[00:06:44] --- stevegt has become available [00:47:05] --- manfred furuholmen has left [00:52:39] --- RedZBear has left [00:55:01] --- dwbotsch has become available [00:57:16] --- dwbotsch has left [00:59:45] --- dwbotsch has become available [01:07:34] --- dev-zero@jabber.org has left [01:56:36] --- dev-zero@jabber.org has become available [02:01:00] --- manfred furuholmen has become available [03:43:10] --- dev-zero@jabber.org has left [04:52:35] --- dev-zero@jabber.org has become available [06:55:43] --- dmontuori has become available [07:53:19] --- Simon Wilkinson has left [07:55:23] --- Simon Wilkinson has become available [08:00:58] --- dev-zero@jabber.org has left [08:02:06] --- dev-zero@jabber.org has become available [08:13:36] shadow? [08:23:24] --- SecureEndpoints has become available [08:24:51] hi [08:57:33] --- Moose has become available [08:59:42] hi moose [09:20:23] --- dev-zero@jabber.org has left [09:30:28] I'm now in a position to be able to sign the OpenAFS builds that come from lochranza. [09:30:42] They're going to be signed with C19E4A0A - "OpenAFS builds from lochranza" [09:31:13] Hopefully that's clear enough that it only indicates build integrity, and doesn't offer any personal guarantees. [09:40:40] --- manfred furuholmen has left [09:41:59] --- reuteras has left [09:45:22] its good enough for me [09:48:09] I've just spent a few hours automating all of the build stuff. There's now a progress report available from http://lochranza.inf.ed.ac.uk/ [09:49:29] notes somewhere on how you're doing the builds? [09:49:56] src/packaging/RedHat/mockbuild.pl [09:50:04] tx. [09:51:45] http://blob.inf.ed.ac.uk/sxw/2008/04/23/building-an-openafs-release-for-redhat-linux/ has, essentially, the steps involved in making a release [09:52:17] that's what I was looking for, tx. [09:52:25] btw, do you really want mockbuild in python? [09:52:37] .pl is perl [09:52:49] I know a guy who is interested in AFS who likes python. that's probably the kind of thing he would enjoy [09:52:55] There's some python embedded in there. [09:53:05] secureendpoints: true, but there is some python in there to deal with yum. [09:53:08] Essentially, we have to drive yum in ways that it won't let us do from the command line. [09:53:15] ok [09:53:24] and a comment that says: # Words cannot describe how gross this is. Yum no longer provides usable # output, so we need to call the python interface. At some point this # probably means this script should be entirely rewritten in python, # but this is not that point. [09:53:26] and there is no perl module? [09:53:35] No perl module :( [09:54:01] well, that will be my reason to drink today [09:54:20] I suspect there would be a general maintainability issue if it were in python. I think most folk in these parts know perl, but don't know much python. [09:54:21] speaking of which, time to go drink [09:54:24] you need a reason to drink? [09:54:36] being of age is gonna be boring :) [09:54:47] summatusmentis: you learn fast ;-) [09:54:51] bye [09:55:03] simon - np. but if you want to go the python route, let me know [10:11:26] --- stevegt has left: Lost connection [10:14:54] --- haba has become available [10:16:32] Good everyone. Here the weekend has begun. [10:17:25] The weekend is trying to begin here. But I'm resisting. [10:18:21] weekend? the weekend doesn't start until I'm done w/ classes (3+ hours) [10:19:06] I think you'll find that there are a lot of people in parts east of you that my disagree. [10:19:16] summatusmentis: that's what you get from living in the east ;) [10:19:45] or west depending how you look at it :) :) [10:21:33] However the AFS question of the day is how long vldb volume info (this volume can be found there) takes to time out. Even a day does not seem to be long enough. Needed to run fs checkv everywhere. That is anoying. [10:21:46] it is not the weekend here, but my brain is dead [10:22:03] that ... sounds like a bug with the cachemgr [10:22:20] could your cache be too big? [10:22:42] (file server down 2 days, volume restored on other server after one day, client still in the dark) [10:23:04] huh [10:23:24] that definitely should have timed out, from what I know (though i don't have to remind you, i am not a coder) [10:23:30] The cache is probably big enough so that the info does not get thrown out for "full" reason. [10:24:07] iirc there is a timeout but i'll be damned if i can remember what it is. lemme see if i can figure it out. [10:25:14] (client is Linux and 1.4.0 to 1.4.6-ish - makes no difference) [10:26:15] Original server is presumably still down, so the client won't be getting back 'volume moved' messages from that? [10:26:21] OH [10:26:29] OH OH OH this is THAT bug [10:26:32] Original server is DEAD [10:26:52] Was that bug fixed recently or is it still around? [10:26:56] cache doesn't recognize the vol as moved [10:26:59] recently fixed [10:27:02] in 1.4.8 iirc [10:27:21] it just thinks it's busy, so it refuses to do the update [10:28:22] Thanks Moose. But starting an empty server would fix it? [10:28:36] --- stevegt has become available [10:28:44] my butt says yes, but my butt is stupid and drinks more than derrick. [10:28:46] It might. Not sure what happens with sysid [10:28:58] (with same sysid) [10:29:26] However, I can wait till midnight when daily runs which now has an fs checkv. [10:59:50] --- dev-zero@jabber.org has become available [11:44:56] jaltman you alive? [11:45:03] or even derrick maybe [11:45:11] or someone else who speaks KfW [11:45:31] jaltman left from drinks a while ago. [11:45:34] s/from/for/ [11:45:39] cripes, it isn't even 3pm here [11:45:49] i'll try zephyr :) [11:46:19] At 18:54:21: "time to go to drink" [11:46:48] That would of course be another time.... 2 hours ago. [11:53:27] turned out that the machien was running parallels and it, in turn, couldn't see the network. [11:53:34] the mac could, but not the virtual windoze box on top [11:54:07] In general, I find being able to see the network helps Kerberos work better. [11:54:18] wow, really? I NEVER THOUGHT OF THAT [11:54:31] :-P [11:55:33] We now have 1.5.55 RPM builds for Linux ... [11:56:46] scary [11:57:14] whee? [11:57:26] (unless 1.5.55 is more stable on Unix than previous 1.5.x's) [11:58:11] The idea is to have something that it's easier for people to test. I'm using the 1.5.55 client on a few machines, to play with disconnected stuff. Doesn't seem that unstable, apart from the bugs I know about ... [11:58:46] It's similar to what Russ is doing on Debian. [11:59:18] the hokey kokey? [12:08:34] --- stevegt has left [12:37:23] --- stevegt has become available [12:40:00] --- dev-zero@jabber.org has left: Replaced by new connection [12:40:01] --- dev-zero@jabber.org has become available [13:37:01] --- Russ has become available [14:01:48] Moose: what do you need? [14:03:00] nothing anymore, thanks :) [14:03:23] someone had some really odd problem with KfW, turned out that parallels was futzing up their network, and it all had nothing to do with KfW after all [14:03:39] although I did get them to upgrade from running an ancient copy of Leash to the latest NetIDMgr [14:03:56] ok [14:05:47] the beer was excellent. I recommend this year's Samuel Smith's Winter Welcome Ale [14:09:27] --- Moose has left [14:21:53] --- SecureEndpoints has left: Replaced by new connection [14:23:21] --- SecureEndpoints has become available [15:26:56] --- dmontuori has left [15:29:11] --- dev-zero@jabber.org has left [16:53:45] --- stevegt has left: Replaced by new connection [16:53:45] --- stevegt has become available [17:39:51] --- edgester has become available [18:01:53] --- stevegt has left [19:32:45] I have a vague recollection that someone has a prototype of hostafsd (ie, 'my local filesystem is a fileserver')..is that jhutz, shadow, ?? [19:34:31] "hostafsd" is the _name_ of my software that does that. [19:35:25] See /afs/cs.cmu.edu/project/systems-jhutz/hostafsd but IIRC Derrick did or was going to do some hacking and may have something someplace else. [19:36:50] There is also /afs/cs.cmu.edu/project/systems-jhutz/tafssrv which is the "trivial AFS server", and attempts to solve a different but similar problem. It doesn't export the local filesystem, but rather exports (read-only) a set of volumes defined by configuration, which can contain files backed by files in the local filesystem. [19:50:35] is that something you'd like someone to poke at/help with? [19:52:44] Neither project has gotten much attention in some time. If you have interest in them, feel free. [19:53:09] Max is interested, so I'll aim him in your general direction. [19:55:10] Feel free to aim him directly at the directories I just mentioned. [19:57:07] * stevenjenkins nods. done. [19:57:09] tx. [19:59:38] --- edgester has left [20:05:57] --- stevegt has become available [20:13:54] btw, ls: /afs/cs.cmu.edu/project/systems-jhutz/hostafsd: No such device [20:14:14] it's on a system I don't normally use, though, so still verifying the client is properly set up [20:39:01] --- Russ has left: Disconnected [20:39:38] --- Russ has become available [21:10:46] That'd be a problem on your end, I think. [21:22:32] --- Simon Wilkinson has left [23:34:14] --- manfred furuholmen has become available