[00:25:55] --- pod has left [02:06:44] --- pod has become available [02:31:19] --- Simon Wilkinson has become available [02:36:09] I suspect we should thing about doing a .1 release, if there's stuff we need for 2.6.34 that isn't in 1.4.12 [02:53:30] --- abo has left [03:31:33] --- shadow@gmail.com/owlB02D7AA7 has become available [03:34:40] --- abo has become available [03:35:16] so we're at some point going to need enhanced CellServDB so we can better specify multihomed, ipv6, etc, dbservers. while a discussion is obviously needed, i'm thinking openafs-info or openafs-devel rather than afs3-stds? like, others are free to implement it but it's not really something which fits with what we standardize. comments? [03:37:11] I'd agree. [03:38:08] I think the CellServDB is implementation specific. I think there's also a question about whether we should be supporting CellServDB at all, going forwards. Or whether all of these shiny new features should be implemented using SRV records. [03:39:12] the db servers themselves need a config file. does't have to be cellservdb but dns isn't it [03:39:50] e.g. "which dbserver is lowesthost" etc etc [03:42:51] Yeah. I think that's just local configuration, though. [03:43:03] (and yes, I agree that it's not DNS) [03:43:26] could be. the problem is there are cases i want to override dns on local hosts where i need a config file also, which probably onveys some of the same info, e.g. more than what we can convey now [03:45:19] I think we should perhaps come up with a new config file format, probably based around the INI file parser that's in gerrit. [03:45:52] It would be good, from a usability point of view, to more clearly separate the configuration that defines a cell from the server's point of view, from that which defines all of the cells that a client can see. [03:46:09] Having them both be CellServDB, but in different locations, just seems to cause confusion and woe for new users. [03:46:40] yeah, i suppose that's true [03:51:33] --- jaltman has left: Replaced by new connection [03:51:34] --- jaltman has become available [04:21:20] --- Simon Wilkinson has left [05:13:55] --- Simon Wilkinson has become available [05:52:45] --- haba has become available [06:02:36] --- kaj has left [06:15:01] --- meffie has become available [06:36:45] --- abo has left [06:41:22] --- abo has become available [06:49:59] --- jaltman has left: Disconnected [06:54:08] --- reuteras has left [06:54:45] --- haba has left [07:28:01] --- deason has become available [07:40:24] --- haba has become available [07:45:36] --- jaltman has become available [07:57:42] --- meffie has left [08:28:13] --- Simon Wilkinson has left [08:28:25] --- Simon Wilkinson has become available [08:33:18] --- jaltman has left: Disconnected [08:35:24] --- haba has left [08:36:00] --- haba has become available [08:49:58] --- jaltman has become available [09:03:47] > enhanced CellServDB I think -devel. And yeah; what we're really talking about is mostly ubik config, not for clients [09:33:39] --- haba has left [09:35:43] --- haba has become available [09:44:18] --- kaj has become available [10:02:59] --- jaltman has left: Disconnected [10:05:21] --- kaj has left [10:05:22] --- kaj has become available [10:10:20] --- haba has left [10:26:54] --- kaj has left [10:30:12] --- kaj has become available [10:44:01] --- matt has become available [11:30:55] --- Russ has become available [12:22:17] --- Simon Wilkinson has left [13:18:30] --- deason has left [13:20:24] --- deason has become available [14:28:11] --- jaltman has become available [15:32:19] --- mdionne has become available [15:35:29] --- deason has left [16:00:53] --- andersk@mit.edu/dr-wily is now known as andersk [16:39:13] --- jaltman has left: Disconnected [17:48:38] --- jaltman has become available [18:41:22] --- Jeffrey Altman has become available [19:17:05] --- jaltman has left: Disconnected [19:26:29] For a memcache cache manager, what would the value of 'afiles' default to? [19:30:20] --- mdionne has left [19:31:26] --- deason has become available [19:51:29] I appear to be smashing my stack somewhere in GetDCache ... avc is 0x28 ... this is annoying. [19:52:36] avc is a register. perhaps you might want to make it, well, not. it might help. [19:54:05] I thought gcc basically ignored the register keyword, these days. Or do you mean make debugging easier? [19:54:56] the latter [19:55:45] and i am going to run away, my body is being all gastric distress again [19:55:47] (The actual crash was trying to dereference 0x68 on line 1775: ReleaseReadLock(&tdc->tlock) But tdc is reasonable. [19:55:53] Bummer. [19:58:45] I have slightly different xen kernels in dom0 and domu--I'm suspicious that's not a great idea [19:59:21] Er, is that apropos of anything in particular, Matt? [19:59:32] no, sorry, mix [19:59:45] need to rest my eyes [20:13:29] --- matt has left [20:45:38] --- Born Fool has become available [21:20:41] --- Born Fool has left [22:01:20] --- reuteras has become available [22:32:09] --- deason has left [23:12:32] --- kaj has left [23:15:06] --- haba has become available [23:44:54] --- Russ has left: Disconnected [23:49:59] --- haba has left [23:55:28] --- haba has become available