[GE issues] [Issue 3126] New - Scheduler crashes with got NULL element for CE_requestable

ravinallan ravichandra.nallan at sun.com
Fri Sep 4 16:18:24 BST 2009


http://gridengine.sunsource.net/issues/show_bug.cgi?id=3126
                 Issue #|3126
                 Summary|Scheduler crashes with got NULL element for CE_request
                        |able
               Component|gridengine
                 Version|6.1
                Platform|All
                     URL|
              OS/Version|All
                  Status|NEW
       Status whiteboard|
                Keywords|
              Resolution|
              Issue type|DEFECT
                Priority|P3
            Subcomponent|kernel
             Assigned to|ravinallan
             Reported by|ravinallan






------- Additional comments from ravinallan at sunsource.net Fri Sep  4 08:18:19 -0700 2009 -------
The sge scheduler crashes in the following scenario:

 0) create a sample complex, say memctrl
maintrunk $  qconf -sc | grep memctrl
memctrl        mc         MEMORY      >=    YES         NO         0        0

 1) add a simple rq
maintrunk $  qconf -srqs
{
   name         test
   description  NONE
   enabled      TRUE
   limit        to memctrl=0
}

 2) now DELETE the complex:
maintrunk $  qconf -mc
....
"/var/tmp/pid-12770-sraq8y" 50 lines, 4156 characters
sgetest at vx86-v65xk-blr03 removed "memctrl" from complex list

 3) and submit any job,
maintrunk $  qsub examples/jobs/sleeper.sh
maintrunk $  qstat
error: commlib error: can't connect to service (Connection refused)
error: unable to contact qmaster using port 6444 on host "vx86-v65xk-blr03"

from the qmaster running:
critical error: !!!!!!!!!! got NULL element for CE_requestable !!!!!!!!!!
Abort

Interestingly restart of qmaster logs this:
maintrunk $  /export/home/sgetest/test1/maintrunk/bin/sol-x86/sge_qmaster
error: "resource_quota" references not existing complex attribute "memctrl" <------
read job database with 0 entries in 0 seconds
qmaster hard descriptor limit is set to 65536
qmaster soft descriptor limit is set to 65536
qmaster will use max. 65516 file descriptors for communication
qmaster will accept max. 99 dynamic event clients
starting up GE 6.2u3 (sol-x86)

------------------------------------------------------
http://gridengine.sunsource.net/ds/viewMessage.do?dsForumId=36&dsMessageId=215769

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



More information about the gridengine-users mailing list