[GE users] Re; {GE users] Stuck in pending

Reuti reuti at staff.uni-marburg.de
Tue Dec 6 22:20:22 GMT 2005


Hi,

Am 06.12.2005 um 22:44 schrieb markdenni at comcast.net:

> The version of SGE is: 6.0u6
> The output of qstat -j when sge_execd is not running on rs101  
> mentions that
> queue instance rs101... is temporarily not available. But when I  
> start it again,
> queue instance rs101 is not even mentioned. This is a mystery to me.
>
> In Host Configuration -> execution host  I have set Access to test  
> (of which I am the only member), XAccess to NONE, PROJECTS to NONE,  
> and XProjects to all the projects the other users use so their jobs  
> won't run on this server.)  Perhaps I am not doing this correctly.
>
> When rs101 is not running the output of qstat is:
>  % qstat -j 129791
> job_number:                 129791
> exec_file:                  job_scripts/129791
> submission_time:            Thu Dec  1 19:21:33 2005
> owner:                      markd
> uid:                        1605
> group:                      iosd
> gid:                        1000
> acco unt:                    sge
> mail_options:               n
> mail_list:                  markd at rs100.sjs.agilent.com
> notify:                     FALSE
> job_name:                   newtest
> priority:                   1
> jobshare:                   0
> script_file:                newtest
> verify_suitable_queues:     3
> version:                    7
> scheduling info:            queue instance  
> "all.q at rsl019.sjs.agilent.com" dropped because it is temporarily  
> not available
>                             queue instance  
> "all.q at rs101.sjs.agilent.com" dropped because it is temporarily not  
> available
>                          &n bsp;  queue instance  
> "all.q at rsl020.sjs.agilent.com" dropped because it is full
>                             queue instance  
> "all.q at rsl022.sjs.agilent.com" dropped because it is full
>                             has no permission for host "global"
>
> When sgeexecd is running on rs101, then it isn't even mentioned in  
> qstat -j:
>  % qstat -j 129791
> job_number:                 129791
> exec_file:                  job_scripts/129791
> submission_time:            Thu Dec  1 19:21:33 2005
> owner:                      markd
> uid:                        1605
> group:                      iosd
> gid:                        1000
> acco unt:                    sge
> mail_options:               n
> mail_list:                  markd at rs100.sjs.agilent.com
> notify:                     FALSE
> job_name:                   newtest
> priority:                   1
> jobshare:                   0
> script_file:                newtest
> verify_suitable_queues:     3
> version:                    7
> scheduling info:            queue instance  
> "all.q at rsl019.sjs.agilent.com" dropped because it is temporarily  
> not available
>                             queue instance  
> "all.q at rsl020.sjs.agilent.com" dropped because it is full
>                             queue in stance  
> "all.q at rsl022.sjs.agilent.com" dropped because it is full
>                             has no permission for host "global"
>
> Thanks in advance for any suggestions.

in the first case: if I understand you in the right, this is the  
output without sgeexec running - well, therefore it's temporarily not  
available.

in the second case: you get two outputs for the scheduling info. One  
is the overall scheduler information about certain queue instances on  
the hosts, the other output is for the job in question (which you see  
only for pending jobs). As the job is already running, you only get  
the first part. And maybe there are still slots left on rs101 and all  
is in best order. To me it looks fine.

Cheers - Reuti

>
> - Mark


---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
For additional commands, e-mail: users-help at gridengine.sunsource.net




More information about the gridengine-users mailing list