[GE users] user permission troubles when submitting with qsub

Shane Dixon shane.dixon at atmel.com
Fri Dec 19 22:29:17 GMT 2008


    [ The following text is in the "UTF-8" character set. ]
    [ Your display is set for the "ISO-8859-10" character set.  ]
    [ Some characters may be displayed incorrectly. ]

I'm having trouble with an error message regarding permissions.  I have
an sgeadmin user (uid 4439) that runs the sgemaster on the master
server.  Whenever I submit a job to it as anyone, it tries to create the
simple.sh.o18 file and then it fails because of permissions.  The .o18
file goes into the user's home directory.  This happens with all users.
It looks like it's trying to create the file as sgeadmin instead of as
the user even though the user submitted it.

I'm not quite sure what's the "proper" way to resolve this?  I tried
running the sgemaster as root, but the shepherd process still gets
called as sgeadmin.  How do I work around this?

--
Shane

Job 23 caused action: Job 23 set to ERROR
 User        = bsmith
 Queue       = test at server
 Host        = server
 Start Time  = <unknown>
 End Time    = <unknown>
failed opening input/output file:12/19/2008 14:42:58 [4439:16105]:
error: can't open output file "/home/bsmith/simple.sh.o23": Permi
Shepherd trace:
12/19/2008 14:42:58 [4439:16103]: shepherd called with uid = 4439, euid
= 4439
12/19/2008 14:42:58 [4439:16103]: starting up 6.1u5
12/19/2008 14:42:58 [4439:16103]: warning: starting not as superuser
(uid=4439)
12/19/2008 14:42:58 [4439:16103]: setpgid(16103, 16103) returned 0
12/19/2008 14:42:58 [4439:16103]: no prolog script to start
12/19/2008 14:42:58 [4439:16105]: pid=16105 pgrp=16105 sid=16105 old
pgrp=16103 getlogin()=<no login set>
12/19/2008 14:42:58 [4439:16105]: reading passwd information for user
'bsmith'
12/19/2008 14:42:58 [4439:16103]: forked "job" with pid 16105
12/19/2008 14:42:58 [4439:16105]: setosjobid: uid = 4439, euid = 4439
12/19/2008 14:42:58 [4439:16103]: child: job - pid: 16105
12/19/2008 14:42:58 [4439:16105]: setting limits
12/19/2008 14:42:58 [4439:16105]: RLIMIT_CPU setting: (soft 0^HINFINITY
hard 0^HINFINITY) resulting: (soft 0^HINFINITY hard 0^HINFINITY)
12/19/2008 14:42:58 [4439:16105]: RLIMIT_FSIZE setting: (soft 0^HINFINITY
hard 0^HINFINITY) resulting: (soft 0^HINFINITY hard 0^HINFINITY)
12/19/2008 14:42:58 [4439:16105]: RLIMIT_DATA setting: (soft 0^HINFINITY
hard 0^HINFINITY) resulting: (soft 0^HINFINITY hard 0^HINFINITY)
12/19/2008 14:42:58 [4439:16105]: RLIMIT_STACK setting: (soft 0^HINFINITY
hard 0^HINFINITY) resulting: (soft 0^HINFINITY hard 0^HINFINITY)
12/19/2008 14:42:58 [4439:16105]: RLIMIT_CORE setting: (soft 0^HINFINITY
hard 0^HINFINITY) resulting: (soft 0^HINFINITY hard 0^HINFINITY)
12/19/2008 14:42:58 [4439:16105]: RLIMIT_VMEM/RLIMIT_AS setting: (soft
0^HINFINITY hard 0^HINFINITY) resulting: (soft 0^HINFINITY hard 0^HINFINITY)
12/19/2008 14:42:58 [4439:16105]: RLIMIT_RSS setting: (soft 0^HINFINITY
hard 0^HINFINITY) resulting: (soft 0^HINFINITY hard 0^HINFINITY)
12/19/2008 14:42:58 [4439:16105]: setting environment
12/19/2008 14:42:58 [4439:16105]: Initializing error file
12/19/2008 14:42:58 [4439:16105]: switching to intermediate/target user
12/19/2008 14:42:58 [4439:16105]: tried to change uid/gid without being
root
12/19/2008 14:42:58 [4439:16105]: try running further with uid=4439
12/19/2008 14:42:58 [4439:16105]: closing all filedescriptors
12/19/2008 14:42:58 [4439:16105]: further messages are in "error" and
"trace"
12/19/2008 14:42:58 [4439:16105]: error: can't open output file
"/home/bsmith/simple.sh.o23": Permission denied
12/19/2008 14:42:58 [4439:16103]: wait3 returned 16105 (status: 6656;
WIFSIGNALED: 0,  WIFEXITED: 1, WEXITSTATUS: 26)
12/19/2008 14:42:58 [4439:16103]: job exited with exit status 26
12/19/2008 14:42:58 [4439:16103]: reaped "job" with pid 16105
12/19/2008 14:42:58 [4439:16103]: job exited not due to signal
12/19/2008 14:42:58 [4439:16103]: job exited with status 26
12/19/2008 14:42:58 [4439:16103]: now sending signal KILL to pid -16105
12/19/2008 14:42:58 [4439:16103]: no tasker to notify
12/19/2008 14:42:58 [4439:16103]: failed starting job
12/19/2008 14:42:58 [4439:16103]: no epilog script to start

Shepherd error:
12/19/2008 14:42:58 [4439:16105]: error: can't open output file
"/home/bsmith/simple.sh.o23": Permission denied

Shepherd pe_hostfile:
server 1 test at server <NULL>

------------------------------------------------------
http://gridengine.sunsource.net/ds/viewMessage.do?dsForumId=38&dsMessageId=93447

To unsubscribe from this discussion, e-mail: [users-unsubscribe at gridengine.sunsource.net].



More information about the gridengine-users mailing list