This used to work .
I now replaced the host certificate of my CE as it expired (I am using temporal)
As usual I installed the new one at /etc/.grid-security with the right permits . As usual : I isntalled the CE certificates (rpm) at the Wns and the UI and the WMS-CE
I rerun the yaim at the CE and all the WNs
When I submit directly from UI-> CE by glite-ce-job-submit the job is submitted but it is queued at the CE and does not reach any WN ( I have also rebooted the relevant machines)
The qstat at the CE shows W(wait) state - that is the submit is blocked there :
Job id Name User Time Use S Queue
------------------------- ---------------- --------------- -------- - -----
593.eladby-temp cream_682969244 kzvo001 0 W kzvo
The WN have nothing at the /home/kzvo001 ( as it should)
The UI machine shows (glite .. _status) :
****** JobID=[https://eladby-temp.haifa.il.ibm.com:8443/CREAM682969244]
Status = [IDLE]
Notice that during the CE yaim the Bnotifier have failed ( despte the reboot and prior : srevice gLite stop) . But in the past the submit worked despite this Failute !
Note: I can submit a job using 'qsub' from CE to a WN ! (but not from the UI as explained above)
In the CE Trustmanger.log I see nothing relevant .
No firewall exist . Paswordless login and reverse nslookup exist between the WNs and CE and UI
Any idea where to look for the error or what is missing ?
|