[00:27:33] --- pod has left [00:38:32] --- pod has become available [03:52:18] --- reuteras has left [03:52:48] --- reuteras has become available [04:12:22] --- mdionne has become available [04:25:01] --- mdionne has left [04:54:29] --- haba has become available [05:02:11] --- Simon Wilkinson has become available [05:26:35] --- haba has left [05:26:52] --- haba has become available [05:27:15] --- haba has left [05:34:02] --- Simon Wilkinson has left [05:36:04] --- shadow has become available [05:40:36] --- shadow has left [06:09:42] --- mdionne has become available [06:43:43] --- Kelli Ireland has left [06:45:08] --- abo has left [06:52:15] --- deason has become available [06:52:24] --- haba has become available [06:58:54] --- Russ has become available [07:02:22] --- Kelli Ireland has become available [07:21:20] --- mmeffie has left [07:23:56] --- haba has left [07:25:25] --- haba has become available [07:37:21] --- matt has become available [07:47:28] --- jaltman has become available [07:58:21] --- reuteras has left [08:33:27] --- asedeno has left [08:46:49] --- kaj has left [09:00:30] * Russ is doing the final build of 1.4.12.1+dfsg-1 for Debian upload now. [09:16:39] "A callback on Windows is defined by cm_HaveCallback() returning TRUE. Please use cm_HaveCallback() to test for callback status instead of testing individual fields." [09:16:58] What about the distinction between has callback promise, and is stat'd? [09:21:04] I see that the CM_SCACHEFLAG_STATD is unlikely to be correct--but is there an equivalent? [09:21:19] unlikely==it isn't updated :) [09:24:06] there is no distinction in the Windows CM [09:25:02] status info is only current if a callback promise is active [09:31:46] check [09:33:59] --- asedeno has become available [09:34:08] --- asedeno has left [09:48:29] --- asedeno has become available [09:49:53] What's the maximum number of sites for a volume? 16? [09:50:48] <_cclausen> doesn't it depend if you are using namei or inode? [09:51:32] I think that would be clones rather than sites. Sites are a VLDB limitation, IIRC. [09:51:43] <_cclausen> ah, ok [09:52:20] russ - 16, but 3 of those are taken, iirc. [09:52:29] ie, 1 is the 'clone', 1 is backup. [09:52:44] 1 is the volume itself, so the practical limit on replica sites is 13. [09:52:55] I think. I'd have to look at that again as it's been a few yrs.. [09:53:13] derrick, tom, or andrew might be more current on that. [09:53:40] Ah, that's right, there's the RW, the BK, and the transient clone, so you can have 13 ROs. [09:54:31] and there are sites where that max # is a real limit for them. [09:54:47] uh. so why are patches which were submitted still in gerrit's "open", but showing submitted? [09:54:49] although multi-cell (e.g., openEfs) [09:54:54] can help with that. [09:57:49] shadow: maybe because it hasn't been merged/cherry-picked? [10:02:47] A similar sort of question -- what's the limit on how many IP addresses that a fileserver can register with the VLDB? [10:03:39] shadow: it's marked merge pending in Gerrit. That's weird. That's a new one on me. [10:04:04] Might be something we need Simon to look at. [10:04:47] tom said 16... I thought it might be 15 [10:05:18] yeah, 15 as far as I know, VL_MAXIPADDRS_PERMH [10:05:24] Danke. [10:05:36] I thought he might have just mentioned that, but I didn't pay enough attention. [10:13:41] there's no cherry-pick button. something weird is up [10:14:07] all the changes were "submitted", but i guess something ailed [10:15:59] hmm, "merge pending" [10:18:27] yeah. question is why [10:22:29] I've seen that before. After a gerrit restart the submitted patchsets were merged [10:23:52] Should I kick gerrit? [10:31:54] --- deason has left [10:32:07] --- jaltman has left: Disconnected [10:32:33] --- haba has left [10:33:03] yeah [10:33:24] Doing that now. [10:34:11] Yup, that did it. [10:35:56] --- matt has left [10:36:29] --- sanket has become available [10:36:53] cclausen: Cool. Works. [10:38:45] Okay, my question again. How do I force the function afs_UFSRead in src/afs/VNOPS/afs_vnop_read.c to be called. I am trying to put some printk's to test a few branches. It is called when the kernel wants to read data from cache to kernel space. [10:45:38] No activity. [10:52:43] <_cclausen> unfortunarely, everyone just went to lunch [10:53:01] <_cclausen> (most of the people in here are at the AFS workshop: http://workshop.openafs.org ) [11:14:29] Have you tried ... reading a file in /afs while using a disk-based cache? [11:15:55] I found an alternatvie. Flushed all cache for my volume and then re-read [11:16:00] that seemed to work [11:35:28] --- haba has become available [11:49:51] --- jaltman has become available [11:59:11] --- deason has become available [12:10:49] --- sanket has left [12:19:12] --- phalenor has left [12:34:50] > How do I force the function afs_UFSRead er. how do you force a read from disk? by needing data which is on disk... [12:35:50] do you sometimes bypass that from the vfs-level cache or something, though? [12:37:14] (if that's it, you can drop those caches by echoing something to /proc/sys/vm/drop_caches IIRC) [12:37:37] "something" is "3" I think; I forget where it's documented what different values do [12:39:32] Andrew: "not if you're on !linux" [12:41:57] yeah yeah; I'm trying to listen to a talk and type at the same time; it's a wonder I say anything understandable [12:42:48] Fair enough. [12:43:18] :) you meant just the drop_caches thing, right? [12:43:29] Yeah. [12:45:10] same problem. trying to listen, compile, debug, and read and reply here. i'm doomed. [12:46:37] deason, shadow, et al: http://blogs.hbr.org/bregman/2010/05/how-and-why-to-stop-multitaski.html [12:47:07] (I write as I'm IM'ing, writing RT tickets, reviewing proposals, etc...) [12:47:58] didn't do marijuana but i have slept like ass. so, i was already boned, steven [12:48:05] steven.jenkins: yeah, and as you send links to us while I'm trying to listen to someone else :) [12:48:19] absolutely. I was well aware of the irony of sending the link.. [12:56:35] --- matt has become available [12:57:02] gerrit is doing something I don't understand in my lastest updated changeset on 2011 [12:57:19] it's showing only added files--but diff -r in the local repo shows all changes I expect. [12:57:35] this is after a git apply and add ala derrick's formula discussed here on 5/23 [13:00:48] 'git add' needs to be for all files modified or added, not just completely new files [13:00:52] or use 'commit -a' [13:01:23] 'show'/'log -p' should be a more accurate view of what you're about to push [13:01:46] 'show --stat' or 'log --stat' would give the summary thing [13:02:09] I'm a dumbbbbbbss [13:02:11] ass [13:31:32] --- kaj has become available [14:44:02] --- haba has left [14:46:33] --- haba has become available [15:08:12] --- deason has left [15:10:13] --- haba has left [15:10:49] --- mdionne has left [15:12:37] --- matt has left [15:13:25] --- Kelli Ireland has left [15:13:27] --- jaltman has left: Disconnected [15:32:43] --- Simon Wilkinson has become available [15:33:56] --- Russ has left: Disconnected [15:38:48] --- Simon Wilkinson has left [15:38:59] --- Simon Wilkinson has become available [15:40:48] --- Kelli Ireland has become available [15:41:27] --- mdionne has become available [16:04:01] --- Kelli Ireland has left [16:06:58] --- mdionne has left [16:13:51] --- phalenor has become available [16:50:54] --- Simon Wilkinson has left [17:45:31] --- Russ has become available [17:45:50] --- matt has become available [17:54:41] --- jaltman has become available [18:59:30] --- matt has left [19:36:07] --- jaltman has left: Disconnected [19:54:03] The newly imported Heimdal config parser fails to build for me. [19:54:29] Lots of errors: missing errno.h include, missing TRUE and FALSE, strcspn not declared, etc. [19:54:54] Enough that I think there's something more fundamental broken for me than just a few warnings on a different platform, but I'm not sure what happened. [20:04:38] --- Kelli Ireland has become available [20:12:46] Pushing a fix to Gerrit as 2040, although I think we also need a replacement for a missing strtok_r. Heimdal provides one in libroken, but we don't provide one currently. [20:21:18] --- matt has become available [20:28:58] --- Born Fool has become available [20:31:48] --- jaltman has become available [20:39:22] And looking at it more, I suspect it is just platform -- Linux is more careful than most platforms about giving you only the prototypes from headers that you asked for and not leaking prototypes from other headers. [20:43:12] --- matt has left [20:45:27] --- Kelli Ireland has left [20:55:53] --- Born Fool has left [21:29:29] * Russ should go to bed, rather than pushing three versions of a patch because I keep seeing stupid mistakes. [21:50:59] --- Russ has left: Disconnected [22:07:56] you're doing better than me. i fucked up. [22:11:22] --- reuteras has become available [22:12:24] --- mmeffie has become available [23:19:41] --- Simon Wilkinson has become available [23:19:42] --- matt has become available [23:21:55] Russ; I suspect it is platform. Derrick and I probably both verified on Mac OS X. [23:42:48] --- matt has left [23:48:42] --- kaj has left