move everything to new trunk directory
This is a very large commit from Rodrigo Lazo (rlazo) for Google SoC This is the latest patch on the job execution series. As far as I could test it, now scire should be able to create, distribute, execute and gather the output of a job with the new models. Also the jobs register their status on the database (pending, downloaded, running, failed, finished or cancelled). So now all the backend code needed to "watch" a job is on its place. Right now a job gets executed but dies on the post processing stage (at least that is what I understand from the error message) Just move the GACL_functions.py from server/ to server/modules and apply this patch and you'll how myltiple entries are created for each member of the group in jobs_clients and on job_history. I've take some precautions regarding concurrency adding some locks, I'm not sure they are enough so please comment on that
adding changes from Rodrigo Lazo (rlazo) for Google SoC: modified add_job.php to make use of the new scheduling logic, plus other minor changes
updates from Rodrigo Lazo for Google SoC: Fixed the backend and UI code so it can work with the latest modification to the db (client_status, jobs_status tables). Also, the job requesting mechanism uses the new run scheduling model (crontab-like).
changed the config_dir to /etc/scire Added some code from Rodrigo Lazo (rlazo) for Google SoC: Modified some function headers, added a "verbose" option to print a lot of messages while running. If this verbose option isn't good tell me so I remove it and just keep it for my working copy.
throwing in a temp hack until the DB supports return_output flag.
adding code to return output if the return_output flag from the scripts table is set to true. this will aid in the post-completion processing to follow.
Just removing some comments cruft.
some bugfixes and debugging added/uncommented. origjob is causing some issues b/c it doesn't exist in the DB schema yet. added semi-proper cleanup after job completion added line to repr(output) so xmlrpclib doesn't barf
Adding in very large changes from mdisney. Will include large commit msg in other commit.
added a debugging line and uncommented some debugging lines. info is good.
Lots of updates from Matt Disney (mdisney) see his svn for more detailed logs
cleaned up sys_info()
updates from mdisney: added a client function called sys_info() which basically populates a small dictionary with system information which will be passed to a method on the server to register/add the client. Then the system information will be added into the database along with the certificate information. added the beginnings of the code to add/register the client in scired.py
fix stupid mistake with data -> script_data
updating to the latest. very confusing changes here from last ver. job/script are now returned as dicts.
testing code in client
Updating the scire server and client programs with their current development copy. All changes from previous version have been made by Matt Disney as part of the Summer of Code. These are somewhat large updates involving database connectivity using python's ADOdb module and function calls between the server and client. Many thanks to Matt for the contribution.
small code cleanup
don't print cert digest in client
remove .svnignore files
add .svnignore files
"working" version of xmlrpc client module
update of new xmlrpc classes
initial commit of secure xmlrpc client module (btw, xmlrpclib and httplib suck)
generate cert/key when needed check server cert's digest against stored digest
bah, duft file
SSL-enabled server and basic client
|Powered by ViewVC 1.1.20|