On Nov 14, 2007, at 10:17 PM, Jacki wrote: [...]
######################################## # dirvish-runall runs dirvish parallel # #########################################current state#dirvish-runall process all vaults like listed in master.conf under "runall:". this not fine, if you backup over a speed-limited and not so reliable connection. a "big" backup in the middle of the list blocks "small" backups at the end. especially if the connection is lost.#target state#all vaults within the "runall"-list are processed parallel, the maximum number of processes can be limited by an option in master.conf. for instance "runall-parallel: 3"with "runall-parallel: 1" its the same like beforewith "runall-parallel: 5" five vaults are backuped parallel at maximum, and if there are two "big" ones - no problem with "runall-parallel: -1" every "dirvish" command gets its own process#implementation# should be no problem with the perl "fork" and "wait" functions the "mail" generation could also be done - but sorting is an issue! "runall-parallel: 1" should stay default for backward compatibility
Hi Jacki,I submitted a patch for comments to the mailing list several months ago for this. If you'd like to check it out I'd appreciate any comments; I didn't get any comments on it before. The only suggestion in your above `target' that I didn't do in the patch was the `-1' parallel setting, though this would be easy to add.
http://www.dirvish.org/pipermail/dirvish/2007-September/001467.htmlAlso, since it seems you've thought about this a bit, I'm curious why you think this should use fork/wait -- this is the same method I used, but I'm curious about advantages/disadvantages to perl threads. My only reason for using fork/wait was because my distro didn't compile in threads support in perl by default, so I went with a method that would work best on my platform.
-- Eric Searcy OSU Open Source Lab
PGP.sig
Description: This is a digitally signed message part
_______________________________________________ Dirvish mailing list [email protected] http://www.dirvish.org/mailman/listinfo/dirvish
