[01:03:31] --- abo has left [01:03:35] --- abo has become available [01:21:27] --- kula has become available [01:44:18] --- Brandon Allbery has left [02:44:06] --- Brandon Allbery has become available [02:53:16] --- Simon Wilkinson has left [05:35:06] --- mvitale has become available [06:18:09] --- shadow@gmail.com/barnowl5ABA57FD has left [06:18:16] --- shadow@gmail.com/barnowl5ABA57FD has become available [06:22:45] --- ballbery has become available [06:51:06] --- mvitale has left [06:53:56] --- mvitale has become available [07:01:25] Okay. (I did read that once, I wonder if I still have it handy.) The securityIndex has the feature of being less hassle to get to in a client/server-independent way, too. [07:19:30] --- Brandon Allbery has left [07:19:41] --- Brandon Allbery has become available [07:28:15] --- Brandon Allbery has left [07:28:21] --- Brandon Allbery has become available [07:35:13] --- Simon Wilkinson has become available [07:42:06] --- deason has become available [09:02:45] --- Brandon Allbery has left [09:02:52] --- Brandon Allbery has become available [09:52:38] --- shadow@gmail.com/barnowl5ABA57FD has left [09:52:47] --- shadow@gmail.com/barnowl5ABA57FD has become available [10:26:38] --- Simon Wilkinson has left [11:40:26] --- Simon Wilkinson has become available [11:43:02] --- Simon Wilkinson has left [12:53:57] --- Simon Wilkinson has become available [14:38:07] --- ballbery has left [14:45:24] deason: the "POSIX_ME_HARDER OS" would (probably) not write data to disk on clean shutdown after fileserver crash. [14:47:22] where are you getting that from? what thing in posix? [14:47:53] Basically the only place "stable storage" appears in posix is in fsync() -- neither close nor write are sufficient. [14:48:12] so any such system would also lose data on open(), write(), close() [14:48:24] Presumably. [14:49:46] I don't think that's worth mentioning; anything like that would be rather... ephemeral, and you'd lost most of your afs data regardless of sync settings [14:49:53] er, and you'd lose [14:50:40] I'm not going to argue very hard about it; it's just that since we claim to support a wide spread of OSes, I'm wary of making assumptions about the kernel VM layer like this. [14:55:23] --- mvitale has left [15:24:11] --- Simon Wilkinson has left [17:11:59] --- mvitale has become available [17:24:59] deason: I'm pretty sure that FreeBSD ZFS has the same eventual-consistency guarantee that Solaris ZFS does, though I don't have a good reference to cite handy. [17:30:42] --- deason has left [18:49:43] --- jaltman/FrogsLeap has left: Disconnected [18:50:03] --- jaltman/FrogsLeap has become available [18:52:48] --- mvitale has left [18:53:05] --- mvitale has become available [18:53:14] --- mvitale has left [18:58:02] --- mvitale has become available [19:08:15] --- mvitale has left [19:10:45] --- mvitale has become available [19:27:24] --- mvitale has left [21:13:27] --- jaltman/FrogsLeap has left: Disconnected [21:22:15] --- jaltman/FrogsLeap has become available