On Mon, Feb 13, 2012 at 02:56:54PM +0400, Hrachya Astsatryan wrote:
> Dear all,
>
> The AM-05-YSU site supports ARMGRID, OPS, DTEAM VOs.
> When we try to submit a job under the OPS VO, we get an error related to the glexec.
> Could you, please, help us to overcome the problem.
> I send to you the following information for your consideration.
>
>
> Regards,
> Hrach
>
>
> 2012-02-13 07:41:01,310 FATAL - Authorization error: Failed to get the
> local user id via glexec
>
> *
> UI SUBMISSION ERROR:
> 2012-02-13 10:55:14,411 FATAL - Authorization error: Failed to get the local user id via glexec
>
Hi,
In your lcmaps logfile:
> LCMAPS 1: 2012-02-13.08:18:02-18239 : lcmaps_gridlist(): We have a gline.
> LCMAPS 1: 2012-02-13.08:18:02-18239 : lcmaps_gridlist(): userid found: .ops
> LCMAPS 1: 2012-02-13.08:18:02-18239 : lcmaps_gridlist(): usernameprefix: ops
> LCMAPS 1: 2012-02-13.08:18:02-18239 : lcmaps_gridlist(): newglobusidp: /C=AM/O=ArmeSFo/O=IIAP NAS RA/OU=HPC Laboratory/CN=Hrachya Astsatryanops
> LCMAPS 1: 2012-02-13.08:18:02-18239 : gridmapdir_userid: Found an existing lease: ops029
> LCMAPS 1: 2012-02-13.08:18:02-18239 : lcmaps_gridlist(): final userid found: ops029
> LCMAPS 1: 2012-02-13.08:18:02-18239 : lcmaps_plugin_voms_poolaccount-plugin_run(): found username: ops029
> LCMAPS 0: 2012-02-13.08:18:02-18239 : lcmaps_plugin_voms_poolaccount-plugin_run(): no valid user account found with the name "ops029"
The same message is repeated many times.
It means that the pool account "ops029" does not exist on the cream CE,
but there is a mapping between your DN and ops029 in
/etc/grid-security/gridmapdir.
I connected with uberftp to ce.ysu-cluster2.grid.am using the dteam VO,
and indeed in your /etc/passwd there is no 'ops029'. The ops group has
only the accounts ops001,ops002,...,ops025 and ops040.
I have read in a previous thread that you had a similar problem with the
armgrid accounts. I don't know how you ended up like this but there
might be more inconsistencies between gridmapdir and /etc/passwd.
I see two options:
- You try to find all inconsistencies and resolve them
by hand (https://wiki.egi.eu/wiki/Tools/Manuals/TS03 might help)
- A more drastic solution would be to remove *everything* that is inside
/etc/grid-securitu/gridmapdir/ and rerun YAIM. However this would erase
all current mappings. So you should do it only after disabling job
submission and only if there are no jobs running on the CE.
Cheers,
Kyriakos
|