx [132.248.32.144] 22 (ssh) open


So, as I wrtoe you before, the poblem seesm to be on the ss
comunication. When I install SoGe with the "-csp" option, that's run
well ! So my solution for now is to use CSP. This implies more stuff for
configurong the users of the cluster. But that's what I have to do,
because in debian 9.4 i never could install SoGE. I try it more tha 10
times with different machines, and never happens.
Maybe if you could send me your compile binaries to try here?

Regards


-- 
-- Jérôme
Le baiser est la plus sûre façon de se taire en disant tout.
(Guy de Maupassant)
___
users mailing list
users@gridengine.org
https://gridengine.org/mailman/listinfo/users

Re: [gridengine users] install_qmaster issue

[Jerome]

Le 21/03/2018 à 06:39, Arnau a écrit :
> Hi,
> 
> I've tried to reproduce your issue in AWS. I did not build the packages
> but installed from github as you did.
> I have to say that everything work perfectly:
> 
> sh scripts/bootstrap.sh
> ./aimk -no-java -no-jni -only-core
> setenv SGE_ROOT /opt/sge/
> scripts/distinst -local -allall
> chown -R sgeadmin:sgeadmin /opt/sge/
> ./install_qmaster
> 

Dear Arnau.

It's seems that SoGE don't like me.. I do a fresh install of debian 9.4,
and install all of the package needed to compile sge But i get this
error during the aimk process:

checking for strsignal... yes
checking for special C compiler options needed for large files... no
checking for _FILE_OFFSET_BITS value needed for large files... no
configure: creating ./config.status
config.status: creating sge_config.h
_C_O_R_E__S_Y_S_T_E_M_
../common/Makefile:71: ../common/common_dependencies: No such file or
directory
../security/gss/Makefile:37: ../security/gss/gss_dependencies: No such
file or directory
../libs/uti/Makefile:163: ../libs/uti/uti_dependencies: No such file or
directory
../libs/uti2/Makefile:47: ../libs/uti2/uti2_dependencies: No such file
or directory
../libs/cull/Makefile:97: ../libs/cull/cull_dependencies: No such file
or directory
../libs/comm/lists/Makefile:97:
../libs/comm/lists/commlistslib_dependencies: No such file or directory
../libs/comm/examples/Makefile:70:
../libs/comm/examples/commexamples_dependencies: No such file or directory
../libs/comm/Makefile:143: ../libs/comm/commlib_dependencies: No such
file or directory
../libs/sgeobj/Makefile:166: ../libs/sgeobj/sgeobj_dependencies: No such
file or directory
../libs/gdi/Makefile:83: ../libs/gdi/gdi_dependencies: No such file or
directory
../libs/evc/Makefile:74: ../libs/evc/evc_dependencies: No such file or
directory
../libs/evm/Makefile:75: ../libs/evm/evm_dependencies: No such file or
directory
../libs/mir/Makefile:87: ../libs/mir/mir_dependencies: No such file or
directory
../libs/sched/Makefile:111: ../libs/sched/sched_dependencies: No such
file or directory
../libs/spool/Makefile:91: ../libs/spool/spool_dependencies: No such
file or directory
../libs/spool/loader/Makefile:75:
../libs/spool/loader/spoolloader_dependencies: No such file or directory
../libs/spool/berkeleydb/Makefile:79:
../libs/spool/berkeleydb/spoolb_dependencies: No such file or directory
../libs/spool/classic/Makefile:73:
../libs/spool/classic/spoolc_old_dependencies: No such file or directory
../libs/spool/dynamic/Makefile:75:
../libs/spool/dynamic/spoold_dependencies: No such file or directory
../libs/spool/flatfile/Makefile:93:
../libs/spool/flatfile/spoolc_dependencies: No such file or directory
../libs/japi/Makefile:117: ../libs/japi/japi_dependencies: No such file
or directory
../libs/juti/Makefile:76: ../libs/juti/juti_dependencies: No such file
or directory
../libs/ijs/Makefile:39: ../libs/ijs/ijs_dependencies: No such file or
directory
../security/coshepherd/Makefile:73:
../security/coshepherd/coshepherd_dependencies: No such file or directory
../daemons/common/Makefile:98: ../daemons/common/common_dependencies: No
such file or directory
../daemons/execd/Makefile:114: ../daemons/execd/execd_dependencies: No
such file or directory
../daemons/qmaster/Makefile:168:
../daemons/qmaster/qmaster_dependencies: No such file or directory
../daemons/shadowd/Makefile:83: ../daemons/shadowd/shadowd_dependencies:
No such file or directory
../daemons/shepherd/Makefile:98:
../daemons/shepherd/shepherd_dependencies: No such file or directory
../clients/common/Makefile:76: ../clients/common/common_dependencies: No
such file or directory
../clients/qacct/Makefile:78: ../clients/qacct/qacct_dependencies: No
such file or directory
../clients/qalter/Makefile:77: ../clients/qalter/qalter_dependencies: No
such file or directory
../clients/qconf/Makefile:87: ../clients/qconf/qconf_dependencies: No
such file or directory
../clients/qdel/Makefile:76: ../clients/qdel/qdel_dependencies: No such
file or directory
../clients/qhost/Makefile:84: ../clients/qhost/qhost_dependencies: No
such file or directory
../clients/qquota/Makefile:87: ../clients/qquota/qquota_dependencies: No
such file or directory
../clients/qmod/Makefile:76: ../clients/qmod/qmod_dependencies: No such
file or directory
../clients/qrsh/Makefile:78: ../clients/qrsh/qrsh_dependencies: No such
file or directory
../clients/qsh/Makefile:79: ../clients/qsh/qsh_dependencies: No such
file or directory
../clients/qstat/Makefile:90: ../clients/qstat/qstat_dependencies: No
such file or directory
../clients/qsub/Makefile:83: ../clients/qsub/qsub_dependencies: No such
file or directory
../clients/qevent/Makefile:77: ../clients/qevent/qevent_dependencies: No
such file or directory
../clients/qrsub/Makefile:82: ../clients/qrsub/qrsub_dependencies: No
such file or directory
../clients/qrdel/Makefile:81: ../clients/qrdel/qrdel_dependencies: No
such file or directory
../clients/qrstat/Makefile:85: ../clients

Re: [gridengine users] install_qmaster issue

[Jerome]

Le 16/03/2018 à 03:41, Arnau a écrit :
> Hi,
> 
> is the name resolution working as expected? is xx.yy.zz = 
> invitado.uuab.ibt.unam.mx  ? is the
> resolution of  invitado.uuab.ibt.unam.mx
>  to 10.0.6.50 ? are you using fqdn or
> short names (ignore_fqdn) ?
> 

Dear Arnau

I've made some try to understand why this issue. But i failed ...

So, i instaled in te same machine a debian 8.10 version, install the
soge paquages v 8.1.9 and run into the qmaster and execd instalation
without any problem.

So i install this machine with debian 9.4. When i try to install the
soge package v 8.1.9, i've got this error:

sge depends on db5.3-util; however:
  Package db5.3-util is not installed.
 sge depends on libhwloc5 (>= 1.10.0); however:
  Package libhwloc5 is not installed.
 sge depends on libjemalloc1 (>= 2.1.1); however:
  Package libjemalloc1 is not installed.
 sge depends on libmunge2 (>= 0.5.8); however:
  Package libmunge2 is not installed.
 sge depends on libssl1.0.0 (>= 1.0.0); however:
  Package libssl1.0.0 is not installed.
 sge depends on libxm4 (>= 2.3.4); however:
  Package libxm4 is not installed.


So i decide to compile the SoGE source from gitlab, version
"9b99ef86b09516976fdad5cbd39af7066766" .

I could obtain the 3 deb packages sge sge-common sge-doc and install then.

The issue occurs at the same point of the configuration:

"Adding admin and submit hosts"

i get back:

error: commlib error: got read error (closing
"invitado.uuab.ibt.unam.mx/qmaster/1")
unable to contact qmaster using port 6444 on host
"invitado.uuab.ibt.unam.mx"

If i decide to not add some admin host, the same arror occurs
during "Creating the default  queue and  hostgroup"

error: commlib error: got read error (closing
"invitado.uuab.ibt.unam.mx/qmaster/1")
unable to contact qmaster using port 6444 on host
"invitado.uuab.ibt.unam.mx"

Command failed: ./bin/lx-amd64/qconf -Ahgrp /tmp/hostqueue635
Probably a permission problem. Please check file access permissions.
Check root read/write permission. Check if SGE daemons are running.

And the program exit.

In other hand, i've check that the service of qmaster is on on the
machine, with "nc":

$ nc -zv invitado.uuab.ibt.unam.mx 20-7000

  invitado:
Tue Mar 20 17:33:19 2018

invitado.uuab.ibt.unam.mx [10.0.6.50] 6444 (sge-qmaster) open
invitado.uuab.ibt.unam.mx [10.0.6.50] 22 (ssh) open

Next i I folow the  grid engine TroubleShooting, and select all of the
posible dl value (1 ->10), and save the log file, when i was trying to
run the above qconf command. I don't know if you want to check out some
of then?

Hope tat you could help us in this issue.

Regards


-- 
-- Jérôme
La violencia es el último recurso del incompetente.
(Fundación, Isaac Azimov)
___
users mailing list
users@gridengine.org
https://gridengine.org/mailman/listinfo/users

Re: [gridengine users] install_qmaster issue

[Jerome]

Le 15/03/2018 à 08:21, Arnau a écrit :
> Hi, 
> 
> when do you get this error?
> is the qmaster really running? can you check if there is anything
> running in port 6444? 
> if it crashes try to enable debug
> http://gridscheduler.sourceforge.net/howto/troubleshooting.html .
> 
> Best,
> 
> 2018-03-14 22:15 GMT+01:00 Jerome  >:
> 
> Dear all
> 
> I've trying to installing SoGE on a fres Debian 9.4 serveur. I've to
> compile for my own the deb pacakge due to change in the libssl library
> between 1.0 and 1.1. Y download the ultimate from gitlab version
> 9b99ef86. All compile well
> 
> Now, during the process nstalation, i got an issue that i can't resolve,
> using install_qmaster:
> 
> error: commlib error: got read error (closing "xx.yy.zz/qmaster/1")
> unable to contact qmaster using port 6444 on host "xx.yy.zz"
> 
> Same message in the log of the sge master:
> 
> 03/14/2018 15:01:14|  main|xx|I|starting up SGE 8.1.9 (lx-amd64)
> 03/14/2018 15:01:57|listen|xx|E|commlib error: got select error (Broken
> pipe)
> 

Dear Arnau

I do some test with the debug on sge_qmaster. When the dl value is 4,
i'v got this message:

  9815   4846  listener000 commlib error: got select error (Broken
pipe)  9816   4846  listener000 <-- sge_gdi2_get_any_request()
../libs/gdi/sge_gdi2.c 711 }

I've read the C file and check that is inside the gdi library...

I'm sorry but that all tat i can check for now...

Hopa that could help¿

Regars



-- 
-- Jérôme
Ça doit être chouette de tout savoir comme ça!
C'est dommage que ça ne lui sert plus à rien, à papa,
puisqu'il ne va plus à l'école.
(Histoires inédites du Petit Nicolas, Goscinny & Sempé)
___
users mailing list
users@gridengine.org
https://gridengine.org/mailman/listinfo/users

Re: [gridengine users] install_qmaster issue

[Jerome]

Le 15/03/2018 à 08:21, Arnau a écrit :
> Hi, 
> 
> when do you get this error?
> is the qmaster really running? can you check if there is anything
> running in port 6444? 
> if it crashes try to enable debug
> http://gridscheduler.sourceforge.net/howto/troubleshooting.html .
> 

Dear Arnau

The error is durin the process instalation of the master host. It power
up the qmaster daemon. The issue occurs when it's trying to configure
the admin host, and after when it try to configure the basic queue
configuration.
The qmaster is running , and more, i can connect to it with a simple
telnet xx.yy 6444.

I will do the debuging process and advise you what occurs.

Thank's a lot to answer me!

Regards


-- 
-- Jérôme
Il est bon de lire entre les lignes, cela fatigue moins les yeux.
(Sacha Guitry)
___
users mailing list
users@gridengine.org
https://gridengine.org/mailman/listinfo/users