gridscheduler-developers Mailing List for Open Grid Scheduler
Brought to you by:
rayrayson,
ron_chen_123
You can subscribe to this list here.
2010 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(1) |
---|---|---|---|---|---|---|---|---|---|---|---|---|
2011 |
Jan
|
Feb
(2) |
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
(2) |
Nov
|
Dec
|
2012 |
Jan
(1) |
Feb
|
Mar
|
Apr
|
May
(2) |
Jun
(1) |
Jul
|
Aug
|
Sep
|
Oct
(4) |
Nov
(1) |
Dec
|
2013 |
Jan
|
Feb
(3) |
Mar
|
Apr
|
May
|
Jun
(5) |
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
2022 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
(1) |
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
From: Tony R. <to...@to...> - 2022-06-13 19:37:08
|
I've just upgraded from Ubuntu 20.04 LTS to 22.04 and it broke gridengine. Specifically /var/lib/gridengine/util/arch makes this call: # libc=/lib/x86_64-linux-gnu/libc.so.6 libc_string=`$libc | head -n 1` On Ubuntu 20.04 LTS this works (output below), on 22.04 libc is not executable: root@think1:/home/tonyr/exp# ls -lta /lib/x86_64-linux-gnu/libc.so.6 -rw-r--r-- 1 root root 2216304 Mar 4 02:54 /lib/x86_64-linux-gnu/libc.so.6 It is sufficient to silently ignore the error, the rest of the code finds the right answer. That is, make the line: libc_string=`$libc | head -n 1` 2> /dev/null There are no doubt better solutions, this one has the advantage that it's minimal. Hope this helps. Tony think0 tonyr: /lib/x86_64-linux-gnu/libc.so.6 GNU C Library (Ubuntu GLIBC 2.31-0ubuntu9.9) stable release version 2.31. Copyright (C) 2020 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. Compiled by GNU CC version 9.4.0. libc ABIs: UNIQUE IFUNC ABSOLUTE For bug reporting instructions, please see: <https://bugs.launchpad.net/ubuntu/+source/glibc/+bugs>. think0 tonyr: /lib/x86_64-linux-gnu/libc.so.6 GNU C Library (Ubuntu GLIBC 2.31-0ubuntu9.9) stable release version 2.31. Copyright (C) 2020 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. Compiled by GNU CC version 9.4.0. libc ABIs: UNIQUE IFUNC ABSOLUTE For bug reporting instructions, please see: <https://bugs.launchpad.net/ubuntu/+source/glibc/+bugs>. |
From: Ryan L. <ry...@be...> - 2013-06-28 17:15:07
|
Hi Rayson, I was just wondering if patches were being actively reviewed, bugs being looked at, etc. We've run into some issues and I'll document and report them if someone were to actually read them. :) Is the SVN repo at SourceForge not the primary repository for OGS/GE? Are you (scalablelogic?) developing things in house and then pushing those changes to SourceForge upon new releases? If so, is your in house repo public? Ryan On Thu, Jun 27, 2013 at 9:37 PM, Rayson Ho <ray...@gm...> wrote: > HI Ryan, > > What particular features are you looking for? We are creating software > and solutions around Open Grid Scheduler / Grid Engine, and when we > are ready, we will release a new OGS/GE version. > > BTW, I was on the call with a partner this morning (EST), we > definitely will announce & release something interesting for SC13. > > Rayson > > > > On Thu, Jun 27, 2013 at 5:03 PM, Ryan Lovett <ry...@be...> wrote: > > Hello, > > > > Is development on Open Grid Scheduler still active? A sourceforge wiki > page > > was just modified but the svn tree hasn't seen any commits lately. This > > isn't a criticism, just wondering... > > > > Ryan > > > > > ------------------------------------------------------------------------------ > > This SF.net email is sponsored by Windows: > > > > Build for Windows Store. > > > > http://p.sf.net/sfu/windows-dev2dev > > _______________________________________________ > > Gridscheduler-developers mailing list > > Gri...@li... > > https://lists.sourceforge.net/lists/listinfo/gridscheduler-developers > > > |
From: William H. <w....@uc...> - 2013-06-28 15:13:42
|
On Fri, 2013-06-28 at 14:24 +0000, Esztermann, Ansgar wrote: > Hello List, > > I am currently trawling through OGS sources in search of two annoying[1] bugs[2]. In daemons/qmaster/sge_sched_thread.c, select_assign_debit(), I came across this code: > > if (is_start) { > /* ... */ > result = sge_select_parallel_environment(&a, pe_list); > } > > if (result == DISPATCH_NOT_AT_TIME) { > if (is_reserve) { > /* ... */ > result = sge_select_parallel_environment(&a, pe_list); > > if (result == DISPATCH_OK) { > result = DISPATCH_NOT_AT_TIME; /* this job got a reservation */ > } > } else { > result = DISPATCH_NEVER_CAT; > } > > This looks like jobs that cannot start right away and do not want a reservation either are not considered startable -- ever (even their category gets blocked). > Am I missing something? > I believe that grid engine throws all this away at the start of the next scheduling cycle so never just means not now and not at a predictable time in the future (since it didn't request a reservation). Jobs in the same category should be blocked since that is the point of categories AIUI. Basically just a badly named constant... William > Thanks, > > A. > > 1: Jobs are sometimes not started: http://gridengine.org/pipermail/users/2013-June/006279.html > 2: Invalid XML from qstat: http://gridengine.org/pipermail/users/2013-February/005528.html |
From: Esztermann, A. <Ans...@mp...> - 2013-06-28 14:24:35
|
Hello List, I am currently trawling through OGS sources in search of two annoying[1] bugs[2]. In daemons/qmaster/sge_sched_thread.c, select_assign_debit(), I came across this code: if (is_start) { /* ... */ result = sge_select_parallel_environment(&a, pe_list); } if (result == DISPATCH_NOT_AT_TIME) { if (is_reserve) { /* ... */ result = sge_select_parallel_environment(&a, pe_list); if (result == DISPATCH_OK) { result = DISPATCH_NOT_AT_TIME; /* this job got a reservation */ } } else { result = DISPATCH_NEVER_CAT; } This looks like jobs that cannot start right away and do not want a reservation either are not considered startable -- ever (even their category gets blocked). Am I missing something? Thanks, A. 1: Jobs are sometimes not started: http://gridengine.org/pipermail/users/2013-June/006279.html 2: Invalid XML from qstat: http://gridengine.org/pipermail/users/2013-February/005528.html -- Ansgar Esztermann DV-Systemadministration Max-Planck-Institut für biophysikalische Chemie, Abteilung 105 |
From: Rayson Ho <ray...@gm...> - 2013-06-28 04:37:32
|
HI Ryan, What particular features are you looking for? We are creating software and solutions around Open Grid Scheduler / Grid Engine, and when we are ready, we will release a new OGS/GE version. BTW, I was on the call with a partner this morning (EST), we definitely will announce & release something interesting for SC13. Rayson On Thu, Jun 27, 2013 at 5:03 PM, Ryan Lovett <ry...@be...> wrote: > Hello, > > Is development on Open Grid Scheduler still active? A sourceforge wiki page > was just modified but the svn tree hasn't seen any commits lately. This > isn't a criticism, just wondering... > > Ryan > > ------------------------------------------------------------------------------ > This SF.net email is sponsored by Windows: > > Build for Windows Store. > > http://p.sf.net/sfu/windows-dev2dev > _______________________________________________ > Gridscheduler-developers mailing list > Gri...@li... > https://lists.sourceforge.net/lists/listinfo/gridscheduler-developers > |
From: Ryan L. <ry...@be...> - 2013-06-27 21:04:02
|
Hello, Is development on Open Grid Scheduler still active? A sourceforge wiki page was just modified but the svn tree hasn't seen any commits lately. This isn't a criticism, just wondering... Ryan |
From: Nicolas J. <nj...@pa...> - 2013-02-13 19:04:38
|
Hi, Please find attached a small patch that fix the build on NetBSD -current, where all kernel memory grovelers now need to define _KMEMUSER symbol in order to make uvmexp structures visible (used in sge_loadmem function). Thanks. -- Nicolas Joly Biology IT Center Institut Pasteur, Paris. |
From: Nicolas J. <nj...@pa...> - 2013-02-08 18:57:45
|
Hi, We recently started using quotas on our cluster, and found that .sge_qquota suppoprt was broken. Here follow a sample test that illustrate this : njoly@lanfeust [~]> echo "-u *">.sge_qquota njoly@lanfeust [~]> qquota OGS/GE 2011.11p1 usage: qquota [options] [...] wc_project_list wc_project[,wc_project,...] error: ERROR! invalid option argument "*" Checking the corresponding source code show that the first token in the file is always discarded ... The attached patch, against HEAD, fix the problem by only discarding the first word on the command line (= program name) and not while parsing files. Thanks. -- Nicolas Joly Biology IT Center Institut Pasteur, Paris. |
From: Nicolas J. <nj...@pa...> - 2013-02-08 18:47:35
|
On Mon, Nov 19, 2012 at 11:27:10AM +0100, Nicolas Joly wrote: > > While testing RQS with memory complexes, i noticed that qquota fails > to report correctly non integer dynamic limits ... [...] > Checking the code, clients/common/sge_qquota.c:qquota_output() do > unconditionally report dynamic limit as a signed int value; which is > just wrong. > > The attached patch against trunk, which unconditionally calls > sge_get_dominant_stringval() for both static and dynamic limit fix the > problem. The limit string is correctly set for all supported types. Hi, Please find attached, a new patch which includes an update for sge_resource_quota(5) man page to note that MEMORY type comlexes can be used for RQS; the previous code fix patch remaining unchanged. Thanks. -- Nicolas Joly Biology IT Center Institut Pasteur, Paris. |
From: Nicolas J. <nj...@pa...> - 2012-11-19 10:27:08
|
Hi, While testing RQS with memory complexes, i noticed that qquota fails to report correctly non integer dynamic limits ... Here follow a few step that show the problem. 1) Create a memory consumable complex. njoly@petaure [~]> qconf -sc #name shortcut type relop requestable consumable default urgency #---------------------------------------------------------------------------------------- mem_rsvd mr MEMORY <= YES YES 0 0 2) Add a it to nodes with a huge default value. njoly@petaure [~]> qconf -se petaure.lan hostname petaure.lan load_scaling NONE complex_values NONE [...] njoly@petaure [~]> qconf -aattr exechost complex_values mem_rsvd=999G petaure.lan nj...@pe...n modified "petaure.lan"in exechost list hostname petaure.lan load_scaling NONE complex_values mem_rsvd=999G 3) Create a RQS to dynamically set this limit on each node. njoly@petaure [~]> qconf -srqs { name mem_rsvd_limit description Dynamically compute mem_rsvd limit enabled TRUE limit hosts {*} to mem_rsvd=$mem_total } 4) Submit a job that require this new resource and watch the corresponding qquota output. njoly@petaure [~]> qhost HOSTNAME ARCH NCPU LOAD MEMTOT MEMUSE SWAPTO SWAPUS ------------------------------------------------------------------------------- global - - - - - - - petaure nbsd-x86_64 2 0.21 1.9G 354.7M 4.0G 0.0 njoly@petaure [~]> qrsh -l mem_rsvd=1G sleep 60 njoly@petaure [~]> qquota resource quota rule limit filter -------------------------------------------------------------------------------- mem_rsvd_limit/1 mem_rsvd=1.000G/2002 hosts petaure njoly@petaure [~]> qquota -xml <?xml version='1.0'?> <qquota_result xmlns="http://gridscheduler.svn.sourceforge.net/viewvc/gridscheduler/trunk/source/dist/util/resources/schemas/qquota/qquota.xsd?revision=11"> <qquota_rule name='mem_rsvd_limit/1'> <hosts>petaure</hosts> <limit resource='mem_rsvd' limit='2002665472' value='1.000G'/> </qquota_rule> </qquota_result> Here the reported should be formatted like all other ones ... Checking the code, clients/common/sge_qquota.c:qquota_output() do unconditionally report dynamic limit as a signed int value; which is just wrong. The attached patch against trunk, which unconditionally calls sge_get_dominant_stringval() for both static and dynamic limit fix the problem. The limit string is correctly set for all supported types. njoly@petaure [~]> qquota resource quota rule limit filter -------------------------------------------------------------------------------- mem_rsvd_limit/1 mem_rsvd=1.000G/1.86 hosts petaure njoly@petaure [~]> qquota -xml <?xml version='1.0'?> <qquota_result xmlns="http://gridscheduler.svn.sourceforge.net/viewvc/gridscheduler/trunk/source/dist/util/resources/schemas/qquota/qquota.xsd?revision=11"> <qquota_rule name='mem_rsvd_limit/1'> <hosts>petaure</hosts> <limit resource='mem_rsvd' limit='1.865G' value='1.000G'/> </qquota_rule> </qquota_result> Hope this helps, Regards. -- Nicolas Joly Biology IT Center Institut Pasteur, Paris. |
From: Rayson Ho <ray...@gm...> - 2012-10-26 05:15:59
|
Thanks a lot, Marco, for the patch. Some of our users reported this problem on Debian-based systems (that also includes Ubuntu) a few months ago. But we develop on RHEL based systems (CentOS & Oracle Linux) and thus it works for us thus far... I think it makes sense as the dl* functions are used, and thus we should add it to the Makefile. Rayson On Fri, Oct 26, 2012 at 12:54 AM, Marco Schmidt <mar...@gm...> wrote: > Hello Open Grid Scheduler Developer! > > First I would like to say "Thanks for the great job!!!" to all Open Grid > Scheduler Developer! > > While compiling on Ubuntu/Debian I got an error and a solution for it. > First the error: > > gcc -DSGE_ARCH_STRING=\"linux-x64\" -O3 -Wall -Wstrict-prototypes > -DUSE_POLL -DLINUX -DLINUX_NATIVE -DLINUXX64 -DLINUXX64 -D_GNU_SOURCE > -DGETHOSTBYNAME_R6 -DGETHOSTBYADDR_R8 -DHAS_VSNPRINTF -DHAS_IN_PORT_T > -DLOAD_OPENSSL -I/build/berkeleydb/include/ -DTARGET_64BIT > -DSPOOLING_classic -DSECURE -DCOMPILE_DC -D__SGE_COMPILE_WITH_GETTEXT__ > -D__SGE_NO_USERMAPPING__ -DTHREADBINDING -DHWLOC -I../common -I../libs > -I../libs/uti -I../libs/juti -I../libs/gdi -I../libs/japi > -I../libs/sgeobj -I../libs/cull -I../libs/rmon -I../libs/comm > -I../libs/comm/lists -I../libs/sched -I../libs/evc -I../libs/evm > -I../libs/mir -I../libs/lck -I../daemons/common -I../daemons/qmaster > -I../daemons/execd -I../clients/common -I. -I/local/jdk1.7.0_09//include > -I/local/jdk1.7.0_09//include/linux -o test_resource_utilization > -L/build/berkeleydb/lib/ -L. -Wl,-rpath,\$ORIGIN/../../lib/linux-x64 > test_resource_utilization.o sge_mt_init.o libsched.a libgdi.a > libsgeobj.a libsgeobjd.a libcull.a -ldl libcomm.a libcommlists.a > libuti.a liblck.a librmon.a -lhwloc -ljemalloc -lm -lpthread > libcomm.a(cl_ssl_framework.o): In function `cl_com_ssl_framework_cleanup': > cl_ssl_framework.c:(.text+0x2f24): undefined reference to `dlclose' > libcomm.a(cl_ssl_framework.o): In function `cl_com_ssl_setup_connection': > cl_ssl_framework.c:(.text+0x38aa): undefined reference to `dlopen' > cl_ssl_framework.c:(.text+0x38c9): undefined reference to `dlsym' > cl_ssl_framework.c:(.text+0x38fa): undefined reference to `dlsym' > cl_ssl_framework.c:(.text+0x394a): undefined reference to `dlsym' > cl_ssl_framework.c:(.text+0x3997): undefined reference to `dlsym' > cl_ssl_framework.c:(.text+0x39e4): undefined reference to `dlsym' > libcomm.a(cl_ssl_framework.o):cl_ssl_framework.c:(.text+0x3a31): more > undefined references to `dlsym' follow > libcomm.a(cl_ssl_framework.o): In function `cl_com_ssl_setup_connection': > cl_ssl_framework.c:(.text+0x5a62): undefined reference to `dlerror' > collect2: ld returned 1 exit status > make: *** [test_resource_utilization] Error 1 > not done > > Next the solution: > "dlsym" is part of libdl, which is referenced in the line "-ldl". I am > not sure, why, but my solution was to put the "-ldl" at the end of the > line. > The modification of the corresponding Makefile is in the attachment as a > svn diff. > > Greetings ... > Marco Schmidt > > > > ------------------------------------------------------------------------------ > Everyone hates slow websites. So do we. > Make your web apps faster with AppDynamics > Download AppDynamics Lite for free today: > http://p.sf.net/sfu/appdyn_sfd2d_oct > _______________________________________________ > Gridscheduler-developers mailing list > Gri...@li... > https://lists.sourceforge.net/lists/listinfo/gridscheduler-developers > |
From: Marco S. <mar...@gm...> - 2012-10-26 05:10:54
|
Hello Open Grid Scheduler Developer! First I would like to say "Thanks for the great job!!!" to all Open Grid Scheduler Developer! While compiling on Ubuntu/Debian I got an error and a solution for it. First the error: gcc -DSGE_ARCH_STRING=\"linux-x64\" -O3 -Wall -Wstrict-prototypes -DUSE_POLL -DLINUX -DLINUX_NATIVE -DLINUXX64 -DLINUXX64 -D_GNU_SOURCE -DGETHOSTBYNAME_R6 -DGETHOSTBYADDR_R8 -DHAS_VSNPRINTF -DHAS_IN_PORT_T -DLOAD_OPENSSL -I/build/berkeleydb/include/ -DTARGET_64BIT -DSPOOLING_classic -DSECURE -DCOMPILE_DC -D__SGE_COMPILE_WITH_GETTEXT__ -D__SGE_NO_USERMAPPING__ -DTHREADBINDING -DHWLOC -I../common -I../libs -I../libs/uti -I../libs/juti -I../libs/gdi -I../libs/japi -I../libs/sgeobj -I../libs/cull -I../libs/rmon -I../libs/comm -I../libs/comm/lists -I../libs/sched -I../libs/evc -I../libs/evm -I../libs/mir -I../libs/lck -I../daemons/common -I../daemons/qmaster -I../daemons/execd -I../clients/common -I. -I/local/jdk1.7.0_09//include -I/local/jdk1.7.0_09//include/linux -o test_resource_utilization -L/build/berkeleydb/lib/ -L. -Wl,-rpath,\$ORIGIN/../../lib/linux-x64 test_resource_utilization.o sge_mt_init.o libsched.a libgdi.a libsgeobj.a libsgeobjd.a libcull.a -ldl libcomm.a libcommlists.a libuti.a liblck.a librmon.a -lhwloc -ljemalloc -lm -lpthread libcomm.a(cl_ssl_framework.o): In function `cl_com_ssl_framework_cleanup': cl_ssl_framework.c:(.text+0x2f24): undefined reference to `dlclose' libcomm.a(cl_ssl_framework.o): In function `cl_com_ssl_setup_connection': cl_ssl_framework.c:(.text+0x38aa): undefined reference to `dlopen' cl_ssl_framework.c:(.text+0x38c9): undefined reference to `dlsym' cl_ssl_framework.c:(.text+0x38fa): undefined reference to `dlsym' cl_ssl_framework.c:(.text+0x394a): undefined reference to `dlsym' cl_ssl_framework.c:(.text+0x3997): undefined reference to `dlsym' cl_ssl_framework.c:(.text+0x39e4): undefined reference to `dlsym' libcomm.a(cl_ssl_framework.o):cl_ssl_framework.c:(.text+0x3a31): more undefined references to `dlsym' follow libcomm.a(cl_ssl_framework.o): In function `cl_com_ssl_setup_connection': cl_ssl_framework.c:(.text+0x5a62): undefined reference to `dlerror' collect2: ld returned 1 exit status make: *** [test_resource_utilization] Error 1 not done Next the solution: "dlsym" is part of libdl, which is referenced in the line "-ldl". I am not sure, why, but my solution was to put the "-ldl" at the end of the line. The modification of the corresponding Makefile is in the attachment as a svn diff. Greetings ... Marco Schmidt |
From: Rayson Ho <ray...@gm...> - 2012-10-25 22:52:59
|
Thanks, and it will be applied to the 2012 release! Rayson On Thu, Oct 25, 2012 at 1:46 PM, Nicolas Joly <nj...@pa...> wrote: > > Hi, > > Please find a small patch, against SVN trunk, that make mem_free value > much more realistic on NetBSD. > > It simply adds the file cache used memory to the "free" memory count > which can be used by applications. > > Thanks. > > -- > Nicolas Joly > > Biology IT Center > Institut Pasteur, Paris. > > ------------------------------------------------------------------------------ > Everyone hates slow websites. So do we. > Make your web apps faster with AppDynamics > Download AppDynamics Lite for free today: > http://p.sf.net/sfu/appdyn_sfd2d_oct > _______________________________________________ > Gridscheduler-developers mailing list > Gri...@li... > https://lists.sourceforge.net/lists/listinfo/gridscheduler-developers > |
From: Nicolas J. <nj...@pa...> - 2012-10-25 17:44:50
|
Hi, Please find a small patch, against SVN trunk, that make mem_free value much more realistic on NetBSD. It simply adds the file cache used memory to the "free" memory count which can be used by applications. Thanks. -- Nicolas Joly Biology IT Center Institut Pasteur, Paris. |
From: <joc...@we...> - 2012-06-27 08:05:00
|
<html><head></head><body><div style="font-family: Verdana;font-size: 12.0px;"><div>Hello Open Grid Scheduler Team<br/></div><div><br/></div><div>At the moment I try to compile open Grid Scheduler under Windows 7 64 bit. I do it, as descript in this message<br/></div><div>http://sourceforge.net/mailarchive/message.php?msg_id=27637625<br/></div><div><br/></div><div>I have encountert the problems and get stuck as discript with following error message:<br/></div><div><br/></div><div>Building in directory: /dev/fs/E/HPC/GE2011.11/source<br/>making in WIN32_X86/ for WIN32_X86 at host Desktop<br/>_________S_G_E__D_E_P_E_N_D_______________<br/>gmake: Nothing to be done for `all'.<br/>____W_I_N_3_2__A_P_P_L_I_C_A_T_I_O_N_S____<br/>cl.exe /MT /D WINDOWS /D _CONSOLE /D _CRT_SECURE_NO_WARNINGS /W3 /WX /nologo /c /EHsc -I'/dev/fs/E/Microsoft Visual Studio 10.0/VC/include' -I'..\\extra_h' -I'..\\common' -I'..\\utilbin' -I'..\\libs\\juti' -c ../utilbin/windows/SGE_Helper_Service/Communication.cpp<br/>Communication.cpp<br/>../utilbin/windows/SGE_Helper_Service/Communication.cpp(33) : fatal error C1083: Datei (Include) kann nicht ge”ffnet werden: "afxtempl.h": No such file or directory<br/>not done<br/><br/></div><div><br/></div><div>I compile with the option:<br/></div><div> ./aimk -no-java -no-jni -no-secure -spool-classic -only-depend<br/></div><div><br/></div><div>After some research concerning the "afxtempl.h", I found, that this is a header file of the Microsoft MFC library:<br/></div><div>http://msdn.microsoft.com/de-de/library/bk77x1wx%28v=vs.90%29.aspx<br/></div><div><br/></div><div>I switched from Visual Studio Express 2010 to Visual Studio Express 2005, but I didn't get this header file. I think, I need a professional version of Microsoft Visual Studio, do install this pack<br/></div><div>http://www.microsoft.com/en-us/download/details.aspx?id=6922<br/></div><div><br/></div><div>I tried to download the header files from<br/></div><div>http://www.cppdoc.com/example/mfc/classdoc/MFC/<br/></div><div><br/></div><div>Now I get this error message<br/></div><div><br/></div><div>Building in directory: /dev/fs/E/HPC/GE2011.11/source<br/>making in WIN32_X86/ for WIN32_X86 at host Desktop<br/>_________S_G_E__D_E_P_E_N_D_______________<br/>gmake: Nothing to be done for `all'.<br/>____W_I_N_3_2__A_P_P_L_I_C_A_T_I_O_N_S____<br/>cl.exe /MT /D WINDOWS /D _CONSOLE /D _CRT_SECURE_NO_WARNINGS /W3 /WX /nologo /c /EHsc -I'E:\Microsoft_Visual_Studio_9\VC\include' -I'..\\extra_h' -I'..\\common' -I'..\\utilbin' -I'..\\libs\\juti' -I'..\\afx' -c ../utilbin/windows/SGE_Helper_Service/Communication.cpp<br/>Communication.cpp<br/>..\\afx\afx.inl(231) : error C2220: Warnung wird als Fehler interpretiert, es wurde keine object-Datei generiert.<br/>..\\afx\afx.inl(231) : warning C4244: 'return': Konvertierung von 'const time_t' in 'LONG', m”glicher Datenverlust<br/>..\\afx\afx.inl(233) : warning C4244: 'return': Konvertierung von 'const time_t' in 'LONG', m”glicher Datenverlust<br/>..\\afx\afx.inl(237) : warning C4244: 'return': Konvertierung von 'const time_t' in 'LONG', m”glicher Datenverlust<br/>..\\afx\afx.inl(241) : warning C4244: 'return': Konvertierung von 'const time_t' in 'LONG', m”glicher Datenverlust<br/>..\\afx\afxtempl.h(776) : warning C4346: 'CList<TYPE,ARG_TYPE>::CNode': Abh„ngiger Name ist kein Typ<br/> Pr„fix mit 'typename' zum Angeben eines Typs<br/>..\\afx\afxtempl.h(776) : error C2143: Syntaxfehler: Es fehlt ';' vor '*'<br/>..\\afx\afxtempl.h(776) : error C4430: Fehlender Typspezifizierer - int wird angenommen. Hinweis: "default-int" wird von C++ nicht untersttzt.<br/>..\\afx\afxtempl.h(776) : fatal error C1903: Weiterverarbeitung nach vorherigem Fehler nicht m”glich; Kompilierung wird abgebrochen.<br/>not done<br/><br/></div><div>So I tried the no-win32 option<br/><div> ./aimk -no-java -no-jni -no-secure -spool-classic -no-wine32 -only-depend<br/></div><div><br/></div><div>This runs with success. Now<br/></div><div>./scripts/zerodepend<br/></div><div><br/></div><div>And<br/> <div> ./aimk -no-java -no-jni -no-secure -spool-classic -no-wine32 depend<br/></div><div><br/></div><div>It breaks with<br/></div><div><br/></div><div>checking for gcc... gcc<br/>checking whether the C compiler works... no<br/>configure: error: in `/dev/fs/E/HPC/GE2011.11/source/3rdparty/qmake/WIN32_X86':<br/>configure: error: C compiler cannot create executables<br/>See `config.log' for more details.<br/>gmake: Makefile: No such file or directory<br/>gmake: *** No rule to make target `Makefile'. Stop.<br/>not done<br/></div></div><br/></div><div>Can you give me a hint, how I should proceed further? My goal is to be able to send jobs to a qmaster host and to execute jobs on a windows 7 64 bit machine. <br/></div><div>Or can you give me I location, where I can get the binarys for Windows?<br/></div><div><br/></div><div>Best regards<br/></div><div> Jochen Rath<br/></div></div> <br><br><table cellpadding="0" cellspacing="0" border="0"><tr><td style="font-family:verdana; font-size:12px; line-height:17px;border-top:1px solid #000000">Ihr WEB.DE Postfach immer dabei: die kostenlose WEB.DE Mail App für iPhone und Android. <br><a href="https://produkte.web.de/freemail_mobile_startseite/"><b>https://produkte.web.de/freemail_mobile_startseite/</b></a></td></tr></table> </body></html> |
From: Rayson Ho <ray...@gm...> - 2012-05-29 14:15:11
|
Thanks Nicolas. We will review the patch and if everything goes well, we will add it into OGS/GE 2011.11 update 1! Rayson On Tue, May 29, 2012 at 8:28 AM, Nicolas Joly <nj...@pa...> wrote: > > Hi, > > We just discovered that setting a cluster queue with hosts but zero > slots give wrong load report for that queue : > > njoly@lanfeust [~]> qstat -g c > CLUSTER QUEUE CQLOAD USED RES AVAIL TOTAL aoACDS cdsuE > -------------------------------------------------------------------------------- > empty nan 0 0 0 0 0 0 > > Looking into the corresponding code show that there is a division by > zero in that case (clients/common/sge_cqueue_qstat.c). > > The attached patch do fix it by not marking load as available if there > is no slots in the cluster queue. It then report `-NA-'instead of > bogus nan previously. > > njoly@lanfeust [~]> qstat -g c > CLUSTER QUEUE CQLOAD USED RES AVAIL TOTAL aoACDS cdsuE > -------------------------------------------------------------------------------- > empty -NA- 0 0 0 0 0 0 > > Hope this helps, > Regards. > > -- > Nicolas Joly > > Projects and Developments in Bioinformatics > Institut Pasteur, Paris. > > ------------------------------------------------------------------------------ > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and > threat landscape has changed and how IT managers can respond. Discussions > will include endpoint security, mobile security and the latest in malware > threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > _______________________________________________ > Gridscheduler-developers mailing list > Gri...@li... > https://lists.sourceforge.net/lists/listinfo/gridscheduler-developers > |
From: Nicolas J. <nj...@pa...> - 2012-05-29 12:48:01
|
Hi, We just discovered that setting a cluster queue with hosts but zero slots give wrong load report for that queue : njoly@lanfeust [~]> qstat -g c CLUSTER QUEUE CQLOAD USED RES AVAIL TOTAL aoACDS cdsuE -------------------------------------------------------------------------------- empty nan 0 0 0 0 0 0 Looking into the corresponding code show that there is a division by zero in that case (clients/common/sge_cqueue_qstat.c). The attached patch do fix it by not marking load as available if there is no slots in the cluster queue. It then report `-NA-'instead of bogus nan previously. njoly@lanfeust [~]> qstat -g c CLUSTER QUEUE CQLOAD USED RES AVAIL TOTAL aoACDS cdsuE -------------------------------------------------------------------------------- empty -NA- 0 0 0 0 0 0 Hope this helps, Regards. -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris. |
From: Nicolas J. <nj...@pa...> - 2012-01-24 12:07:17
|
Hi, The attached patch do fix OGS build on NetBSD. It moves, in aimk, CFLAGS definition before DEPEND_FLAGS which use it. Thanks. -- Nicolas Joly Projects and Developments in Bioinformatics Institut Pasteur, Paris. |
From: Rayson Ho <ray...@gm...> - 2011-10-07 06:44:47
|
Hi Leonardo, This is fixed in the code for the next release of Grid Scheduler. I first encountered it last year as Fedora is one of my main platforms (the other is Oracle Linux 5.6), but I did not expect Freetype2 to be picked up by RH 6 so quickly -- so I guess it is time to for me to install Oracle Linux 6 (as RH6 is too expensive and CentOS 6 was really late). Anyway, back to the fix, I have a quick fix which worked for me in SGE 6.2u5p2. You can get it by running: % wget http://gridscheduler.sourceforge.net/patches/aimk % chmod u+x aimk Then continue to compile rest of the qmon code: % ./aimk -no-java -no-jni -no-secure -spool-classic -no-dump The full fix will come when we release the next release of Grid Scheduler. Rayson On Thu, Oct 6, 2011 at 12:01 PM, Leonardo Pereira <leo...@gm...> wrote: > Hello all, > > I was trying to compile the Open Grid Scheduler on Scientific Linux 6.1 > (kernel 2.6.32), but an error ocurred. > I followed these sequence: > > ./aimk -no-java -no-jni -no-secure -spool-classic -no-dump -only-depend > ok > > ./scripts/zerodepend > ok > > ./aimk -no-java -no-jni -no-secure -spool-classic -no-dump depend > ok > > ./aimk -no-java -no-jni -no-secure -spool-classic -no-dump > get this error: > > . > . > . > > _________Q_M_O_N__________________________ > gcc -c -DSGE_ARCH_STRING=\"lx26-x86\" -O3 -Wall -Werror -Wstrict-prototypes > -DUSE_POLL -DLINUX -DLINUX86 -DLINUX86_26 -D_GNU_SOURCE -DGETHOSTBYNAME_R6 > -DGETHOSTBYADDR_R8 -I/off_home/gridengine/db-4.4.20/lx26-x86/include/ > -DTARGET_32BIT -DSPOOLING_classic -Wno-strict-aliasing > -D_FILE_OFFSET_BITS=64 -DPLPA -DNO_JNI -DCOMPILE_DC > -D__SGE_COMPILE_WITH_GETTEXT__ -D__SGE_NO_USERMAPPING__ -I../Xmt310 > -Wno-strict-prototypes -Wno-error -DHAVE_MEMMOVE -DXMT_HAS_STRERROR > -I../3rdparty/qmon/Xmt310 -I../3rdparty/qmon/Xbae > -I../3rdparty/qmon/iconlist -I../3rdparty/qmon/ltree > -I../3rdparty/qmon/outline -I../3rdparty/qmon/rowcol > -I../3rdparty/qmon/spinbox -I../clients/qmon -I../3rdparty/qmon/tab > -I/usr/X11R6/include -fPIC ../Xmt310/Xmt/All.c > In file included from ../Xmt310/Xmt/All.c:23: > ../Xmt310/Xmt/Xmt.h:56:19: error: Xm/Xm.h: Arquivo ou diretório não > encontrado > In file included from ../Xmt310/Xmt/Xmt.h:357, > from ../Xmt310/Xmt/All.c:23: > ../Xmt310/Xmt/Util.h:46: error: expected ‘=’, ‘,’, ‘;’, ‘asm’ or > ‘__attribute__’ before ‘XmtCreateXmString’ > ../Xmt310/Xmt/Util.h:47: error: expected ‘=’, ‘,’, ‘;’, ‘asm’ or > ‘__attribute__’ before ‘XmtCreateLocalizedXmString’ > In file included from ../Xmt310/Xmt/All.c:27: > ../Xmt310/Xmt/Layout.h:24:26: error: Xm/BulletinB.h: Arquivo ou diretório > não encontrado > make: ** [All.o] Erro 1 > not done > > Can anyone help me? > > Leonardo. > > > > > > > ------------------------------------------------------------------------------ > All the data continuously generated in your IT infrastructure contains a > definitive record of customers, application performance, security > threats, fraudulent activity and more. Splunk takes this data and makes > sense of it. Business sense. IT sense. Common sense. > http://p.sf.net/sfu/splunk-d2dcopy1 > _______________________________________________ > Gridscheduler-developers mailing list > Gri...@li... > https://lists.sourceforge.net/lists/listinfo/gridscheduler-developers > > |
From: Leonardo P. <leo...@gm...> - 2011-10-06 16:01:44
|
Hello all, I was trying to compile the Open Grid Scheduler on Scientific Linux 6.1 (kernel 2.6.32), but an error ocurred. I followed these sequence: ./aimk -no-java -no-jni -no-secure -spool-classic -no-dump -only-depend ok ./scripts/zerodepend ok ./aimk -no-java -no-jni -no-secure -spool-classic -no-dump depend ok ./aimk -no-java -no-jni -no-secure -spool-classic -no-dump get this error: . . . _________Q_M_O_N__________________________ gcc -c -DSGE_ARCH_STRING=\"lx26-x86\" -O3 -Wall -Werror -Wstrict-prototypes -DUSE_POLL -DLINUX -DLINUX86 -DLINUX86_26 -D_GNU_SOURCE -DGETHOSTBYNAME_R6 -DGETHOSTBYADDR_R8 -I/off_home/gridengine/db-4.4.20/lx26-x86/include/ -DTARGET_32BIT -DSPOOLING_classic -Wno-strict-aliasing -D_FILE_OFFSET_BITS=64 -DPLPA -DNO_JNI -DCOMPILE_DC -D__SGE_COMPILE_WITH_GETTEXT__ -D__SGE_NO_USERMAPPING__ -I../Xmt310 -Wno-strict-prototypes -Wno-error -DHAVE_MEMMOVE -DXMT_HAS_STRERROR -I../3rdparty/qmon/Xmt310 -I../3rdparty/qmon/Xbae -I../3rdparty/qmon/iconlist -I../3rdparty/qmon/ltree -I../3rdparty/qmon/outline -I../3rdparty/qmon/rowcol -I../3rdparty/qmon/spinbox -I../clients/qmon -I../3rdparty/qmon/tab -I/usr/X11R6/include -fPIC ../Xmt310/Xmt/All.c In file included from ../Xmt310/Xmt/All.c:23: ../Xmt310/Xmt/Xmt.h:56:19: error: Xm/Xm.h: Arquivo ou diretório não encontrado In file included from ../Xmt310/Xmt/Xmt.h:357, from ../Xmt310/Xmt/All.c:23: ../Xmt310/Xmt/Util.h:46: error: expected ‘=’, ‘,’, ‘;’, ‘asm’ or ‘__attribute__’ before ‘XmtCreateXmString’ ../Xmt310/Xmt/Util.h:47: error: expected ‘=’, ‘,’, ‘;’, ‘asm’ or ‘__attribute__’ before ‘XmtCreateLocalizedXmString’ In file included from ../Xmt310/Xmt/All.c:27: ../Xmt310/Xmt/Layout.h:24:26: error: Xm/BulletinB.h: Arquivo ou diretório não encontrado make: ** [All.o] Erro 1 not done Can anyone help me? Leonardo. |
From: Ron C. <ron...@ya...> - 2011-02-04 20:22:56
|
Hi Chris, Grid Engine already has support for thread & process binding, which means Open Grid Scheduler also has this feature. -Ron --- On Thu, 2/3/11, Chris Jewell <chr...@wa...> wrote: > Hi All, > > Firstly, I'd like to introduce myself to the new > list. I currently work in a statistics department and, > although I'm a statistician primarily, I built and > administer our small departmental cluster. I was a SGE > user for many years on other systems, and so it was natural > for me to use it as the scheduling environment for the said > cluster. > > I'm interested in helping the development of GridScheduler, > particularly with regard to CPU binding, which I see as an > essential feature of a grid scheduler in today's multicore > environments. In fact, I had a thread running a couple > of months ago on the SGE mailing list -- a thread that we > all agreed should be archived for future reference. > I'm not sure what happened to these archives after the > proverbial sinking of the Titanic! > > Anyway, congratulations to the team for creating the > GridScheduler project, and I'll be waiting with baited > breath for news! > > Regards, > > Chris > > -- > Dr Chris Jewell > Department of Statistics > University of Warwick > Coventry > CV4 7AL > UK > Tel: +44 (0)24 7615 0778 > > > > > > > ------------------------------------------------------------------------------ > Special Offer-- Download ArcSight Logger for FREE (a $49 > USD value)! > Finally, a world-class log management solution at an even > better price-free! > Download using promo code Free_Logger_4_Dev2Dev. Offer > expires > February 28th, so secure your free ArcSight Logger TODAY! > http://p.sf.net/sfu/arcsight-sfd2d > _______________________________________________ > Gridscheduler-developers mailing list > Gri...@li... > https://lists.sourceforge.net/lists/listinfo/gridscheduler-developers > |
From: Chris J. <chr...@wa...> - 2011-02-03 10:23:34
|
Hi All, Firstly, I'd like to introduce myself to the new list. I currently work in a statistics department and, although I'm a statistician primarily, I built and administer our small departmental cluster. I was a SGE user for many years on other systems, and so it was natural for me to use it as the scheduling environment for the said cluster. I'm interested in helping the development of GridScheduler, particularly with regard to CPU binding, which I see as an essential feature of a grid scheduler in today's multicore environments. In fact, I had a thread running a couple of months ago on the SGE mailing list -- a thread that we all agreed should be archived for future reference. I'm not sure what happened to these archives after the proverbial sinking of the Titanic! Anyway, congratulations to the team for creating the GridScheduler project, and I'll be waiting with baited breath for news! Regards, Chris -- Dr Chris Jewell Department of Statistics University of Warwick Coventry CV4 7AL UK Tel: +44 (0)24 7615 0778 |
From: Ron C. <ron...@ya...> - 2010-12-31 23:12:22
|
After a busy year in 2010, the project really starts this year -- 2011!! -Ron |