Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

AsyncStageout should be a component #95

Closed
DMWMBot opened this issue Aug 11, 2010 · 8 comments
Closed

AsyncStageout should be a component #95

DMWMBot opened this issue Aug 11, 2010 · 8 comments

Comments

@DMWMBot
Copy link

DMWMBot commented Aug 11, 2010

The AsyncStageout component should be written in a WMAgent components style.

@DMWMBot
Copy link
Author

DMWMBot commented Aug 24, 2010

riahi: The component is written in WMAgent style and ready to be commited. I expect is fine to check the code in cvs. That is fine? Any constraints about the agent naming convention? The following is the workflow:

  1. FileTransferPoller queries the couchDB docs to get needed information
  2. It loops on users to call the ftscpWrapper for them. It gives to ftsWrapper processes (working in parallel): mapping of pfn sources and destinations, userProxy, destination site, list of destination pfn
  3. Each ftscpWapper process cleans the storage space removing the destination pfn got
  4. create ftscp copy jobs splitting by source the mapping of pfn sources and destination in files (since different sources should be in different copy jobs to be transferred using ftscp command)
  5. Call ftscp command for the copy job files created
  6. remove the pfn for succeeded transfers

The following is what I see as next steps :

  • Overload the mark_good method to remove succeeded transfers from couchDB (discuss couchDB fields to use)
  • Discuss from where the user proxy can be got

@drsm79
Copy link

drsm79 commented Aug 24, 2010

metson: Hi,
The code should go into SVN, can we do that next week after the MC code fest? Will that hold you up on other work? I'm not sure by what you mean by steps 2 and 4 - creating the input for ftscp was done via a view and list - have you changed this? I'd like to discuss that further before the code goes into version controll.

Can you open tickets up for the two issues you mention at the end of your comment?
Cheers
Simon

@DMWMBot
Copy link
Author

DMWMBot commented Aug 24, 2010

riahi: Hi,

For the commit I'll wait your green light since you are the boss :-)

For the step 2, the input for ftscp wrapper is done via view and list: mapping of pfn sources and destinations, destination site and list of destination pfn are got from view and list.

For the step 4, due to the fact that different sources can't use the same FTS server, for e.g:

source X and destination Y uses fts server Z
source T and destination Y can't use the fts server Z

That is why the jobs should be split into different copy job files sorted by source. So, each ftscpWrapper process will create needed copy jobs file before calling ftscp command.

I will open a ticket for the two isseus.

Cheers
Hassen

@DMWMBot
Copy link
Author

DMWMBot commented Sep 8, 2010

riahi: Patch to review and commit is enclosed to the ticket.

@drsm79
Copy link

drsm79 commented Sep 9, 2010

metson: (In 8891) Default config file to fix ticket 95

From: Hassen Riahi Hassen.Riahi@cern.ch

@drsm79
Copy link

drsm79 commented Sep 9, 2010

metson: (In 8892) Poller class to fix ticket 95

From: Hassen Riahi Hassen.Riahi@cern.ch

@drsm79
Copy link

drsm79 commented Sep 9, 2010

metson: (In 8893) Main component class to fix ticket 95

From: Hassen Riahi Hassen.Riahi@cern.ch

@drsm79
Copy link

drsm79 commented Sep 9, 2010

metson: (In 8894) Worker class to fix ticket 95

From: Hassen Riahi Hassen.Riahi@cern.ch

This issue was closed.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants