2008-08-22 00:49 folks 2008-08-22 00:50 flips: or put the metadata delta in a contiguous block run 2008-08-22 00:56 delayed allocation is your best friend, the allocator is going to be a pain in the ass 2008-08-22 01:59 true that 2008-08-22 02:01 bh, I don't get your comment about the metadata delta 2008-08-22 02:02 hey 2008-08-22 02:02 oh maybe I'm being stupid 2008-08-22 02:03 like changes in the b-tree itself might benefit from being contiguous 2008-08-22 02:03 you mean for replication? 2008-08-22 02:04 the principle is simple: changes in metadata do not mean a thing 2008-08-22 02:04 for dealing with changes to the b-tree itself, maybe make a distinction in how metadata is written versus data 2008-08-22 02:04 it is only changes in the logical data that have to be replicated 2008-08-22 02:04 the packing for that is known so it might benefit from a special treatment of that case 2008-08-22 02:04 oh, ok, like that log things you wrote up about ? 2008-08-22 02:05 log thing? 2008-08-22 02:05 that is a way of doing atomic commit 2008-08-22 02:05 yes, it matters 2008-08-22 02:05 because the log can contain part of the logical data 2008-08-22 02:05 if it has not been rolled up into the "real" fs structure yet 2008-08-22 02:25 -!- pgquiles(~pgquiles@239.Red-83-41-113.dynamicIP.rima-tde.net) has joined #tux3 2008-08-22 03:33 ACTION is about to head to bed 2008-08-22 08:22 -!- pgquiles(~pgquiles@239.Red-83-41-113.dynamicIP.rima-tde.net) has joined #tux3 2008-08-22 10:25 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #tux3 2008-08-22 10:32 -!- pgquiles(~pgquiles@239.Red-83-41-113.dynamicIP.rima-tde.net) has joined #tux3 2008-08-22 10:33 -!- konrad(~konrad@c-24-16-77-169.hsd1.wa.comcast.net) has joined #tux3 2008-08-22 11:24 -!- MaZe(~MaZe@216-239-45-4.google.com) has joined #tux3 2008-08-22 11:29 ah 2008-08-22 11:30 got through a big jogjam, inode creation is now in a plausible from 2008-08-22 11:30 going to do a call for testing pretty soon 2008-08-22 11:30 ACTION thinks about maze 2008-08-22 11:31 or more properly, a call for reality check on basic algorithms 2008-08-22 11:31 testing/debugging is too easy for the great minds on this channel ;-) 2008-08-22 11:53 ;-) 2008-08-22 12:15 hey, there's a guy who hasn't joined our tux3 LinkedIn List 2008-08-22 12:16 got to engage buttgears 2008-08-22 12:16 http://www.linkedin.com/e/gis/154012 2008-08-22 12:16 buttgears 2008-08-22 12:16 that's a new one to me 2008-08-22 12:17 is that like toosh_drive? 2008-08-22 12:17 geekified version of got to get ass in gear 2008-08-22 12:38 hey 2008-08-22 12:38 hi bh 2008-08-22 22:33 -!- konrad(~konrad@c-24-16-74-109.hsd1.wa.comcast.net) has joined #tux3