Discussion:
[gt-user] GRAM Job submission failed because the job manager failed to create an internal script argument file
Ben Bimber
2014-10-17 20:19:22 UTC
Permalink
Hello,

I am installing globus/gram 6.0 on ubuntu, following these instructions:

http://toolkit.globus.org/toolkit/docs/6.0/admin/quickstart/#q-first

I am trying to run the final command:

globus-job-run labkeyremotepipeline /bin/hostname

to test GRAM. I get this error: "GRAM Job submission failed because the
job manager failed to create an internal script argument file (error code
22)". I have read that this means I need to create a home directory;
however, one exists for my user. Can anyone please suggest other
debugging steps I can take to figure out more specifically what is going
wrong?

I am running this command as the user 'quser', created using adduser. This
user has a home directory (/home/quser). I'm not sure what other user
would be involved here. The globus-gatekeeper service is running as root.
I dont know if jobs submitted are somehow mapped to a different user, which
might be missing a home directory.

Sorry for the basic question and thank you in advance for any help.

-Ben
Ben Bimber
2014-10-17 22:50:16 UTC
Permalink
Hello,

In case it helps, here is the output from globus-gatekeeper.log. I *think*
it is saying I authenticated as globus user "Vagrant User", which maps to
local user 'vagrant'. That's what I would expect. The local user vagrant
has a standard home directory, which is why I find it surprising I'm
getting the error code 22 message.

Thanks again for any help or debugging ideas. Log below:


TIME: Fri Oct 17 22:44:14 2014
PID: 15493 -- Notice: 6: /usr/sbin/globus-gatekeeper pid=15493 starting at
Fri Oct 17 22:44:14 2014

TIME: Fri Oct 17 22:44:14 2014
PID: 15493 -- Notice: 6: GRAM contact:
labkeyRemotePipeline:2119:/O=Grid/OU=GlobusTest/OU=simpleCA-labkeyremotepipeline/CN=labkeyRemotePipeline

TIME: Fri Oct 17 22:44:14 2014
PID: 15493 -- Notice: 0: GATEKEEPER_ACCT_FD=6
(/var/log/globus-gatekeeper.log)
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 6: Got connection ::ffff:127.0.0.1 at Fri Oct 17
22:44:31 2014

TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 5: Authenticated globus user:
/O=Grid/OU=GlobusTest/OU=simpleCA-labkeyremotepipeline/OU=local/CN=Vagrant
User
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 0: GATEKEEPER_JM_ID
2014-10-17.22:44:31.0000015494.0000000001 for
/O=Grid/OU=GlobusTest/OU=simpleCA-labkeyremotepipeline/OU=local/CN=Vagrant
User on ::ffff:127.0.0.1
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 5: Requested service: jobmanager
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 5: Authorized as local user: vagrant
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 5: Authorized as local uid: 1000
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 5: and local gid: 1000
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 0: executing /usr/sbin/globus-job-manager
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 0: GRID_SECURITY_HTTP_BODY_FD=11
TIME: Fri Oct 17 22:44:31 2014
PID: 15624 -- Notice: 0: Set CONTENT_LENGTH=353
TIME: Fri Oct 17 22:44:31 2014
PID: 15624 -- Notice: 0: Set GATEWAY_INTERFACE to CGI/1.1
TIME: Fri Oct 17 22:44:31 2014
PID: 15624 -- Notice: 0: Set SERVER_NAME to labkeyremotepipeline
TIME: Fri Oct 17 22:44:31 2014
PID: 15624 -- Notice: 0: Set SERVER_PORT to 2119
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 0: Read 148 bytes from proxy pipe
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 0: Child 15624 started
Post by Ben Bimber
Hello,
http://toolkit.globus.org/toolkit/docs/6.0/admin/quickstart/#q-first
globus-job-run labkeyremotepipeline /bin/hostname
to test GRAM. I get this error: "GRAM Job submission failed because the
job manager failed to create an internal script argument file (error code
22)". I have read that this means I need to create a home directory;
however, one exists for my user. Can anyone please suggest other
debugging steps I can take to figure out more specifically what is going
wrong?
I am running this command as the user 'quser', created using adduser.
This user has a home directory (/home/quser). I'm not sure what other user
would be involved here. The globus-gatekeeper service is running as root.
I dont know if jobs submitted are somehow mapped to a different user, which
might be missing a home directory.
Sorry for the basic question and thank you in advance for any help.
-Ben
Joseph Bester
2014-10-20 15:27:21 UTC
Permalink
There are a couple issues with the debian packages that I'm debugging right now. I'll post here when update packages are available.
Post by Ben Bimber
Hello,
In case it helps, here is the output from globus-gatekeeper.log. I *think* it is saying I authenticated as globus user "Vagrant User", which maps to local user 'vagrant'. That's what I would expect. The local user vagrant has a standard home directory, which is why I find it surprising I'm getting the error code 22 message.
TIME: Fri Oct 17 22:44:14 2014
PID: 15493 -- Notice: 6: /usr/sbin/globus-gatekeeper pid=15493 starting at Fri Oct 17 22:44:14 2014
TIME: Fri Oct 17 22:44:14 2014
PID: 15493 -- Notice: 6: GRAM contact: labkeyRemotePipeline:2119:/O=Grid/OU=GlobusTest/OU=simpleCA-labkeyremotepipeline/CN=labkeyRemotePipeline
TIME: Fri Oct 17 22:44:14 2014
PID: 15493 -- Notice: 0: GATEKEEPER_ACCT_FD=6 (/var/log/globus-gatekeeper.log)
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 6: Got connection ::ffff:127.0.0.1 at Fri Oct 17 22:44:31 2014
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 5: Authenticated globus user: /O=Grid/OU=GlobusTest/OU=simpleCA-labkeyremotepipeline/OU=local/CN=Vagrant User
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 0: GATEKEEPER_JM_ID 2014-10-17.22:44:31.0000015494.0000000001 for /O=Grid/OU=GlobusTest/OU=simpleCA-labkeyremotepipeline/OU=local/CN=Vagrant User on ::ffff:127.0.0.1
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 5: Requested service: jobmanager
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 5: Authorized as local user: vagrant
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 5: Authorized as local uid: 1000
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 5: and local gid: 1000
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 0: executing /usr/sbin/globus-job-manager
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 0: GRID_SECURITY_HTTP_BODY_FD=11
TIME: Fri Oct 17 22:44:31 2014
PID: 15624 -- Notice: 0: Set CONTENT_LENGTH=353
TIME: Fri Oct 17 22:44:31 2014
PID: 15624 -- Notice: 0: Set GATEWAY_INTERFACE to CGI/1.1
TIME: Fri Oct 17 22:44:31 2014
PID: 15624 -- Notice: 0: Set SERVER_NAME to labkeyremotepipeline
TIME: Fri Oct 17 22:44:31 2014
PID: 15624 -- Notice: 0: Set SERVER_PORT to 2119
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 0: Read 148 bytes from proxy pipe
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 0: Child 15624 started
Hello,
http://toolkit.globus.org/toolkit/docs/6.0/admin/quickstart/#q-first
globus-job-run labkeyremotepipeline /bin/hostname
to test GRAM. I get this error: "GRAM Job submission failed because the job manager failed to create an internal script argument file (error code 22)". I have read that this means I need to create a home directory; however, one exists for my user. Can anyone please suggest other debugging steps I can take to figure out more specifically what is going wrong?
I am running this command as the user 'quser', created using adduser. This user has a home directory (/home/quser). I'm not sure what other user would be involved here. The globus-gatekeeper service is running as root. I dont know if jobs submitted are somehow mapped to a different user, which might be missing a home directory.
Sorry for the basic question and thank you in advance for any help.
-Ben
Ben Bimber
2014-10-20 15:35:51 UTC
Permalink
Thank you. If I can provide any information / test anything that would
help please let me know.
Post by Joseph Bester
There are a couple issues with the debian packages that I'm debugging
right now. I'll post here when update packages are available.
Post by Ben Bimber
Hello,
In case it helps, here is the output from globus-gatekeeper.log. I
*think* it is saying I authenticated as globus user "Vagrant User", which
maps to local user 'vagrant'. That's what I would expect. The local user
vagrant has a standard home directory, which is why I find it surprising
I'm getting the error code 22 message.
Post by Ben Bimber
TIME: Fri Oct 17 22:44:14 2014
PID: 15493 -- Notice: 6: /usr/sbin/globus-gatekeeper pid=15493 starting
at Fri Oct 17 22:44:14 2014
Post by Ben Bimber
TIME: Fri Oct 17 22:44:14 2014
labkeyRemotePipeline:2119:/O=Grid/OU=GlobusTest/OU=simpleCA-labkeyremotepipeline/CN=labkeyRemotePipeline
Post by Ben Bimber
TIME: Fri Oct 17 22:44:14 2014
PID: 15493 -- Notice: 0: GATEKEEPER_ACCT_FD=6
(/var/log/globus-gatekeeper.log)
Post by Ben Bimber
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 6: Got connection ::ffff:127.0.0.1 at Fri Oct 17
22:44:31 2014
Post by Ben Bimber
TIME: Fri Oct 17 22:44:31 2014
/O=Grid/OU=GlobusTest/OU=simpleCA-labkeyremotepipeline/OU=local/CN=Vagrant
User
Post by Ben Bimber
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 0: GATEKEEPER_JM_ID
2014-10-17.22:44:31.0000015494.0000000001 for
/O=Grid/OU=GlobusTest/OU=simpleCA-labkeyremotepipeline/OU=local/CN=Vagrant
User on ::ffff:127.0.0.1
Post by Ben Bimber
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 5: Requested service: jobmanager
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 5: Authorized as local user: vagrant
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 5: Authorized as local uid: 1000
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 5: and local gid: 1000
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 0: executing /usr/sbin/globus-job-manager
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 0: GRID_SECURITY_HTTP_BODY_FD=11
TIME: Fri Oct 17 22:44:31 2014
PID: 15624 -- Notice: 0: Set CONTENT_LENGTH=353
TIME: Fri Oct 17 22:44:31 2014
PID: 15624 -- Notice: 0: Set GATEWAY_INTERFACE to CGI/1.1
TIME: Fri Oct 17 22:44:31 2014
PID: 15624 -- Notice: 0: Set SERVER_NAME to labkeyremotepipeline
TIME: Fri Oct 17 22:44:31 2014
PID: 15624 -- Notice: 0: Set SERVER_PORT to 2119
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 0: Read 148 bytes from proxy pipe
TIME: Fri Oct 17 22:44:31 2014
PID: 15623 -- Notice: 0: Child 15624 started
Hello,
http://toolkit.globus.org/toolkit/docs/6.0/admin/quickstart/#q-first
globus-job-run labkeyremotepipeline /bin/hostname
to test GRAM. I get this error: "GRAM Job submission failed because the
job manager failed to create an internal script argument file (error code
22)". I have read that this means I need to create a home directory;
however, one exists for my user. Can anyone please suggest other
debugging steps I can take to figure out more specifically what is going
wrong?
Post by Ben Bimber
I am running this command as the user 'quser', created using adduser.
This user has a home directory (/home/quser). I'm not sure what other user
would be involved here. The globus-gatekeeper service is running as root.
I dont know if jobs submitted are somehow mapped to a different user, which
might be missing a home directory.
Post by Ben Bimber
Sorry for the basic question and thank you in advance for any help.
-Ben
Loading...