Hi Christian, I went for the second option. Now everything is working fine.
Thank you, Samatha [EMAIL PROTECTED] wrote: > Hello, > > This is called by 4.0.5, right? > /opt/globus-4.0.5/libexec/globus-gridmap-and-execute [...] <-- This would be > wrong in my opinion > While this is called by the older version: > /opt/globus/libexec/globus-gridmap-and-execute [...] > > In my opinion, this is a small bug with globus, so that it does not use the > GLOBUS_LOCATION environment variable when calling globus-gridmap-and-execute, > but the --prefix variable that was given at configure time. > > So what the old version executes matches the entries in the sudoers file, > while the new one does not. That is what I think causes the problems for you > with sudo. It's clear that your sudo configuration does not work if > /opt/globus-4.0.5 is called but /opt/globus is in the sudoers file. So your > problem is not really the sudoers configuration (though you could resolve the > issues here too by appending the version), but in the way that the command is > executed by globus in the first place. > > I had similar problems, and what I did was the following: > 1) install 4.0.3 to /opt/globus (direct installation, ./configure > --prefix=/opt/globus) > 2) moved /opt/globus to /opt/globus-4.0.3, installed 4.0.5 with ./configure > --prefix=/opt/globus-4.0.5 > 3) created a symlink from /opt/globus to /opt/globus-4.0.5 > > The GLOBUS_LOCATION and so on are /opt/globus, but I think that the fact that > I used --prefix=/opt/globus-4.0.5 causes globus to hard-code the path to > globus-gridmap-and-execute to /opt/globus-4.0.5, the prefix directory, > instead of using /opt/globus, the GLOBUS_LOCATION. > I have not reinstalled 4.0.5 with --prefix=/opt/globus and then moved it to > /opt/globus-4.0.5 to verify this though. If you can, maybe you want to give > this a try. > > In my opinion, you've got two options: > 1) Add the versions to the sudoers file, that is, fix the mismatch at the > sudo end. (workaround, but simple) > 2) Fix the way the script is called by globus, that is fix the mismatch at > the globus end (clean solution, but may require a reinstallation, or messing > with who-knows-how-many globus files). You could either edit the paths > wherever they are hard-coded (maybe grep -R for globus-4.0.5 or > globus-gridmap-and-execute), or you could reinstalling globus with > --prefix=/opt/globus, then move it to /opt/globus-4.0.5 and create the > symlink. > > Hope it's clearer what I mean now, > Christian > > -----Ursprüngliche Nachricht----- > Von: Samatha Kottha [mailto:[EMAIL PROTECTED] > Gesendet: Dienstag, 2. Oktober 2007 10:22 > An: Gabriel Mateescu; [EMAIL PROTECTED]; Assfalg,Christian (AIV) BIP-DE-B > Cc: Helmut Heller; [email protected] > Betreff: Re: [gt-user] Sudo is misconfigured to run the > globus-job-manager-script.pl script > > > Dear all, > > I could not check this before as machines are heavily used. I came to > the conclusion that something is wrong with GT 4.0.5. Please correct me > if I am wrong. We have two types of servers: romeo (IA-64, Globus 4.0.2, > and Java 1.4.2) and hector (X86-64, Globus 4.0.4, and Java 1.6). Both > are running nicely with old versions of GT. Both have the same entries > in their sudo file: > > #Globus GRAM entries > globus ALL=(ALL,!root) NOPASSWD: > /opt/globus/libexec/globus-gridmap-and-execute -g > /etc/grid-security/grid-mapfile > /opt/globus/libexec/globus-job-manager-script.pl * > globus ALL=(ALL,!root) NOPASSWD: > /opt/globus/libexec/globus-gridmap-and-execute -g > /etc/grid-security/grid-mapfile > /opt/globus/libexec/globus-gram-local-proxy-tool * > > The GLOBUS_LOCATION is /opt/globus which is a symlink. The sudoer have > no problems with symlink in the older versions. Now, I compiled and > installed GT 4.0.5 into a different directory. Changed the symlink to > the new location. Everything (gsissh, GridFTP, and RFT) is working > except job submission. Here is what you see at the client side: > > [EMAIL PROTECTED]:~> globusrun-ws -submit -S -F > https://romeo.urz.tu-dresden.de:8443/wsrf/services/ManagedJobFactoryService > -s -c /bin/date > Delegating user credentials...Done. > Submitting job...Done. > Job ID: uuid:2cc1e60a-70be-11dc-b9e1-080069149999 > Termination time: 10/03/2007 08:05 GMT > Current job state: Failed > Destroying job...Done. > Cleaning up any delegated credentials...Done. > globusrun-ws: Job failed: Error code: 200 > Sudo is misconfigured to run the globus-job-manager-script.pl script for > user zimd0022. > > [EMAIL PROTECTED]:~> > > *The entries in romeo's container file are below: * > > 2007-10-02 10:05:04,977 DEBUG exec.StateMachine > [RunQueueThread_5,runScript:2987] running script submit > 2007-10-02 10:05:04,998 DEBUG exec.JobManagerScript [Thread-18,run:208] > Executing command: > /usr/bin/sudo -H -u zimd0022 -S > /opt/globus-4.0.5/libexec/globus-gridmap-and-execute -g > /etc/grid-security/grid-mapfile /opt/globus-4.0.5/libexec/globus-jo > b-manager-script.pl -m fork -f > /opt/globus-4.0.5/tmp/gram_job_mgr52269.tmp -c submit > 2007-10-02 10:05:05,160 DEBUG exec.JobManagerScript [Thread-18,run:225] > first line: null > 2007-10-02 10:05:05,161 DEBUG exec.JobManagerScript [Thread-18,run:335] > failure message: Sudo is misconfigured to run the > globus-job-manager-script.pl scri > pt for user zimd0022. > 2007-10-02 10:05:05,162 DEBUG exec.JobManagerScript > [Thread-18,setDone:345] script is done, setting done flag > 2007-10-02 10:05:05,163 DEBUG exec.StateMachine > [RunQueueThread_5,processSubmitState:1168] Done waiting for submit script > 2007-10-02 10:05:05,164 DEBUG exec.StateMachine > [RunQueueThread_5,processSubmitState:1176] script return code: 200 > 2007-10-02 10:05:05,165 DEBUG exec.StateMachine > [RunQueueThread_5,processSubmitState:1181] script return code means error! > 2007-10-02 10:05:05,177 DEBUG exec.StateMachine > [RunQueueThread_5,createFaultFromErrorCode:3131] Creating fault from > error code 200 > 2007-10-02 10:05:05,177 WARN exec.StateMachine > [RunQueueThread_5,createFaultFromErrorCode:3270] Unhandled fault code 200 > 2007-10-02 10:05:05,178 DEBUG exec.StateMachine > [RunQueueThread_5,createFaultFromErrorCode:3271] Offending Script > Command: submit > 2007-10-02 10:05:05,184 DEBUG utils.FaultUtils > [RunQueueThread_5,createFault:422] Script Command: submit > 2007-10-02 10:05:05,196 DEBUG utils.FaultUtils > [RunQueueThread_5,makeFault:460] Fault Class: class > org.globus.exec.generated.FaultType > 2007-10-02 10:05:05,196 DEBUG utils.FaultUtils > [RunQueueThread_5,makeFault:461] Resource Key: > {http://www.globus.org/namespaces/2004/10/gram/job}ResourceID > =2e3cb5a0-70be-11dc-b1ea-b1743b772918 > 2007-10-02 10:05:05,196 DEBUG utils.FaultUtils > [RunQueueThread_5,makeFault:462] Description: Error code: 200 > 2007-10-02 10:05:05,197 DEBUG utils.FaultUtils > [RunQueueThread_5,makeFault:463] Cause: java.lang.Exception: Sudo is > misconfigured to run the globus-job-man > ager-script.pl script for user zimd0022. > 2007-10-02 10:05:05,197 DEBUG utils.FaultUtils > [RunQueueThread_5,makeFault:464] State when failure occurred Unsubmitted > 2007-10-02 10:05:05,197 DEBUG utils.FaultUtils > [RunQueueThread_5,makeFault:466] Script Command: submit > 2007-10-02 10:05:05,198 DEBUG utils.FaultUtils > [RunQueueThread_5,makeFault:467] GT2 Error Code: 200 > 2007-10-02 10:05:05,225 DEBUG utils.FaultUtils > [RunQueueThread_5,makeFault:514] setting fault cause > 2007-10-02 10:05:05,227 DEBUG utils.FaultUtils > [RunQueueThread_5,makeFault:519] Script Command: submit > > *So, I took the sudo command and directly executed and gave the globus > password. * > > [EMAIL PROTECTED]:~> /usr/bin/sudo -H -u zimd0022 -S > /opt/globus-4.0.5/libexec/globus-gridmap-and-execute -g > /etc/grid-security/grid-mapfile > /opt/globus-4.0.5/libexec/globus-job-manager-script.pl -m fork -f > /opt/globus-4.0.5/tmp/gram_job_mgr32745.tmp -c submit > Password: > Sorry, try again. > Password: > Sorry, try again. > Password: > Sorry, try again. > /usr/bin/sudo: 3 incorrect password attempts > [EMAIL PROTECTED]:~> > [EMAIL PROTECTED]:~> ls -altrh /opt/ > total 24K > ........... > drwxr-xr-x 16 globus globus 4.0K 2007-09-27 15:38 globus-4.0.5 > drwxr-xr-x 16 globus globus 4.0K 2007-10-01 11:47 globus-4.0.2 > lrwxrwxrwx 1 root root 12 2007-10-01 11:59 globus -> globus-4.0.5 > drwxr-xr-x 17 root root 4.0K 2007-10-01 11:59 . > [EMAIL PROTECTED]:~> > > *Now on hector: * > > [EMAIL PROTECTED]:~> globusrun-ws -submit -S -F > https://hector.zih.tu-dresden.de:8443/wsrf/services/ManagedJobFactoryService > -s -c /bin/date > Delegating user credentials...Done. > Submitting job...Done. > Job ID: uuid:a9a0a49a-70be-11dc-bf1f-080069149999 > Termination time: 10/03/2007 08:08 GMT > Current job state: Failed > Destroying job...Done. > Cleaning up any delegated credentials...Done. > globusrun-ws: Job failed: Error code: 201 > Script stderr: > zimd0022's password: > > [EMAIL PROTECTED]:~> > > *The entries in container log: * > > 2007-10-02 10:08:32,890 DEBUG exec.StateMachine > [RunQueueThread_13,runScript:2987] running script submit > 2007-10-02 10:08:32,890 DEBUG exec.JobManagerScript [Thread-18,run:208] > Executing command: > /usr/bin/sudo -H -u zimd0022 -S > /opt/globus-4.0.5/libexec/globus-gridmap-and-execute -g > /etc/grid-security/grid-mapfile /opt/globus-4.0.5/libexec/globus-jo > b-manager-script.pl -m fork -f > /opt/globus-4.0.5/tmp/gram_job_mgr63230.tmp -c submit > 2007-10-02 10:08:32,983 DEBUG exec.JobManagerScript [Thread-18,run:225] > first line: null > 2007-10-02 10:08:32,984 DEBUG exec.JobManagerScript [Thread-18,run:335] > failure message: Script stderr: > zimd0022's password: > 2007-10-02 10:08:32,984 DEBUG exec.JobManagerScript > [Thread-18,setDone:345] script is done, setting done flag > 2007-10-02 10:08:32,985 DEBUG exec.StateMachine > [RunQueueThread_13,processSubmitState:1168] Done waiting for submit script > 2007-10-02 10:08:32,986 DEBUG exec.StateMachine > [RunQueueThread_13,processSubmitState:1176] script return code: 201 > 2007-10-02 10:08:32,986 DEBUG exec.StateMachine > [RunQueueThread_13,processSubmitState:1181] script return code means error! > 2007-10-02 10:08:32,986 DEBUG exec.StateMachine > [RunQueueThread_13,createFaultFromErrorCode:3131] Creating fault from > error code 201 > 2007-10-02 10:08:32,986 WARN exec.StateMachine > [RunQueueThread_13,createFaultFromErrorCode:3270] Unhandled fault code 201 > 2007-10-02 10:08:32,987 DEBUG exec.StateMachine > [RunQueueThread_13,createFaultFromErrorCode:3271] Offending Script > Command: submit > 2007-10-02 10:08:32,991 DEBUG utils.FaultUtils > [RunQueueThread_13,createFault:422] Script Command: submit > 2007-10-02 10:08:32,997 DEBUG utils.FaultUtils > [RunQueueThread_13,makeFault:460] Fault Class: class > org.globus.exec.generated.FaultType > 2007-10-02 10:08:32,998 DEBUG utils.FaultUtils > [RunQueueThread_13,makeFault:461] Resource Key: > {http://www.globus.org/namespaces/2004/10/gram/job}ResourceI > D=aaadb440-70be-11dc-bd72-842811339d49 > 2007-10-02 10:08:32,998 DEBUG utils.FaultUtils > [RunQueueThread_13,makeFault:462] Description: Error code: 201 > 2007-10-02 10:08:32,998 DEBUG utils.FaultUtils > [RunQueueThread_13,makeFault:463] Cause: java.lang.Exception: Script stderr: > zimd0022's password: > 2007-10-02 10:08:32,998 DEBUG utils.FaultUtils > [RunQueueThread_13,makeFault:464] State when failure occurred Unsubmitted > 2007-10-02 10:08:32,998 DEBUG utils.FaultUtils > [RunQueueThread_13,makeFault:466] Script Command: submit > 2007-10-02 10:08:32,999 DEBUG utils.FaultUtils > [RunQueueThread_13,makeFault:467] GT2 Error Code: 201 > 2007-10-02 10:08:33,006 DEBUG utils.FaultUtils > [RunQueueThread_13,makeFault:514] setting fault cause > 2007-10-02 10:08:33,007 DEBUG utils.FaultUtils > [RunQueueThread_13,makeFault:519] Script Command: submit > > *Result of sudo command direct execution. * > > [EMAIL PROTECTED]:~> /usr/bin/sudo -H -u zimd0022 -S > /opt/globus-4.0.5/libexec/globus-gridmap-and-execute -g > /etc/grid-security/grid-mapfile > /opt/globus-4.0.5/libexec/globus-job-manager-script.pl -m fork -f > /opt/globus-4.0.5/tmp/gram_job_mgr63230.tmp -c submit > Password: > Sorry, try again. > Password: > Sorry, try again. > Password: > Sorry, try again. > /usr/bin/sudo: 3 incorrect password attempts > [EMAIL PROTECTED]:~> > > [EMAIL PROTECTED]:~> ls -altrh /opt/ > total 16K > ............... > drwxr-xr-x 16 globus globus 4.0K 2007-09-20 01:43 globus-4.0.4 > lrwxrwxrwx 1 root root 12 2007-10-02 09:49 globus -> globus-4.0.5 > drwxr-xr-x 16 globus globus 4.0K 2007-10-02 09:50 globus-4.0.5 > [EMAIL PROTECTED]:~> > > So, I set the globus to again to globus-4.0.4. > > hector:/opt # rm globus > hector:/opt # ln -s globus-4.0.4 globus > hector:/opt # ls -altrh > total 16K > ....... > drwxr-xr-x 16 globus globus 4.0K 2007-09-20 01:43 globus-4.0.4 > drwxr-xr-x 16 globus globus 4.0K 2007-10-02 09:50 globus-4.0.5 > lrwxrwxrwx 1 root root 12 2007-10-02 10:15 globus -> globus-4.0.4 > hector:/opt # > > *At client: * > > [EMAIL PROTECTED]:~> globusrun-ws -submit -S -F > https://hector.zih.tu-dresden.de:8443/wsrf/services/ManagedJobFactoryService > -s -c /bin/date > Delegating user credentials...Done. > Submitting job...Done. > Job ID: uuid:d39adff8-70bf-11dc-a4b2-080069149999 > Termination time: 10/03/2007 08:16 GMT > Current job state: Active > Current job state: CleanUp-Hold > Tue Oct 2 10:16:52 CEST 2007 > Current job state: CleanUp > Current job state: Done > Destroying job...Done. > Cleaning up any delegated credentials...Done. > [EMAIL PROTECTED]:~> > > *Container log: > > *2007-10-02 10:16:52,525 DEBUG exec.StateMachine > [RunQueueThread_0,runScript:2883] running script submit > 2007-10-02 10:16:52,525 DEBUG exec.JobManagerScript [Thread-14,run:208] > Executing command: > /usr/bin/sudo -H -u zimd0022 -S > /opt/globus/libexec/globus-gridmap-and-execute -g > /etc/grid-security/grid-mapfile /opt/globus/libexec/globus-job-manager-sc > ript.pl -m fork -f /opt/globus/tmp/gram_job_mgr18847.tmp -c submit > 2007-10-02 10:16:52,686 DEBUG exec.JobManagerScript [Thread-14,run:225] > first line: GRAM_SCRIPT_JOB_ID:d4e0e07e-70bf-11dc-aba6-00151714069c:1045 > 2007-10-02 10:16:52,686 DEBUG exec.JobManagerScript [Thread-14,run:228] > Read line: GRAM_SCRIPT_JOB_ID:d4e0e07e-70bf-11dc-aba6-00151714069c:1045 > 2007-10-02 10:16:52,686 DEBUG exec.JobManagerScript [Thread-14,run:240] > Received local job ID d4e0e07e-70bf-11dc-aba6-00151714069c:1045 > 2007-10-02 10:16:52,686 DEBUG exec.JobManagerScript [Thread-14,run:228] > Read line: GRAM_SCRIPT_JOB_STATE:2 > 2007-10-02 10:16:52,687 DEBUG exec.JobManagerScript [Thread-14,run:335] > failure message: null > 2007-10-02 10:16:52,688 DEBUG exec.JobManagerScript > [Thread-14,setDone:345] script is done, setting done flag > 2007-10-02 10:16:52,688 DEBUG exec.StateMachine > [RunQueueThread_0,processSubmitState:1105] Done waiting for submit script > 2007-10-02 10:16:52,688 DEBUG exec.StateMachine > [RunQueueThread_0,processSubmitState:1129] script return code: 0 > 2007-10-02 10:16:52,689 DEBUG exec.StateMachine > [RunQueueThread_0,processSubmitState:1161] script returned job state: Active > 2007-10-02 10:16:52,690 DEBUG > ManagedJobResourceImpl.d39adff8-70bf-11dc-a4b2-080069149999 > [RunQueueThread_0,getResourceDatum:217] getting resource datum localJobId > 2007-10-02 10:16:52,690 DEBUG > ManagedJobResourceImpl.d39adff8-70bf-11dc-a4b2-080069149999 > [RunQueueThread_0,getResourceDatum:223] Obtaining lock on resourceData > 2007-10-02 10:16:52,690 DEBUG > ManagedJobResourceImpl.d39adff8-70bf-11dc-a4b2-080069149999 > [RunQueueThread_0,getResourceDatum:226] Obtained lock on resourceData > 2007-10-02 10:16:52,690 DEBUG > ManagedJobResourceImpl.d39adff8-70bf-11dc-a4b2-080069149999 > [RunQueueThread_0,getResourceDatum:266] Releasing lock on resourceData > > *I see the only difference is that in case of GT4.0.4 the symlink is not > resolved but sudo is working perfectly as you can see. Sorry for the > long mail. But now any tips for me. > > Cheers, > Samatha > > * > > -- Samatha Kottha Zentrum für Informationsdienste und Hochleistungsrechnen (ZIH) Technische Universität Dresden Tel: (+49) 351 463-38776 Room 1019 Fax: (+49) 351 463-38245 Noethnitzer Straße 46 01187 Dresden Germany
