[00:09:46] --- Russ has left: Disconnected [00:11:34] --- dev-zero@jabber.org has become available [00:18:35] --- jaltman has become available [00:20:54] --- abo has become available [00:22:24] --- dev-zero@jabber.org has left: Replaced by new connection [00:22:24] --- dev-zero@jabber.org has become available [00:24:21] --- dev-zero@jabber.org has left: Replaced by new connection [00:24:22] --- dev-zero@jabber.org has become available [00:53:49] --- dev-zero@jabber.org has left: Replaced by new connection [00:53:50] --- dev-zero@jabber.org has become available [01:08:27] --- Claudio Bisegni has become available [01:08:28] --- dev-zero@jabber.org has left: Replaced by new connection [01:08:28] --- dev-zero@jabber.org has become available [01:09:15] --- Claudio Bisegni has left [01:09:19] --- Claudio Bisegni has become available [01:09:44] --- Claudio Bisegni has left [01:09:46] --- Claudio Bisegni has become available [01:09:52] --- Claudio Bisegni has left [01:10:37] --- Claudio Bisegni has become available [01:12:17] --- dev-zero@jabber.org has left: Replaced by new connection [01:12:18] --- dev-zero@jabber.org has become available [01:29:10] --- dev-zero@jabber.org has left: Replaced by new connection [01:29:11] --- dev-zero@jabber.org has become available [01:50:59] --- dev-zero@jabber.org has left [02:27:05] --- Simon Wilkinson has left [02:59:03] --- Claudio Bisegni has left [03:44:39] --- jaltman has left: Disconnected [03:50:58] --- haba has become available [03:53:18] --- Simon Wilkinson has become available [04:01:59] --- Simon Wilkinson has left [06:04:49] --- dev-zero@jabber.org has become available [06:22:50] --- mmeffie has become available [06:33:20] --- dev-zero@jabber.org has left: Replaced by new connection [06:33:21] --- dev-zero@jabber.org has become available [06:41:25] --- dev-zero@jabber.org has left: Replaced by new connection [06:41:26] --- dev-zero@jabber.org has become available [06:47:36] --- dev-zero@jabber.org has left: Replaced by new connection [06:47:37] --- dev-zero@jabber.org has become available [06:52:18] --- deason has become available [08:42:47] --- haba has left [10:05:54] --- Russ has become available [10:25:56] --- jaltman has become available [10:26:54] --- stevenjenkins has left [10:26:55] --- stevenjenkins has become available [11:12:16] --- deason has left [11:12:43] --- deason has become available [11:23:19] --- jaltman has left: Disconnected [12:10:04] --- jaltman has become available [12:19:37] --- sxw mobile has become available [12:47:44] --- sxw mobile has left [12:49:18] --- deason has left [12:49:23] --- deason has become available [12:51:42] would it be out of line to delay creating the FSYNC thread until after volume attachment? [12:51:59] I get a little confused when an fsync request fails and tells me a volume doesn't exist, when it does exist but hasn't been (pre)attached yet [12:56:44] That's nonsensical in DAFS, since volumes don't get preattached. And IMHO it is too likely to have unforeseen effects to be done in 1.4.x [12:57:06] I mean after preattachment in dafs [12:57:17] in general, wait until VInit==2, instead of waiting for VInit==1 [12:57:35] and 1.5.x only [12:58:20] Oh, hm. At what point does the fileserver start answering RPC's? [13:00:41] after (pre)attachment, and after restoring state in dafs [13:00:48] It seems like there might be a circular dependency here, or at least a chance for unpleasant behavior, related to the salvage server needing the FSYNC interface to check out volumes to salvage them. [13:00:55] fiveminutecheck &co are started after then, too [13:01:02] But, I'm not an expert on that code. [13:01:42] Also, do we depend on binding the FSYNC socket as an exclusion mechanism to keep two fileservers from running at once? [13:01:58] oh, well, I can assure that it works in the trivial case, because I just tried doing it [13:02:01] hm, not sure about that [13:03:03] and, in order to have a fileserver-initiated salvage for anything, you need to be listening to RPCs [13:04:06] since we don't schedule a salvage until we actually try to do something with it, so I don't think there'd be a problem with trying to salvage [13:13:43] okay, no, binding to 7000 occurs before any volume package init, so that is what prevents multiple fileservers at least for now [13:14:24] hello, does anyone know the status of kolya's rx-spec. is going to be part of the afs-standardization process? [13:14:44] and I was slightly mistaken; for dafs we don't start answering rpcs until after preattachment, but for non-dafs we do start; at least at the rx_StartServer level [13:15:01] "for non-dafs we do start before" [13:16:46] kolya's rx-spec is what it is. nominally, afs3-stds standardizes on top of rx, not in it. i have a proposal for rx which i will propose to afs3-stds and a couple known-external users just in case, but there's no "proper" vessel. [13:22:13] hmm, ok. it is a very useful document (although out of date is some places if i recall) [13:22:39] i'm not arguing that in the slightest :) [13:26:36] out of curiosity, who are other users? [13:27:47] the ones i have in mind should be on afs3-stds and probably aren't, but there are certain;y non-AFS rx users: adm comes to mind [13:28:53] I think Rx is a suitable topic for standardization. But, like the RPC-level protocols, the main focus is on ongoing change; the existing protocols don't have to be standardized under that process before we can use, build on, or modify them. [13:29:15] stage also uses rx. [13:29:54] And I have an SSH public key database that uses it, though I want to retire that one of these days in favor of something using remctl. [13:30:18] rx is a fine topic for standardization. but not (necessarily) by an organization which exists to standardize AFS3 RPCs [13:30:37] (i have no better idea for an org to use, mind you) [13:31:06] my intention was to publish e.g. draft-brashear-rx- just as further afs3 drafts will be draft-person-afs3-whatever [13:33:15] do we have rights to use, take over maintanership of kolya's document? i suppose he wouldn't mind. [13:34:43] we could ask. pretty sure he even still reads zephyr [13:36:15] No, I think rx is an OK topic for afs3-stds. It certainly was designed for and is mostly used by AFS, and there's plenty of precedent for standards orgs taking responsibility for building blocks below their main work, even when those blocks are reusable elsewhere, if no one else is taking ownership. [13:38:04] it seems reasonable [14:17:24] later. [14:17:35] --- mmeffie has left [14:31:33] --- sxw mobile has become available [14:35:34] --- sxw mobile has left [14:38:33] --- sxw mobile has become available [15:04:33] --- dev-zero@jabber.org has left [15:13:36] --- deason has left [15:47:10] --- Simon Wilkinson has become available [16:06:52] --- sxw mobile has left [16:36:44] --- sxw mobile has become available [16:54:22] --- deason has become available [16:58:43] --- sxw mobile has left [17:26:29] --- Simon Wilkinson has left [17:27:32] --- sxw mobile has become available [17:41:54] --- dwbotsch has left [17:42:38] --- dwbotsch has become available [17:43:28] --- sxw mobile has left [17:54:20] --- kula has left [17:58:11] --- sxw mobile has become available [18:12:21] --- sxw mobile has left [18:13:15] --- sxw mobile has become available [18:28:22] --- sxw mobile has left [18:28:48] --- sxw mobile has become available [18:28:54] --- kula has become available [18:42:42] --- sxw mobile has left [18:57:39] --- sxw mobile has become available [19:13:17] --- sxw mobile has left [19:29:07] --- sxw mobile has become available [19:43:28] --- Jeffrey Altman has left: Replaced by new connection [19:44:09] --- sxw mobile has left [20:00:00] --- sxw mobile has become available [20:18:47] --- sxw mobile has left [20:21:44] --- sxw mobile has become available [20:38:02] --- sxw mobile has left [20:42:42] --- sxw mobile has become available [20:52:15] --- sxw mobile has left [21:23:40] --- sxw mobile has become available [21:39:03] --- sxw mobile has left [22:00:06] --- sxw mobile has become available [22:03:47] --- deason has left [22:04:14] --- deason has become available [22:07:21] --- sxw mobile has left [22:12:08] --- sxw mobile has become available [22:16:34] --- deason has left [22:23:22] --- sxw mobile has left [22:24:53] --- sxw mobile has become available [22:37:52] --- sxw mobile has left [22:52:40] --- sxw mobile has become available [23:08:22] --- sxw mobile has left [23:24:20] --- sxw mobile has become available [23:32:45] --- jaltman has left: Disconnected [23:47:16] --- sxw mobile has left [23:57:26] --- sxw mobile has become available [00:00:18] --- sxw mobile has left