WMS 3.4.1 EMI2 precertification report
Testing server on: cream-51.pd.infn.it (SL5_64)
First installed and configured a EMI2 UI. Then installed update to glite-wms-ui-command (
glite-wms-ui-commands-3.4.0-5.sl5.x86_64.rpm
) from this repository:
http://etics-repository.cern.ch/repository/pm/registered/repomd/id/685c7797-b6c2-4587-964e-b3d479998aa2/sl5_x86_64_gcc412EPEL/index.html
built from WMS UI etics configuration emi-wms-ui_R_3_4_0_6.
Tried some simple job submission (without and with ISB)
dorigoa@cream-51 17:12:35 ~>rpm -qa | grep commands|grep wms
glite-wms-ui-commands-3.4.0-5.sl5
dorigoa@cream-51 17:12:41 ~>glite-wms-job-submit -a -e https://cream-01.pd.infn.it:7443/glite_wms_wmproxy_server sleep.jdl
Connecting to the service https://cream-01.pd.infn.it:7443/glite_wms_wmproxy_server
====================== glite-wms-job-submit Success ======================
The job has been successfully submitted to the WMProxy
Your job identifier is:
https://devel09.cnaf.infn.it:9000/VfRdyI099ZXzjrtGfE3QYA
==========================================================================
dorigoa@cream-51 17:12:43 ~>glite-wms-job-status https://devel09.cnaf.infn.it:9000/VfRdyI099ZXzjrtGfE3QYA
======================= glite-wms-job-status Success =====================
BOOKKEEPING INFORMATION:
Status info for the Job : https://devel09.cnaf.infn.it:9000/VfRdyI099ZXzjrtGfE3QYA
Current Status: Ready
Destination: lcgceatlas.dnp.fmph.uniba.sk:8443/cream-pbs-dteam
Submitted: Mon Jan 21 17:12:43 2013 CET
==========================================================================
dorigoa@cream-51 17:12:51 ~>cat sleep.jdl
[
Executable = "/bin/echo";
Arguments = "ciao";
MyProxyServer="";
]
dorigoa@cream-51 17:14:06 ~>cat wms_ISB.jdl
[
Executable = "/bin/sleep" ;
Arguments = "15";
#InputSandbox = {"/etc/group"};
InputSandbox = {"/home/dorigoa/mysandbox.txt", "/home/dorigoa/mysandbox2.txt"};
#InputSandboxBaseURI = "gsiftp://cream-23.pd.infn.it/var/";
AllowZippedISB = true;
requirements = RegExp("cream.*", other.GlueCEUniqueID);
RetryCount = 0;
ShallowRetryCount = 1;
]
dorigoa@cream-51 17:14:25 ~>glite-wms-job-submit -a -e https://cream-01.pd.infn.it:7443/glite_wms_wmproxy_server wms_ISB.jdl
Connecting to the service https://cream-01.pd.infn.it:7443/glite_wms_wmproxy_server
====================== glite-wms-job-submit Success ======================
The job has been successfully submitted to the WMProxy
Your job identifier is:
https://devel09.cnaf.infn.it:9000/BnYdX-_NytVUbMopFuG18g
==========================================================================
dorigoa@cream-51 17:14:37 ~>glite-wms-job-status https://devel09.cnaf.infn.it:9000/BnYdX-_NytVUbMopFuG18g
======================= glite-wms-job-status Success =====================
BOOKKEEPING INFORMATION:
Status info for the Job : https://devel09.cnaf.infn.it:9000/BnYdX-_NytVUbMopFuG18g
Current Status: Ready
Destination: ce203.cern.ch:8443/cream-lsf-grid_2nh_dteam
Submitted: Mon Jan 21 17:14:36 2013 CET
==========================================================================
LIST OF BUGS
UI
BUGS:
EMI-2 WMS UI cannot handle jobids starting with a '-'. (https://savannah.cern.ch/bugs/?98113
) PRE-CERTIFIED (adorigo 08/01/2013)
The bug has been fixed and verified. In order to check it, please take into account that there's a 2% probability to get the "-" as leading character of the gridjobid. By submitting (in a loop) about 1000 jobs, there's a 100% to get the at least 1 candidate job. The
JDL to use is:
[
Type = "Job";
JobType = "Normal";
Executable = "/bin/hostname";
StdOutput = "hello.out";
StdError = "hello.err";
InputSandbox = {"test_script_120.sh"};
OutputSandbox = {"hello.out","hello.err"};
RetryCount = 0;
ShallowRetryCount = 0;
Requirements =
other.GlueSiteUniqueID == "CERN-PROD" ||
other.GlueCEInfoHostName == "x";
]
Submit, in a loop, about 200-400 jobs with the above
JDL. Redirect the output to a logfile; when the candidate
GridJobID is generated the loop should stop and in the log file you can see a message like this:
terminate called after throwing an instance of 'boost::filesystem::filesystem_error'
what(): boost::filesystem::path: invalid name "-V" in path: "SandboxDir/-V/https_3a_2f_2fwms004.cnaf.infn.it_3a9000_2f-VZRnkQj2MopgqpsvdLjRA/input/group"
./test_98113.sh: line 6: 19300 Aborted glite-wms-job-submit -o GIDS -d pippo -e https://cream-28.pd.infn.it:7443/glite_wms_wmproxy_server my.jdl
This is a loop that should ensure to get the candidate
GridJobID:
#!/bin/bash
set -e
set i=0
for i in 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
do
for k in 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
do
glite-wms-job-submit -o GIDS -a -e https://<YOUR_PREFERRED_WMS_ENDPOINT>:7443/glite_wms_wmproxy_server <YOUR_JDL>.jdl
sleep 1
done
done
It is better to pre-delegate a proxy and use the delegation ID with "-d" option of glite-wms-job-submit instead of "-a" (automatica delegation mean a delegation per job, that would take a lot of time).
After installing new rpm of glite-wms-ui-commands (version 3_4_0_5) the abort shown above shoud disappear.
--
AlviseDorigo - 2013-01-09