[00:33:43] is it at all possible to build a inode fs from 1.4.8 src? [00:33:48] (Sol 10 x86) [00:34:01] s/\ a\ /an/ [00:41:06] Not specifying anything about fs type on Solaris used to give us an inode fs, but now: "This program is compiled with AFS_NAMEI_ENV" is the barf message from salvager... [01:05:39] --- haba has become available [01:06:36] You have no idea what hassle it is to build a inode (not namei) vol-dump for Sol10 Sparc. [01:07:47] So if anyone has a working binary somewhere in AFS, send the path to mansaxel@besserwisser.org. [01:33:09] --- haba has left [02:36:25] --- haba has become available [03:06:31] I have one now; if one bends the build system by undef'ing AFS_NAMEI_ENV it actually builds one. [03:35:56] Why the $%^&*( does vol-dump want the afs-syscall (65). [03:51:48] --- haba has left [04:18:52] --- SecureEndpoints has left: Replaced by new connection [04:37:45] --- haba has become available [05:44:07] haba: something with tokens? [06:06:29] abo: perhaps. Right now I'm treating it as low priority, since I got things working. Fixing something that says "--enable-inode-filerserver" (which would be more like bringing the configure system up-to-date) instead of having --enable-manei-fileserver with that already standard and no way of overriding without going to post-configure hacks is more important. [06:07:00] namei. not manei. new mbp keyboard is good but I'm not used to it yet. [06:21:56] you realize that inode fileserver is the default (currently in cvs) only places where you'd presumably already have it deployed, and thius any fix you might send isn't really interesting, right? [06:23:34] shadow: you mean namei is default. The problem is that it is difficult to build a inode on present Solaris if you want to migrate or find an old HD where there might be data that you want get out. [06:36:16] abo: Yes, probably some token stupidity, volserver does not need tokens to access the /vicep*, why should vol-dump need? [06:36:18] shadow: what harald said, plus it is more a question of "the configure parameters need updating to reflect the fact that namei now is standard" -- because there is a flag to say "I want namei" which is in effect a noop, but there is no "i want inode" which would actually do something. And yes, this has mostly archaeological interest. [06:43:23] i mean namei is the default *where it should be* [06:45:15] as to vol-dump, the inode version will need the system call so it can iopen() [06:46:23] if you want to provide a flag which enables inode, and are careful to make it work correctly given the already existant i want namei, please do. [06:46:31] you can, however, --disable-namei-fileserver [06:46:57] so, "difficult" [06:48:36] > if one bends the build system by undef'ing AFS_NAMEI_ENV i don't consider --disable-namei-fileserver to be particularly much bendage, and at some point it was tested and worked correctly. does it not now? [06:53:16] I'm not sure that the --disable-namei-fileserver works for Sol10 as expected. [06:53:36] it did when inode became the default [06:53:39] After the experiences from mansaxel I doubt it. [06:53:47] namei-by-default-on-sol10-20080816 [06:53:54] added it, and was tested at the time. [06:54:30] it's easy enough to find out. in vol-salvage.c, add a #error protected by #ifdef AFS_NAMEI_ENV after the includes, then try it [06:54:38] or i will when i get my sol10 vm back [06:54:59] I think after we actually got the data back, we may have some time to test it. [06:55:05] ok [06:55:43] I'd like to see vol-dump built and installed by default so that it won't rot. [07:07:06] I can confirm that 1.4.8 on Sol10 x86 gave me namei binaries in any build situation short of "vi afsconfig.h" after a configure run. [09:25:26] --- haba has left [09:44:36] --- SecureEndpoints has become available [12:26:49] --- reuteras has left [12:43:24] --- haba has become available [13:27:53] --- Russ has become available [14:16:21] --- haba has left [15:54:59] --- tkeiser@sinenomine.net/owl has left [19:01:01] --- edgester has become available [19:01:23] --- edgester has left [23:17:34] --- reuteras has become available [23:40:28] --- Russ has left: Disconnected [23:40:40] --- Russ has become available [23:41:50] --- Russ has left: Disconnected [23:43:42] --- haba has become available [23:59:05] --- shadow@gmail.com/owl922BC4C1 has left