They have to exist on system level (passwd/shadow/group files or LDAP, etc).


Am 07.05.2015 um 11:07 schrieb Trevor Gale:
> 
> If I want multiple users to be able to run jobs do I just specify them all in 
> slurm.conf under slurmuser? or do they just have to exist on both nodes with 
> the same UID and GID? I’ve disabled munge completely in my configuration so 
> the only credential check should be UID and GID correct?
> 
> Thanks,
> Trevor
> 
>> On May 7, 2015, at 10:24 AM, Uwe Sauter <[email protected]> wrote:
>>
>>
>> Using "slurm" as slurmuser is totally fine as long as that user exists on 
>> your compute nodes. BUT every user who should be able to
>> submit and run job also has to exist. It is up to you to give the users 
>> access to the compute nodes (e.g. shell entry in passwd file).
>>
>>
>>
>>
>> Am 07.05.2015 um 09:18 schrieb Trevor Gale:
>>>
>>> They do not exist, I used the default “slurm” name as the slurmuser. If I 
>>> create a slurm user would it then run using that accounts privileges? Does 
>>> the user have to exist on both nodes with sudo privileges? or do you think 
>>> this could be caused just by the fact that the user “slurm” doesn’t exist?
>>>
>>> Thanks,
>>> Trevor
>>>> On May 7, 2015, at 8:26 AM, Uwe Sauter <[email protected]> wrote:
>>>>
>>>>
>>>> If you configured Slurm to use Munge for authentication, you'll need to 
>>>> have Munge running…
>>>>
>>>> Please go to http://slurm.schedmd.com and read the quick starting and the 
>>>> troubleshooting guides.
>>>>
>>>>
>>>> Am 06.05.2015 um 22:17 schrieb Trevor Gale:
>>>>>
>>>>> I’ve disabled munge and now my two nodes see each other, but now I get an 
>>>>> error that says “task launch for job 352 failed: Invalid job credential”. 
>>>>> Any idea what might be causing this error? I’ve turned my SlurmdDebug up 
>>>>> to 7 but the log file essentially says the exact same thing as the stdout 
>>>>> when I try and submit a job.
>>>>>
>>>>> Thanks,
>>>>> Trevor
>>>>>
>>>>>> On May 6, 2015, at 9:54 PM, Uwe Sauter <[email protected]> wrote:
>>>>>>
>>>>>>
>>>>>> Configure your SLURM host as NTP server and your compute clients as NTP 
>>>>>> clients,
>>>>>>
>>>>>>
>>>>>> Am 06.05.2015 um 19:20 schrieb Trevor Gale:
>>>>>>> Ive made sure the keys are the same, how do you synchronize the 
>>>>>>> clocks/tell if they're out of synch?
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> On May 6, 2015, at 7:24 PM, Morris Jette <[email protected] 
>>>>>>> <mailto:[email protected]>> wrote:
>>>>>>>
>>>>>>>> Likely inconsistent keys or clocks.
>>>>>>>>
>>>>>>>> On May 6, 2015 4:20:36 PM PDT, Trevor Gale <[email protected] 
>>>>>>>> <mailto:[email protected]>> wrote:
>>>>>>>>
>>>>>>>>
>>>>>>>>  Do you happen to know what might be incorrect with munges 
>>>>>>>> configuration? 
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>      On May 6, 2015, at 6:55 PM, Moe Jette <[email protected] 
>>>>>>>> <mailto:[email protected]>> wrote:
>>>>>>>>
>>>>>>>>
>>>>>>>>      Quoting Trevor Gale <[email protected] 
>>>>>>>> <mailto:[email protected]>>:
>>>>>>>>
>>>>>>>>          Hello,
>>>>>>>>
>>>>>>>>          I am running slum between two desktop computers for testing 
>>>>>>>> purposes and am having a problem where when I launch
>>>>>>>>          jobs from the controlnode (rousseau) I get that the jobs have 
>>>>>>>> been queued and are waiting for resources and that the
>>>>>>>>          required node is not available (down or drained).
>>>>>>>>
>>>>>>>>
>>>>>>>>      The issue is in Munge's configuration, which Slurm user for 
>>>>>>>> authentication.
>>>>>>>>      -- 
>>>>>>>>      Morris "Moe" Jette
>>>>>>>>      CTO, SchedMD LLC
>>>>>>>>      Commercial Slurm Development and Support
>>>>>>>>
>>>>>>>>
>>>>>>>> -- 
>>>>>>>> Sent from my Android device with K-9 Mail. Please excuse my brevity. 

Reply via email to