Note that hypervisor/ right now is also used by masterd, for certain things (only class methods). But in general separating at node level is not a bad idea, and would have avoided the mcpu mistake (see Issue 149).
As for refactoring bdev we can go bdev/drbd or block/drbd (more similar to haskell). It's probably better than "just" noded/drbd as it becomes clearer especially as different ones get added. (eg. is noded/pinkbunny.py a block device or a network technology?) :) Thanks, Guido On Mon, Apr 15, 2013 at 9:05 AM, Helga Velroyen <[email protected]> wrote: > FYI: I am currently working on the storage reporting for file storage and > planned to extract the file storage related stuff from bdev to a separate > module. Let's coordinate a bit, how we should structure the code. > > > On Mon, Apr 15, 2013 at 8:57 AM, Helga Velroyen <[email protected]> wrote: >> >> +1 for a refactoring + daemon-level separation >> >> >> On Mon, Apr 15, 2013 at 7:02 AM, Michele Tartara <[email protected]> >> wrote: >>> >>> On Mon, Apr 15, 2013 at 12:55 PM, Bernardo Dal Seno <[email protected]> >>> wrote: >>>> >>>> I agree with Thomas. Separation at daemon level is good for most of >>>> the things. I mean, there are things that must be shared among >>>> daemons, e.g., configuration, but low-level code as the one dealing >>>> with block devices should not. That forces to clearly define >>>> interfaces (those are shared, of course), and I think it would be >>>> beneficial for the quality of the code. >>> >>> >>> Ok, then. I thought there was more shared code. Let's go for a >>> daemon-level separation. >>> >>> Michele >> >> > -- Guido Trotter Ganeti Engineering Google Germany GmbH Dienerstr. 12, 80331, München Registergericht und -nummer: Hamburg, HRB 86891 Sitz der Gesellschaft: Hamburg Geschäftsführer: Graham Law, Katherine Stephens Steuernummer: 48/725/00206 Umsatzsteueridentifikationsnummer: DE813741370
