**** BEGIN LOGGING AT Fri Apr 01 23:59:56 2005 Apr 02 00:26:56 rwhitby-asleep: done Apr 02 02:21:36 03bzhou * 10unslung/ (Makefile make/py-bittorrent.mk): Apr 02 02:21:36 py-bittorrent 4.0.1 ready for cross testing Apr 02 02:21:36 require build machine to have python2.4 Apr 02 02:21:36 btdownloadcurses.py works Apr 02 02:23:40 enjoy! py-bittorrent actually depends on the latest python.mk which has curses support Apr 02 02:24:54 03bzhou * 10unslung/make/py-bittorrent.mk: added python to _DEPENDS Apr 02 03:24:25 hello Apr 02 03:24:33 is anyone actually here today? Apr 02 03:26:31 hello? Apr 02 03:28:47 dammit... Apr 02 04:01:41 hello Apr 02 04:02:10 hello Apr 02 04:03:27 whew Apr 02 04:03:34 can you give me a hand with something? Apr 02 04:03:48 possilbly Apr 02 04:04:26 after attempting to unsling to a flash disk, the NSLU2 is not responding to anything (even ping). it is as though the ip address has changed. Apr 02 04:04:32 i have not been able to telnet into redboot. Apr 02 04:05:17 what firmware? Apr 02 04:05:24 3.18 Apr 02 04:05:47 have you tried to connect to 192.168.1.77? Apr 02 04:06:11 i cannot even get a ping response from it Apr 02 04:06:24 on 1.77? Apr 02 04:06:30 yup Apr 02 04:06:39 did you change the IP? Apr 02 04:06:45 which is your network adress? Apr 02 04:06:54 0.12 Apr 02 04:07:14 so you couldn't ping 1.77 Apr 02 04:07:38 should change at least one of your systems to 192.168.1.x Apr 02 04:07:54 shouldn't i be able to ping that subnet anyway? Apr 02 04:07:54 an try it again fromthat system Apr 02 04:07:59 or use an alias. Apr 02 04:08:07 alright... i'll disappear for a minute when i make the change... Apr 02 04:08:21 RavenHawk, what is your netmask? Apr 02 04:08:45 255.255.255.0 Apr 02 04:08:58 are we talking about windows or linux? Apr 02 04:09:03 windows Apr 02 04:09:42 Just go add 192.168.0.5 or something to your interface alias list, it'll simplify things. Apr 02 04:09:56 in windows? Apr 02 04:10:18 (to answer the question ,you cannot ping 192.168.0.x from 192.168.1.x if your n etmask is 255.255.255.0 Apr 02 04:10:24 Yes in windows., Apr 02 04:10:41 in the first place, how do i add to an interface alias list? Apr 02 04:10:44 I set mine to be 192.168.{0,1,2,50}.x to support the various vendor defaults. Apr 02 04:10:55 and in the second, as long as there's a router between me and the NSLU2, shouldn't i get there anyway? Apr 02 04:11:13 Oh, you didnt mention the router part.... Apr 02 04:11:50 is your router routing for both 192.168.0.x AND 192.168.1.x Apr 02 04:11:51 well, i am wondering if it is routing properly... it did let me telnet and http to 1.77 two days ago. Apr 02 04:12:05 how is it connected to your router? Apr 02 04:12:11 computer and slug Apr 02 04:12:32 computer by wireless, slug directly to the back of the router via ethernet Apr 02 04:12:49 okay then you dont have a "router" between them. Apr 02 04:13:00 youre pulugged to the "LAN port" right? Apr 02 04:13:09 the NSLU2 is, yes Apr 02 04:13:30 the LAN port is a switch thats generally unrouted unless you are doing something very very strange. Apr 02 04:14:02 the device is a WCG200; its supposed to route. I was able to get to 1.77 via telnet and http two days ago. Apr 02 04:14:24 Okay well I dont know anythingt about that product Apr 02 04:14:26 but i can pretty readily just change the whole system to be on the "1.x" subnet if you think that would help. Apr 02 04:14:41 but if you want to add an alias... Apr 02 04:14:52 i'd rather just be all on the same subnet i think Apr 02 04:15:15 let me pop onto 1.x, i'll be back in a minute... Apr 02 04:15:29 network->properties->tcpip->properties->advanced->ip->add Apr 02 04:15:39 well, it may or may not help you Apr 02 04:15:54 because you'll eventually need to be on the 0.x LAN interface to telnet to redboot. Apr 02 04:16:19 or you can just set your network to 192.168.0.0/255.255.0.0 in which can it'll justu work. Apr 02 04:22:31 oh. that device. Its not routing wireless to lan. Thats bridged internally. its routing your internal LAN packets to the WAN (DOCSIS Modem) Apr 02 04:27:59 ok back Apr 02 04:28:01 that worked, by the way Apr 02 04:28:23 now for the payload: disk2 has an NTFS partition and a damaged ext3 partition. Apr 02 04:28:35 how to repair the ext3 partition using scandisk? Apr 02 04:28:50 unmount it Apr 02 04:28:54 the fsck it Apr 02 04:29:03 fsck is non-destructive? Apr 02 04:29:14 its not mounted, since the NSLU2 doesn't recognize the disk as formatted Apr 02 04:29:19 File System ChecK Apr 02 04:29:37 it fixes whatever issues there are as best it can Apr 02 04:29:53 ~man fsck Apr 02 04:30:03 ~google man fsck Apr 02 04:30:07 Google says man fsck is at http://www.hmug.org/man/8/fsck.html or http://www.rt.com/man/fsck.8.html or http://www.die.net/doc/linux/man/man8/fsck.8.html or http://www.linux.or.jp/JM/html/e2fsprogs/man8/fsck.8.html or http://macslash.org/article.pl?sid=03/01/07/2316238&mode=thread Apr 02 04:32:04 hmmm... it is seeing sdb through sdb4... does that mean its mounted i wonder... Apr 02 04:33:01 how do i tell which device is the partition i want to hit? Apr 02 04:36:23 hello? Apr 02 04:37:06 ok, i'm not timed out and gone... so, this plan to fix the disk through the NSLU2 is going rather poorly... all superblocks corrupt and so forth... Apr 02 04:37:16 is anyone aware of software that can repair an ext3 partition from windows? Apr 02 04:37:34 Hmm Apr 02 04:37:56 if it doesnt work on the nslu2 it aint gonna work on a windows machine. Apr 02 04:38:07 what exactly happened to this partition? Apr 02 04:38:10 well, doesn't work can mean a variety of things Apr 02 04:38:22 i can open and read and write the partition in windows using Mount Anything. Apr 02 04:38:32 I just get an error when i try to resize it using Partition Magic. Apr 02 04:38:41 are you sure you hit the right partition? Apr 02 04:38:44 "bad magic number in super-block" Apr 02 04:38:45 with fsck? Apr 02 04:38:48 i tried all of them Apr 02 04:39:04 do you have busybox installed? Apr 02 04:39:21 not unless it was installed by default Apr 02 04:39:29 okay I cant help you then. Apr 02 04:39:46 can you tell me if the issue might be whether the partition is logical or primary? Apr 02 04:40:05 partn's less than 5 are primary Apr 02 04:40:14 and no, that wont cause fsck to fail Apr 02 04:40:25 i can mark them primary or logical in partition magic; these are logical Apr 02 04:40:27 how did you make this partition? Apr 02 04:40:48 partition magic Apr 02 04:40:49 well, theyre obviously not logical if they are really numbered 1 through 4. Apr 02 04:41:20 What version of PM are you using? Apr 02 04:41:22 i don't see that connection - the NSLU2 is giving them numbers Apr 02 04:41:23 8 Apr 02 04:41:32 sigh Apr 02 04:42:07 i'm rather sure they're marked as logical, as i recall being concerned if that would create a problem down the road Apr 02 04:42:11 I just said. Linux labels logical partitions 5 and above. Apr 02 04:42:20 so if they 1-4, linux doesnt think so. Apr 02 04:42:41 and I would almost blame that on PM because its infamous for munching partition tables beyond repair. Apr 02 04:43:03 well, you're numbering point may be right - or the ext3 partitions may not be showing up in the NSLU2 at all Apr 02 04:43:31 if you have busybox 1.00 installed, I can tell you... Apr 02 04:43:32 which would explain why i get the "bad superblock" error - the only partition its seeing is the one marked primary, which is an NTFS partition. Apr 02 04:43:42 ok, let me install busybox... Apr 02 04:44:07 the onther point is that if they are in fact logical, the nslu2 has no way to addreesas them Apr 02 04:44:18 because the device nodes dont exist for logical partitions. Apr 02 04:44:33 so that's the issue - it isn't seeing them, and i can't fsck them, because they're logical. Apr 02 04:44:43 so i have to mark it as primary, then go back... Apr 02 04:44:45 no problem... Apr 02 04:45:18 or make the logical nodes. Apr 02 04:45:24 shrug Apr 02 04:45:33 "make the logical nodes"? Apr 02 04:45:41 ~google man mknod Apr 02 04:45:45 Google says man mknod is at http://www.hmug.org/man/8/mknod.html or http://www.hmug.org/man/2/mknod.html or http://www.rt.com/man/mknod.1.html or http://www.rt.com/man/mknod.2.html or http://rootr.net/man/man/mknod/8 Apr 02 04:46:13 i think i'm better off convering to primary Apr 02 04:46:25 because at the end of this disaster this disk is going to be my main NSLU2 one Apr 02 04:46:36 so it has to be primary at the end of the day anyway Apr 02 04:47:58 ok, now its primary... lets move it back to the NSLU2... Apr 02 04:48:05 (no need to reboot the NSLU2, right?) Apr 02 04:53:51 "filesystem does not have a UUID" - dangerous or not dangerous do you think? Apr 02 04:54:14 "revision 0 filesystem - fix" - dangerous or not dangerous... Apr 02 04:57:54 checking inodes, blocks and sizes... that's progress... Apr 02 04:58:06 ya Apr 02 05:22:44 "too many illegal blocks in inode" - how big a deal? Apr 02 05:34:23 dammit, i'm smelling a data loss Apr 02 05:34:26 i hate this shit Apr 02 06:41:19 Hi guys, anyone having troubles with fresh openslug builds please? Apr 02 06:41:57 how so? Apr 02 06:42:22 I carefully replicate all the istructions and test an image from ram using redboot but I really cant make it boot Apr 02 06:42:38 test=tested Apr 02 06:43:06 when I exec the kernl it beeps once and then stops Apr 02 06:43:20 the ready/status led keep staying red Apr 02 06:44:22 Openslug 1.8? Apr 02 06:44:31 or 1.7? Apr 02 06:44:42 the latest Apr 02 06:44:47 baked from bk Apr 02 06:44:57 I think it's 1.8 Apr 02 06:45:00 did you do a pull this mornng? Apr 02 06:45:03 sure Apr 02 06:45:19 I pulled like 8 hrs ago Apr 02 06:45:46 repull and rebuild again Apr 02 06:45:56 there were some changes I saw about 8hrs ago Apr 02 06:46:02 uhmm Apr 02 06:46:33 doyou know any faster way then delte /tmp and rebild please? I'm pretty new to oe Apr 02 06:46:50 I don't. Apr 02 06:49:54 nettie: what's in the flash? Apr 02 06:50:07 hangon Apr 02 06:50:13 you need to have a correct mtd partition table in flash to test openslug or unslung from ram Apr 02 06:50:21 so you need to have flashed them once Apr 02 06:50:29 ahhhh Apr 02 06:50:34 ok Apr 02 06:50:47 the default one is in the flash Apr 02 06:50:53 I'll flash then Apr 02 06:51:00 with openslug and see Apr 02 06:51:09 thanx again for the hint, you're very kin Apr 02 06:51:10 d Apr 02 06:56:38 rwhitby-away noway .. same problem Apr 02 06:57:07 rwhitby-away beep and nothing more Apr 02 06:57:17 hangon Apr 02 06:57:21 maybe Apr 02 06:57:33 yes Apr 02 06:57:37 took a bit to boot! Apr 02 06:57:43 great! finally Apr 02 06:57:50 :) Apr 02 06:57:54 what do think about splitting busybox into three packages for unslung? ohne meta package that depends on base and links. base consists of the busybox executable and links has packed all the links in it? Apr 02 06:57:56 good job nettie Apr 02 06:58:17 be sure to add you name to the database of OpenSluggers Apr 02 06:58:44 so people can install busybox-base and have no conflicts wih coreutils or findutils Apr 02 06:59:13 pTweety, unslung package? Apr 02 06:59:25 yes Apr 02 07:00:17 pTweety: good idea Apr 02 07:00:36 shall I check it in so that someone can have a look? Apr 02 07:01:21 if it works, yeah. are you generating the control files from inside the .mk file according to the latest template.mk? Apr 02 07:01:26 ka6sox-zzzz I'll do it eheh Apr 02 07:01:29 sure Apr 02 07:01:38 cool Apr 02 07:01:47 rwhitby-away it's great.. but why you guys didnt use uClibc? Apr 02 07:02:11 nettie: how big is your attached disk? ;-) Apr 02 07:02:16 0 ;) Apr 02 07:02:17 eheh Apr 02 07:02:23 I was planning not using it Apr 02 07:02:25 eheheh Apr 02 07:02:41 you can build for uclibc by a single line change in your local.conf Apr 02 07:02:50 wow Apr 02 07:02:55 I really look forward try it Apr 02 07:02:56 target audience is not diskless Apr 02 07:02:58 :) Apr 02 07:03:08 but we support the diskless usage too Apr 02 07:03:16 so glibc is the default Apr 02 07:04:11 rwhitby-away yeah .. I can understand ;) Apr 02 07:05:17 it's a pretty fully fladge system we get Apr 02 07:05:24 busybox is basically complete Apr 02 07:05:24 nice Apr 02 07:06:15 when you;ve got a 200Gb disk usually attached, busybox becomes less relevant :-) Apr 02 07:06:26 eheh Apr 02 07:07:09 anyone know how it handles aes256 vpn traffic ? Apr 02 07:07:40 the idea is using it as vpn endpoint (network redirector) Apr 02 07:08:35 I need a device to hide somewhere on a pretty close netowrk and use it to connect my remote router and tunnel that network Apr 02 07:08:59 corporate gateway ? ;-) Apr 02 07:09:59 I can not test busybox meta package, cause it has no feed for the dependend packages :-( Apr 02 07:10:11 but I guess it will work Apr 02 07:10:28 well .. I need to access some AXIS DVR's which are located in a yeah call it "corporate network" Apr 02 07:10:29 pTweety: you can set up a local feed on your disk Apr 02 07:10:35 so the idea was a ssl vpn Apr 02 07:10:49 or something else I could tunnel on known open ports Apr 02 07:10:50 nettie: you could just use ssh port forwarding ... Apr 02 07:10:58 yeah +ppp Apr 02 07:11:20 but I read everywhere it's not very reliable solution Apr 02 07:11:37 vpn with ssh+ppp Apr 02 07:11:43 port forwarding isnt an issu Apr 02 07:11:43 eheh Apr 02 07:11:44 issue Apr 02 07:12:25 because it's more complex than that. the dvrs are acutally on a remote network connected to a box which can only be reached by the corporate network Apr 02 07:12:40 so the idea is to vnc on that box and use the dvr Apr 02 07:12:58 instead of a double human-based vnc forward eeheh Apr 02 07:13:08 nod Apr 02 07:16:41 rwhitby does a load average of 1.0 is normal? Apr 02 07:19:37 03rwhitby * 10unslung/ (28 files in 13 dirs): Control file generation for many packages, and addition on microperl target in perl package. Courtesy of Oleg (wl500g). Apr 02 07:20:06 nettie: ixp driver does 4000 or so interrupts per second Apr 02 07:21:34 damn Apr 02 07:22:42 did you try any cpu intensive task? Apr 02 07:22:47 how it performs? Apr 02 07:24:51 not I Apr 02 07:25:09 there's a benchmarks page on the wiki somewhere Apr 02 07:27:48 03rwhitby * 10unslung/make/vsftpd.mk: Bumped the IPK version. Apr 02 07:28:26 03rwhitby * 10unslung/make/perl.mk: Bumped the IPK version. Apr 02 07:28:47 03ptweety * 10unslung/make/busybox.mk: Control now generated; slit into three packages: busybox, busybox-base, busybox-links. Apr 02 07:31:46 03ptweety * 10unslung/sources/busybox/control: Control file is now generated. Apr 02 07:32:20 rwhitby, thinks for the pointer to local feeds Apr 02 07:32:28 this is quite cool for testing Apr 02 07:32:52 indispensible for local testing :-) Apr 02 07:33:55 well, that's it for today Apr 02 07:34:59 night all Apr 02 07:42:46 portmap start/stop/reload/restart script, line 25 there's a --oknodo option which isnt recognized by the start-stop-daemon and preevent portmap to stop. Removing --oknodo solves the issue Apr 02 07:45:02 nettie: you'll need to report that in #oe as it's not an nslu2-specific package Apr 02 07:45:11 ok Apr 02 07:45:28 just wanted to know you guys can replicate the issue Apr 02 07:45:38 reproduce Apr 02 07:46:01 can't at the moment (not near openslug) Apr 02 07:46:35 k Apr 02 07:53:42 microperl now cross-compiles for the slug thanks to Oleg. Apr 02 07:53:58 http://www.foo.be/docs/tpj/issues/vol5_3/tpj0503-0003.html describes microperl Apr 02 07:54:32 rwhitby what's the best way to add packages like iptables vconfig etc..etc.. Apr 02 07:54:46 for openslug or unslung? Apr 02 07:54:49 open Apr 02 07:55:02 are they in OE already? Apr 02 07:55:11 I bealive Apr 02 07:55:27 then just add them to openslug-packages.bb Apr 02 07:56:02 then bb openslug-packages will build ipks for them, and they will end up in the feed next time we update it Apr 02 07:56:18 nice Apr 02 07:57:13 we have a build farm (http://nslu2-linux.thestuffguy.com/gallery/gastronomic) which was bought with donations Apr 02 07:58:07 nice Apr 02 07:58:26 we have a generous community Apr 02 07:59:42 definitely Apr 02 07:59:44 how it works? Apr 02 07:59:59 it's like a box where you guys have access and you build stuff there? Apr 02 08:00:30 yep - they are our official build machines Apr 02 08:00:34 03rwhitby * 10unslung/ (Makefile sources/vsftpd/uclibc-prctl.patch): Forgot the new patch Apr 02 08:00:49 okay so I cant seem to unsling to hdd-data. Apr 02 08:01:01 on my prodslug2 with no serial. Apr 02 08:01:23 what happens? Apr 02 08:01:37 it boots to the jffs2. Apr 02 08:01:39 I unslung to hdd-data on devslug last night after the ipkg fix Apr 02 08:01:52 unless I'm missing something... Apr 02 08:02:01 load the firmware, unsling hdd-data; reboot Apr 02 08:02:09 that's it Apr 02 08:02:28 yep, on my 3rd try, it just plain dont do it. Apr 02 08:02:40 rwhitby what's the best/fastest way to recompile an image? Apr 02 08:02:41 the flagfiles are all there. Apr 02 08:02:42 is .sda1root there in the jffs2, and does /sbin/init exist on /share/hdd/conf? Apr 02 08:02:52 nettie: depends what has changed Apr 02 08:03:02 the -image package is always rebuilt Apr 02 08:03:25 no, but I do have a /share/hdd/data/bin/init . ;-) Apr 02 08:03:28 df Apr 02 08:04:27 I see Apr 02 08:04:27 so no .sda1root? Apr 02 08:04:47 if I want to add iptables with relative kernel modules etc what should I do? Apr 02 08:04:48 it should be there in both the jffs2 and in /share/hdd/data (I put it in both) Apr 02 08:05:03 add it to packages and recompile the image? Apr 02 08:05:06 nettie: update the defconfig in openslug-kernel-... Apr 02 08:05:23 kernel modules are packaged and uploaded to the feed automatically if they are in the defconfig Apr 02 08:06:06 dyoung: let me wipe my sda and try to recreate Apr 02 08:06:17 ok great Apr 02 08:06:54 maybe I also can install pakcaged already uplaoded by 3rd party Apr 02 08:07:35 where's the rep? Apr 02 08:07:39 uploaded by who? Apr 02 08:07:48 ipkg update ; ipkg list Apr 02 08:07:53 cool Apr 02 08:08:13 eheh the point is that I'm not expecting so much stuff working Apr 02 08:08:13 eheh Apr 02 08:08:17 I'm gonna reflash it for good measure. Apr 02 08:08:34 nettie: ye of little faith .... Apr 02 08:09:05 before the ipkg debacle we were just about to go to beta with openslug Apr 02 08:09:11 and we don't go to beta lightly Apr 02 08:12:04 I'm crashed I go bed.. thanx again and see you tomorrow. night Apr 02 08:12:17 ciao Apr 02 08:13:13 dyoung: I unslung to a previously rootfs'd drive, which might have masked something Apr 02 08:15:55 one thing that's gonna have to be made clear in the README is to wait until the quota checking is done if you are unslinging to hdd-data Apr 02 08:16:10 README note above Apr 02 08:16:36 and that's hard to tell, cause it's well after the beep for a large drive Apr 02 08:17:04 maybe we can change the beep to wait for quota checking to finish Apr 02 08:21:00 Hmm Apr 02 08:21:32 did you see that python builds for the wl500g? :-) Apr 02 08:21:49 I might have to learn python - it's seems to be more cross-compile-compatible than perl ... Apr 02 08:22:26 I thought you hated snakes? Apr 02 08:22:43 I'm all in a quandry of which I should learn. Apr 02 08:22:50 (or rather expand my skillset in) Apr 02 08:24:13 I don't hate snakes. Apr 02 08:24:52 I just can't write it yet, and have only a novice to competent reading knowledge of it. Apr 02 08:27:25 dyoung: I figure if we change the &>/dev/null to a 2>/dev/null in the call to rc.quota in rc.local, then that will make it sequential. Apr 02 08:28:01 it would also make unslung packages wait for the disk to become read-write again too. a win/win. Apr 02 08:28:14 rwhitby: python is good Apr 02 08:28:41 eno: yeah, most languages are good when you know them :-) Apr 02 08:29:13 and yes, some are intrinsicly good too Apr 02 08:29:16 i'm a language nuts, and i like python qiute a lot Apr 02 08:32:33 dyoung: yeah, that works well Apr 02 08:33:18 Okay cool. Apr 02 08:33:34 (you'll note that this time around I'm being a perfect idiot for the purposes of testing...) Apr 02 08:47:38 dyoung: seems to still work for me Apr 02 08:47:47 doesnt work for me. Apr 02 08:47:53 (unsling hdd-data to an empty data partition) Apr 02 08:48:01 will try reformatting the disk completely next Apr 02 08:48:04 jffs-hdd-data works. Apr 02 08:48:13 but that one is fairly brainless. Apr 02 08:48:30 is the rootfs being copied correctly? are the .sda1root flag files being created? Apr 02 08:48:39 yes and yes. Apr 02 08:49:02 is it a slow usb disk thing? Apr 02 08:49:16 it might be.... Apr 02 08:49:31 theres are fairly regular disks though. WD. Apr 02 08:49:33 160G Apr 02 08:50:49 03rwhitby 07 * r1.3133.1.1 10openembedded/packages/nslu2-binary-only/ (2 files in 2 dirs): Unslung 4.10-alpha: Added wait-for-quotacheck.patch Apr 02 08:52:19 guess we need some more alpha testers to see whether your problem is local or not? Apr 02 08:53:08 I dunno. Apr 02 08:53:43 I try with a single disk later Apr 02 08:53:52 has that disk ever worked with the switchbox timing before? Apr 02 08:54:08 I had it unslinged for 2.x Apr 02 08:54:09 you said "disks" ? Apr 02 08:54:23 this isn't a multi-lun issue is it? Apr 02 08:54:36 I have 2 identical disks plugged in on port 1 and port 2, so not LUN issue Apr 02 08:54:42 oh Apr 02 08:59:42 gotta love open source. from 0 to 91 packages for the wl500g in five days. Apr 02 09:45:21 what's the wl500g? Apr 02 09:55:47 ~google wl500g Apr 02 09:55:51 Google says wl500g is at http://wl500g.dyndns.org/ or http://forum.chupa.nl/ or http://uk.asus.com/products/communication/wireless/wl-500g/overview.htm or http://www.chupa.nl/forum/forumdisplay.php?forumid=8 or http://paul.luon.net/journal/hacking/WL500g.html Apr 02 09:58:24 the wl500gx (deluxe model) is pretty much a slug with wireless Apr 02 09:59:36 dyoung: we can get rid of /linuxrc now, right? Apr 02 10:16:39 03rwhitby 07 * r1.3139 10openembedded/packages/nslu2-binary-only/unslung-rootfs_2.3r25.bb: Removed /linuxrc from unslung - we have better technology now :-) Apr 02 10:20:27 Yep. Apr 02 10:38:30 I'm tempted to remove the specialised options from the usage message of unsling, and just have "disk1" and "disk2" options which map to hdd-data and flash-data respectively Apr 02 10:38:54 the other options would remain in the code for those in the know Apr 02 10:39:32 I need to find a spare disk to test things out on Apr 02 10:39:41 it doesnt work for me yet with only 1 drive attached Apr 02 10:41:31 what do you think about the disk1/disk2 unsling idea? Apr 02 10:42:45 the list of possible options is rather large. Apr 02 10:43:16 simple disk1 or disk2 is more intuitive for "Joe User". Apr 02 10:43:48 yeah, and with the wait-for-quotacheck patch, putting rootfs and packages on the data partition is safer Apr 02 10:43:59 What do you think about defaulting to maint mode? Apr 02 10:44:22 can't do that - uses too much ram Apr 02 10:44:23 so you actually have to DO something before making a mess. Apr 02 10:44:55 Yes, I know, so the first step would be to unsling to something. Apr 02 10:45:24 so you cant accidentally fill up your jffs2 with junk and make yourself unbootable. Apr 02 10:46:05 interesting idea Apr 02 10:46:20 so put /.ramdisk in the image Apr 02 10:46:22 so if you do nothing, its kind of like stock. Apr 02 10:46:25 and have unsling remove it Apr 02 10:46:44 except with a working GENESYS patch. ;-) Apr 02 10:48:55 I'm not sure how I feel about that either, but it was a passing thought a few weeks ago. Apr 02 10:49:22 hmm - actually we can leave .ramdisk there cause it is a lower priority than the Apr 02 10:49:30 disks Apr 02 10:49:55 but switchbox will currently drop into jffs2 if the disks don't mount. Apr 02 10:50:06 we would need to change that to maintenance instead. Apr 02 10:50:21 and then maintenance to jffs2, then jffs2 to recovery Apr 02 10:51:05 ah - the only problem is that that makes unsling much more complex, cause it has to then mount the jffs2 to set up the flag file. Apr 02 10:51:09 so we won't do it. Apr 02 10:51:23 oh yeah. Apr 02 10:51:52 if people stuff up jff2 before unsling, then they deserve what they get, and they haven't invested much to be too disappointed about a reflash Apr 02 10:52:53 but I will simplify unsling with disk1 and disk2 Apr 02 10:53:02 I wish I could test it. Apr 02 10:55:12 03rwhitby 07 * r1.3140 10openembedded/packages/nslu2-binary-only/ (unslung-rootfs_2.3r25.bb unslung-rootfs-2.3r25/unsling): Simplified unsling for the casual user. Complex options are available but not advertised in the usage message. Apr 02 11:02:48 new copy of 4.10-alpha will be up soon Apr 02 11:02:53 back later Apr 02 12:46:46 Not really a developer question, but more than a general. Can anyone tell me if any of the 3 NFS packages for unslung 3.18beta support the map_static option? I've tried with all of them (unfs3, nfs-utils and nfs-server) and none of them seem to know about the option. Apr 02 14:33:39 ChezDude2010: dunno Apr 02 14:48:16 unsling disk1 works Apr 02 14:49:44 the other thing I'm going to do is remove the /tmp ramdisk and clear /tmp and /mnt/backup on each boot Apr 02 14:50:08 (remove /tmp ramdisk only if you're loading rootfs from external usb disk) Apr 02 15:06:11 I think I will remove maintenance mode altogether. The normal situation (unsling rootfs to a disk) allows for reflashing in normal mode through the web interface. The situation where people want to run from jffs2 can be handled by any of the other upgrade modes (telnet into redboot, or touch .ramdisk, or use upslug). Apr 02 15:43:46 03rwhitby 07 * r1.3141 10openembedded/packages/nslu2-binary-only/ (2 files in 2 dirs): Unslung 4.11-alpha: Removed maintenance mode, simplified advertised unsling arguments to disk1 or disk2. Removed /tmp ramdisk if unslung to an external disk. Clears /tmp and /mnt/backup on each boot. Apr 02 17:38:50 Hmm - seems that removing maintenance mode is not such a good idea. The problem is if you have lots of unslung packages running, and you have very little memory left, then you can upgrade through the web interface. So then you have to remove the hard disk (assuming you don't want to remove all your packages), and then you're back to booting from jffs2 (in which case maintenance mode is required) Apr 02 17:38:50