**** BEGIN LOGGING AT Tue Sep 22 02:59:57 2009 Sep 22 03:07:02 i think i'm gonna pass on lvm for the sheeva tonight, i am more interested in moving data back on a changed md0 starting before i go to bed tonight Sep 22 03:07:41 md0? you have a "freebsd ramdisk device" on your linux? :> Sep 22 03:08:02 /dev/md0, my raid1 Sep 22 03:08:08 oh Sep 22 03:09:23 well... Sep 22 03:09:24 stalker ~ # uname -a; cat /etc/*release Sep 22 03:09:24 Linux stalker 2.6.31-openrd #2 PREEMPT Tue Sep 22 03:58:47 CEST 2009 armv5tel Feroceon 88FR131 rev 1 (v5l) Marvell Ope Sep 22 03:09:28 nRD Client Board GNU/Linux Sep 22 03:09:30 Gentoo Base System release 2.0.1 Sep 22 03:21:14 gentoo makes it easy to strip your system Sep 22 03:21:28 and keep stuff off your system with USE=-X and the like Sep 22 03:21:44 but i love not having to compile Sep 22 03:27:37 morfic: well, i just set up distcc with my desktop Sep 22 03:28:55 it's nice once it's all set up, but not to get going Sep 22 03:30:21 crossdev --target armv5tel-softfloat-linux-gnueabi, install distcc, create some symlinks in /usr/lib/distcc/bin and remove the default g++, gcc, cc and c++ symlinks Sep 22 03:30:35 and set up distcc as you would usually do Sep 22 03:33:37 you did notice i run no gentoo system, not even a sourcemage system, so things now would not be as simple, i am fine, just need to get some more space for /usr/share stuff and the like to make system more usable Sep 22 03:43:38 chunksize on raid1 for my two usb drives is something that makes everyone queasy it seems :) while writing might not be affected, i wonder what impact chunk size of raid1 and block size of xfs will have on read performance Sep 22 03:56:58 "Compliant with the Cortex A8, Sheeva also supports both the ARMv6 and ARMv7 instruction sets" - if that's true, why does `uname -m' show armv5tel? Sep 22 03:59:11 where does it say that.. Sep 22 04:15:50 http://www.marvell.com/technologies/cpu_history/cpu_history.jsp Sep 22 04:22:39 arachnist: ask rabeeh if that is correct with regard to the kirkwood Sep 22 04:23:13 what would those instruction sets do for the binaries on the sheeva if they are indeed supported? Sep 22 04:23:46 arachnist: uname -m' shows armv5tel because that is how it is compiled. Sep 22 04:25:18 morfic: I think that might be marketing hyperbole. The functional specs for the kirkwood mentions ARMv5TE Sep 22 04:25:47 cpuinfo only shows that Sep 22 04:25:49 but ask rabeeh...he should be able to clarify Sep 22 04:26:25 Processor : Feroceon 88FR131 rev 1 (v5l) CPU architecture: 5TE Sep 22 04:27:14 and it does not say that it is compliant with cortex 8. Sep 22 04:27:35 hm Sep 22 04:27:38 soory misread it does Sep 22 04:28:02 well Sep 22 04:28:03 http://dpaste.com/96592/ Sep 22 04:28:05 sorry the advertising stuff does say it is cortex A8 compliant Sep 22 04:28:26 compiling it with "-march=armv6 -falign-functions=1 -falign-jumps=1 -falign-labels=1" and running gives illegal instruction Sep 22 04:28:47 so, i guess it's not armv6 Sep 22 04:29:59 marketing always finds ways to hype stuff, if it was up to marketing, my myTouch3G has 192MB RAM, no need to mention that 64MB is for fb and 32mb for radio, and only 96MB are actually there for linux kernel/OS Sep 22 04:30:09 haha Sep 22 04:30:23 take a look at the functional spec http://www.marvell.com/files/products/embedded_processors/kirkwood/FS_88F6180_9x_6281_OpenSource.pdf Sep 22 04:30:25 they are not lying, they are just not truthful Sep 22 04:30:31 http://www.marvell.com/files/products/embedded_processors/kirkwood/FS_88F6180_9x_6281_OpenSource.pdf Sep 22 04:31:02 http://www.marvell.com/files/products/embedded_processors/kirkwood/FS_88F6180_9x_6281_OpenSource.pdf Sep 22 04:31:40 hmm pasted the link as invisible Sep 22 04:31:54 http://www.marvell.com/files/products/embedded_processors/kirkwood/FS_88F6180_9x_6281_OpenSource.pdf Sep 22 04:31:56 we saw it, 3 times Sep 22 04:32:01 4 times Sep 22 04:32:01 well Sep 22 04:32:11 i guess openrd is not armv6 Sep 22 04:32:14 oh not seeing on my end dumb bug in my client Sep 22 04:32:26 nice, not just me having fun bugs ;P Sep 22 04:32:27 anyone with sheevaplug could check http://dpaste.com/96592/? Sep 22 04:33:51 Illegal instruction Sep 22 04:34:34 ok, moving on Sep 22 04:34:37 apologies for the mutltiple link posts Sep 22 04:34:49 tinker-f595: we survive, barely, but we will Sep 22 04:35:05 haha Sep 22 04:35:08 tired... Sep 22 04:35:10 time for sleep Sep 22 04:35:33 * morfic thinks he could build and test a few raid options before moving files back if moving them the first time finishes not too long after midnight Sep 22 04:35:58 and what time zone are is morfic in? Sep 22 04:38:28 tinker-f595: central time us, germany-7 Sep 22 04:38:47 oh you are just waking up ;) Sep 22 04:38:55 early riser Sep 22 04:38:58 what? Sep 22 04:39:02 23:38 here Sep 22 04:39:13 late sleeper is more like it Sep 22 04:39:29 waiting for rsync -av /home /mnt to finish Sep 22 04:39:32 thought you said germany Sep 22 04:39:32 misread Sep 22 04:39:35 From the "88F6281 Integrated Controller Hardware Specifications"- Compliant with v5TE architecture, as published in the ARM Architect Reference Manual, Second Edition Sep 22 04:40:06 so i can umount /home and mdadm --stop /dev/md0 and the --create it again with changed chunk size Sep 22 04:40:19 21:40 here...going to have an early night Sep 22 04:40:29 tinker-f595: i meant GermanyTime minus seven hours Sep 22 04:40:37 why germany time? Sep 22 04:40:41 west coast Sep 22 04:40:46 easier to remember? Sep 22 04:40:58 yes west coast. Sep 22 04:41:02 i call family a lot, so i think in +7 hours for them Sep 22 04:41:05 pacific time Sep 22 04:41:11 not gmt or utc Sep 22 04:41:32 phases of the moon works better ;) Sep 22 04:44:55 morfic night Sep 22 04:45:45 'nigt tinker-f595 Sep 22 05:38:13 morfic: i'd suggest you try out raid-Z ;) Sep 22 05:40:11 would be nice, sounds a little iffy? rebuilding it i mean? Sep 22 12:00:06 | os = [[Debian]] 9.04 (since rev. 1.2), [[Ubuntu (operating system)|Ubuntu]] 9.04 (before) Sep 22 12:00:12 wtf is debian 9.04 Sep 22 12:01:13 does it come with debian as stock now? Sep 22 12:01:40 ECONTEXT Sep 22 12:01:51 ? Sep 22 12:02:04 where is that info from? Sep 22 12:02:23 wikipedia Sep 22 12:03:24 i thought the only debian was Martyns one? Sep 22 12:04:18 Cause it contradicts itself, saying martin has a debian port Sep 22 12:06:06 I've commented that out Sep 22 12:07:24 Reedy: probably whoever has been editing wikipedia has just been confused. Sep 22 12:07:30 mhmm Sep 22 12:07:38 I've commented it out for the moment Sep 22 12:07:44 Will remove it later Sep 22 12:13:02 isn't there out-of-the-box kernels for it in debian proper now? Sep 22 12:13:15 you don't need anything special, just current debian armel Sep 22 12:18:51 anyone with openrd-client? my second gigabit ethernet port doesn't seem to work properly with a custom kernel Sep 22 12:19:38 in dmesg i get: Sep 22 12:19:39 eth1: link up, 10 Mb/s, half duplex, flow control disabled Sep 22 12:19:56 arachnist: probably wrong phy address Sep 22 12:20:04 arachnist: or something like that Sep 22 12:20:08 as in, mac? Sep 22 12:20:12 phy address Sep 22 12:20:31 what phy addresses are specified in the openrd bsp file? Sep 22 12:20:39 what bsd file? Sep 22 12:20:43 bsp* Sep 22 12:21:25 where can i find it? Sep 22 12:21:37 which kernel patches did you use to build your kernel? Sep 22 12:22:55 0001-ARM-Kirkwood-Marvell-OpenRD-Base-board-support-added.patch, 0001-ARM-Kirkwood-Marvell-OpenRD-Client-board-support-add.patch, 0001-Kirkwood-Initialise-SATA-for-OpenRD-Base.patch and replaced mach-types Sep 22 12:23:01 where do i find those patches? Sep 22 12:23:28 i found them on the http://groups.google.com/group/openrd/topics Sep 22 12:28:38 http://groups.google.com/group/openrd/msg/3ecaba96b58f2187?&q=Kirkwood+Initialise+SATA+for+OpenRD+Base <| here's the exact post with the patches Sep 22 12:30:31 arachnist: that patch doesn't create a eth1 Sep 22 12:30:40 arachnist: so how did you get the eth1? Sep 22 12:30:50 it's just there Sep 22 12:31:24 dmesg said: Sep 22 12:31:25 net eth0: port 0 with MAC address 00:50:43:01:de:52 Sep 22 12:31:25 net eth1: port 0 with MAC address 00:50:43:01:de:53 Sep 22 12:31:47 ah, looking at the wrong ile Sep 22 12:31:52 + .phy_addr = MV643XX_ETH_PHY_ADDR(8), Sep 22 12:31:54 + .phy_addr = MV643XX_ETH_PHY_ADDR(18), Sep 22 12:31:58 that latter one seems unlikely Sep 22 12:32:01 can you email me your dmesg? Sep 22 12:32:17 will a dpaste.com paste be sufficent? Sep 22 12:32:27 sure Sep 22 12:33:23 http://dpaste.com/96692/ Sep 22 12:36:41 think i'll try with "MV643XX_ETH_PHY_ADDR(16)" in the mean time Sep 22 12:37:10 arachnist: right, or try 9 Sep 22 12:37:27 arachnist: can you see what PHY chip is on the board? Sep 22 12:37:36 how can i check it? Sep 22 12:37:42 or if you have phylib enabled, can you check what phylib driver bound to it? Sep 22 12:37:52 arachnist: well, if you have the bare pcb, try to find the ethernet phy chip Sep 22 12:38:04 it should have a part number 88E1??? if it's a marvell phy Sep 22 12:38:13 or there could be two of them Sep 22 12:38:19 but i would guess that there is a single dual-port chip Sep 22 12:38:56 1116_NNC if the diagrams are to be believed Sep 22 12:39:48 1116 is single-port Sep 22 12:39:51 ron_: are there two of them? Sep 22 12:40:11 removing the enclosure will be rather hard with all the cables attached ;> Sep 22 12:40:21 arachnist: then leave it :) Sep 22 12:40:23 on the -client I think so, as there is only a single eth on the -base Sep 22 12:40:34 ron_: so the client has the same 1116, but twice? Sep 22 12:41:15 that's what the circuit diagram in front of me (and empty pads on the -base board) would have be believe Sep 22 12:42:02 ok Sep 22 12:44:10 ron_: where are the schematics? Sep 22 12:44:54 the copy I have came off the CD that came with it ... Sep 22 12:45:10 I thought most of that stuff was (freely) online somewhere though Sep 22 12:45:11 ron_: ok, then can you tell me what lines the CONFIG0-3 pins are hooked up to on each of the 88E1116R chips? Sep 22 12:45:47 i mean, to VSS, VDD, LED0 or LED1 Sep 22 12:48:09 good question... depends on what resistor pads are bridged by default ... Sep 22 12:48:46 hm :) Sep 22 12:49:34 lennert: well, setting it to 16 didn't help Sep 22 12:49:39 will try with "9" now Sep 22 12:53:55 looks like on eth1 C0 -> GND, C1 -> LED1, C2 -> VCC, C3 -> one of (any of) them Sep 22 12:55:10 for eth0 C0 -> GND, C1 -> LED1, C2 -> LED1, C3 -> any Sep 22 12:56:03 so eth0 phy addr is 01000, which is 8 Sep 22 12:56:24 for eth1, phy addr is 11000, which is 24 Sep 22 12:56:44 arachnist: try 24 Sep 22 12:57:05 oh yep, they give exactly those addresses too Sep 22 12:57:40 ok Sep 22 13:02:28 (there should be some kind of distcc for "ld"...) Sep 22 13:06:33 lennert: 24 worked. thanks :) Sep 22 13:07:03 arachnist: cool Sep 22 13:07:11 arachnist: please report that to that guy Sep 22 13:07:27 there's one bit of documentation that's correct then \o/ Sep 22 13:08:37 ron_: thanks for checking the schematics Sep 22 13:09:29 np. I had them right in front of me, so it seemed easy to scroll about and get good answers without guessing too much Sep 22 13:44:45 hm Sep 22 13:45:43 if it wasn't for capacitors C175 and C475 on the openrd-client, it looks like it's be possible to put a 3.5" drive inside the enclosure Sep 22 13:46:08 though heat might be an issue Sep 22 13:46:55 oh Sep 22 13:47:04 and the enclosue's too low Sep 22 13:47:30 * armin76 suggests glueing an external disk Sep 22 13:47:49 or SSD, no? Sep 22 13:47:52 well Sep 22 13:48:02 i have a sata<->esata cable Sep 22 13:48:14 and a external power supply for the 3.5" drive Sep 22 13:48:22 i was just curious Sep 22 13:49:42 (and the enclosure looks just way too cool to dispose of it, imo :> ) Sep 22 13:53:55 you could move those caps to the other side of the board easily enough Sep 22 13:54:03 they're through hole jobbies Sep 22 13:54:34 provided i had any soldiering skills and... what was the "soldiering device" called in english? Sep 22 13:55:39 a soldiering device is probably a rifle, or cannon :) Sep 22 13:56:12 "soldering gun" Sep 22 13:56:26 "iron" is the common word :) Sep 22 13:57:00 iron is the thing that melts and connects, isn't it? Sep 22 13:57:20 (well, at least the common name. not chemically iron) Sep 22 13:57:54 yeah, that's all you'd need to move those though Sep 22 13:59:05 (chemically they are usually mostly copper :) Sep 22 14:27:30 arachnist: openrd support should really go upstream Sep 22 14:28:04 lennert: as in, kernel and u-boot patches? Sep 22 14:28:25 yep Sep 22 14:28:58 so there's one tested version of the code in one pÄ…lace Sep 22 14:29:02 lennert: well, it looked (from the discussion on the google group) that the guy who wrote the patches one of which i needed to modify has submitted them to the kernel devs Sep 22 14:29:20 alright Sep 22 14:29:33 so with that eth1 bug then Sep 22 14:29:59 if it's the ones I'm thinking of, they should be in mainline lk now (and nicolas' orion tree) Sep 22 14:30:41 sending a patch to those guys and/or linux-arm list would be appreciated Sep 22 14:36:56 lennert: dhavel @ einfochips said he was going to upstream the OpenRD Client patches on Sept 17th but nothing has happened Sep 22 14:42:56 arachnist why would you want to use a 3.5 inch drive? They consume more power than a 2.5 inch drive. Also there is 2.5 inch SATA drive header on the board. One of the docs list the spacers needed to mount one. Sep 22 14:45:34 tinker-f595: ok Sep 22 14:45:49 dhaval's been responsive to taking patches, if not to pushing them back out again quickly :) Sep 22 14:46:07 the ones I sent him got pushed through quickly enough though Sep 22 14:46:30 lennert: I was a bit disappointed when that date passed and the code had not be submittted upstream Sep 22 14:47:39 ron_: also be nice to get openrd stuff upstream in uboot Sep 22 14:47:59 yea Sep 22 14:51:04 sure, that's been happening too, prafulla took and pushed some patches from me for that as well Sep 22 14:51:49 there's only so many hours in a day, if you want it to happen faster, the answer should be obvious :) Sep 22 14:52:11 ron_: I have not seen them upstream in the uboot repository. I spent a hour or two reading through it last night Sep 22 14:52:34 ron_: sheevaplug stuff is upstream in the uboot repositiry Sep 22 14:54:07 and dhaval has not put his OpenRD Client patch in his own repository http://repo.or.cz/w/linux-2.6/linux-2.6-openrd.git Sep 22 14:55:19 I've booted prafulla's branch off denx uboot, and sent some patches for it, and got an email from him last week that said he was going to push the next batch to them Sep 22 14:55:49 -client I have no idea about, I just have a somewhat modified -base Sep 22 14:56:06 tinker-f595: one of the reasons being that i already have a 1TB 3.5" sata drive :> Sep 22 14:57:59 arachnist: I have problems attemtping to use eSata. The einfochips uboot has problems with it. Sep 22 14:58:45 tinker-f595: well, i'm booting the system from the internal flash Sep 22 14:59:01 tinker-f595: only /usr will be on the drive, as it didn't fit Sep 22 14:59:13 how much power consumption has openrd with sata disks? someone knows? Sep 22 15:07:51 zumbi depends on the particular disk you are using. Sep 22 15:08:12 tinker-f595: which are your measurements? aprox Sep 22 15:08:49 well depends if the disk is starting, reading, writing or idle Sep 22 15:23:15 tinker-f595: max(power_consumption); Sep 22 15:23:41 tinker-f595: are we talking 5W or 30W or 60W ? Sep 22 15:23:43 as I said it depends on what drive you have Sep 22 15:24:39 tinker-f595: could you tell an aproximated range? Sep 22 15:24:45 zumbi: the openrd will take <6W. a SATA disk could peak at up to 15W, if its a beast. Sep 22 15:25:16 rektide: thanks, so 20W aprox, that's too much for a small solar panel Sep 22 15:26:30 look at 2.5 inch disks...they use much less power than 3.5 Sep 22 15:27:48 depending on your storage requirements & what you're trying to do, a usb flash stick consumes less than a 2.5 Sep 22 15:28:23 a 2.5" will take up to, i dunno, 8w at worst, for a power hungry drive. a usb flash stick will take no more than 1w, i'd think. Sep 22 15:28:36 rektide: i have a 2.5" only USB powered HD, that could do nice low power with sheeva Sep 22 15:29:04 * zumbi was just wondering on how much would be moving to sata Sep 22 15:33:42 rektide never seen a 2,5 inch sata hdd use more than 5 watts at startup Sep 22 15:33:59 idle is less than a watt Sep 22 15:34:33 remember 2,5 inch drives are designed for laptops etc Sep 22 15:51:35 even 7200 rpm drives? Sep 22 15:52:18 i'vee erred on the side of safeness here, but i still wouldnt be surprised by 8w peaks Sep 22 15:57:39 hm Sep 22 15:58:15 too bad the openrd-client doesn't fit in a 5.25" drive bay (like the cdrom drives etc.) Sep 22 15:59:32 (yes, another computer inside a normal computer would be a nice idea imo. provided they had independent power supplies) Sep 22 16:00:56 theres really no reason the openrd has to be so frakking big Sep 22 16:07:09 rektide especially 7200 rpm drives as they use the latest tech Sep 22 16:08:01 how bout 10k rpm drives? ;) Sep 22 16:08:05 rektide: I wish the open rd came with the drive mounting spacers Sep 22 16:08:28 rektide: have you found one that does not vibrate? Sep 22 16:08:54 heh no; i'm not looking, i'm happy with my ms4um 3.5" enclosure Sep 22 16:09:06 rektide: 10k rpm dries are too expensive Sep 22 16:09:17 theyre all enterprise targetted Sep 22 16:09:20 not surprising Sep 22 16:09:32 rektide: you like power hogs ;) Sep 22 16:09:34 it is a bit surprising that they all seem to have such limited capacity Sep 22 16:10:02 well, i'm a bit hellbent on optimizing power consumption Sep 22 16:10:04 just in weird ways Sep 22 16:10:26 well you could go expensive and use an ssd Sep 22 16:10:47 "reliable" ssd are expensive Sep 22 16:10:54 ssds dont make sense yet Sep 22 16:11:19 rektide: notice the quotes around "reliable" Sep 22 16:11:43 the latest ocz's and intels are reliable and moderately priced (comparatively) Sep 22 16:12:07 but in two years, any ssd you buy now will look laughably over expensive, horribly under performant, and woefully under capacity Sep 22 16:13:47 and terribly unreliable Sep 22 16:13:51 rabeeh: around? Sep 22 16:14:22 rektide: can you ping dhavel about the openrd git tree? Sep 22 16:14:38 yeah i'll drop by the newsgroup Sep 22 16:14:42 i'm going to try and push for upstream Sep 22 16:14:56 hey tinker-f595 Sep 22 16:15:07 rektide: thanks I don't want to ping him about it too many times in a row Sep 22 16:15:36 morfic: I am heading off to work in a minute Sep 22 16:15:38 yup, i understand 100% Sep 22 16:15:54 raid1 resyncing plus moving data from 3rd drive to raid makes for a nice 6.something load :) Sep 22 16:16:02 rektide: remind him about the openrd client code as well and uboot please Sep 22 16:16:12 good plans Sep 22 16:16:17 morifc ??? Sep 22 16:16:43 ciao for now Sep 22 16:16:46 tinker-f595: why the ??? do you expect less with a saturated usb bus? Sep 22 16:17:26 morfic wasn't sure of the context of the 6. Sep 22 16:17:34 no units Sep 22 16:18:02 tinker-f595: don't be late for work :) Sep 22 16:18:27 "sory boss, this guy in #openplug was so confusing"... not a good excuse ;) Sep 22 16:21:43 i have yet to run Mono on my openrd Sep 22 16:21:46 i should really do that Sep 22 16:21:59 soon, considering how much code i'm developing specifically for that platform Sep 22 16:22:19 morfic: haha Sep 22 16:22:51 rektide: mono hmmm Sep 22 16:23:09 i've been porting ioctl's and input.h to mono Sep 22 16:23:44 "fun" Sep 22 16:32:44 hm Sep 22 16:33:15 seems like the esata controller on the openrd is a little slower than the sata one in my pc Sep 22 16:33:18 4194304000 bytes (4.2 GB) copied, 48.4195 s, 86.6 MB/s Sep 22 16:33:42 with my pc i get around 100MB/s for reads and writes Sep 22 16:35:46 hello all Sep 22 16:36:08 I built 2.6.31 for the sheevaplug, and hell, the default config didn't have initrd enabled :) Sep 22 16:36:26 About u-boot... I built it with the default sheevaplug config, and I find many feataures missing Sep 22 16:36:31 viric: :D Sep 22 16:36:32 for example, the tab-completion Sep 22 16:36:45 * viric rebuilds 2.6.31 Sep 22 16:36:54 viric: what do you need initrd for, really? :> Sep 22 16:37:12 I'm porting an OS that cares on an initrd Sep 22 16:37:28 And I don't want to throw their work away on that part :) Sep 22 16:37:38 which one? Sep 22 16:37:41 NixOS Sep 22 16:38:07 I don't have any heavier reason than that. Sep 22 16:38:10 arachnist: sata 1.5 or 3? Sep 22 16:39:17 well, it got detected as a 1.5 Sep 22 16:39:19 2009-09-22T16:24:53.704016+02:00 stalker kernel: ata2: SATA link up 1.5 Gbps (SStatus 113 SControl F300) Sep 22 16:40:56 but 1.5gbps =~ 150MB/s Sep 22 16:45:16 cool, switched a jumper on the drive, and now it's 3.0 Sep 22 16:45:17 2009-09-22T18:44:38.964414+02:00 stalker kernel: ata2: SATA link up 3.0 Gbps (SStatus 123 SControl F300) Sep 22 16:46:01 but it didn't change the results at all. anyway, i can live with 80MB/s Sep 22 16:46:32 how do you test that speed? Sep 22 16:46:39 sheeva got no "speaker" to beep with? Sep 22 16:46:59 viric: dd if=/dev/sde1 of=/dev/null bs=10M count=40 Sep 22 16:47:00 looks like he used dd Sep 22 16:47:30 ok Sep 22 16:47:42 try with bs=8k and appropriate count, or better yet try tiobench and bonnie++ Sep 22 16:48:06 you can also test from /dev/zero to /dev/sde1 Sep 22 16:48:10 it's still about as fast as the ssd in my laptop and a little bit faster than the ata drive i left in my desktop Sep 22 16:48:37 morfic: from /dev/zero to /dev/sde1 would cause a disaster ;) Sep 22 16:49:04 arachnist: well, i would mount it and write to a file Sep 22 16:49:20 more interesting that the viric: dd if=/dev/sde1 of=/dev/null bs=10M count=40 proposed Sep 22 16:50:39 well Sep 22 16:50:40 dd if=/dev/sde1 of=/dev/null bs=8k count=51200 Sep 22 16:50:45 419430400 bytes (419 MB) copied, 5.19316 s, 80.8 MB/s Sep 22 16:51:22 how fast are writes on that? Sep 22 16:51:54 dd if=/dev/zero of=/mnt/test.img bs=8k count=51200 Sep 22 16:51:58 419430400 bytes (419 MB) copied, 4.84035 s, 86.7 MB/s Sep 22 16:52:25 reasonable speeds, i'd say Sep 22 16:54:29 /mnt is a 30GB ext3 partition, offset about 10GB from the disk start Sep 22 16:56:59 eewww, lost drives during rsync/resync Sep 22 16:57:05 going back to 2.6.30.6 Sep 22 16:57:21 arachnist: nice speeds, yeah Sep 22 16:57:42 at least i hope it's a kernel issue Sep 22 16:58:14 morfic: back ?where you were? Sep 22 16:59:43 i was on .30.6 but figured when i reboot it anyway might as well boot 2.6.31, but dmesg on that turned me off, so i went to 2.6.30.7 but with that i had lost drives twice now, first i thought it was a powerblip since got some good weather here, but now, i see drives just went bye Sep 22 16:59:58 ? Sep 22 17:00:04 end_request: I/O error, dev sda, sector 963359 Sep 22 17:00:05 Kernel panic - not syncing: Attempted to kill init! Sep 22 17:00:05 now in 2.6.30.6? Sep 22 17:03:42 I'll try 2.6.31 Sep 22 17:03:44 in a little bit Sep 22 17:03:58 booting from nand? Sep 22 17:04:56 no Sep 22 17:05:02 SD by now Sep 22 17:06:52 let me know what your dmesg looks like :) Sep 22 17:07:09 .6 has the only clean dmesg Sep 22 17:07:28 .7 still muttered something about mtdsomething being bad, lesson learned Sep 22 17:15:11 oh nice it oopsed even on .6 now Sep 22 17:19:33 I booted a bit... Sep 22 17:19:36 not perfectly. Sep 22 17:19:43 about mtd? let me see Sep 22 17:19:58 mtd: partition size too small (0) Sep 22 17:19:58 Creating 3 MTD partitions on "orion_nand": Sep 22 17:19:58 0x000000000000-0x000000100000 : "u-boot" Sep 22 17:19:58 0x000000100000-0x000000500000 : "uImage" Sep 22 17:19:58 0x000000500000-0x000020000000 : "root" Sep 22 17:20:31 NAND device: Manufacturer ID: 0xad, Chip ID: 0xdc (Hynix NAND 512MiB 3,3V 8-bit) Sep 22 17:20:34 Scanning device for bad blocks Sep 22 17:20:37 there are some 'bad eraseblocks' Sep 22 17:27:10 2.6.31 works fine here. Sep 22 17:59:01 http://wstaw.org/images/free/2009/09/22/1588fec65934ed53c7515b0f7c27d8.jpeg Sep 22 17:59:23 arachnist: my seagate 500g drives look exactly like that Sep 22 18:02:36 i think i'll need to get some stands for that drive, and possibly a cooling fan :/ Sep 22 18:03:08 (because of the heat) Sep 22 18:03:23 my old samsung 500g drives (4 platters?) get hot too Sep 22 18:03:34 i suspect new 500g drives wouldn't get so hot Sep 22 18:06:57 well Sep 22 18:07:18 that's a 4-platter 1TB drive (not sure if it's 4-platter) Sep 22 20:36:50 can I test the sheeva's memory? I want to know what the segfault was all about mounting md0 earlier. Sep 22 20:38:28 never built the memtest bit in kernel. Sep 22 20:49:15 morficmobile: uboot has a memtest Sep 22 20:49:25 iirc Sep 22 21:58:22 Great, sheevaplug works great finally Sep 22 21:58:39 I'm quite happy! :)) Sep 22 23:23:23 http://en.wikipedia.org/wiki/File:SheevaPlug_with_external_drive_enclosure.jpg Sep 22 23:23:30 Have they moved the usb port on the newer revs? Sep 22 23:25:37 Reedy: why, where is yours? Sep 22 23:25:51 much closer to the network port Sep 22 23:25:58 ie right next to it Sep 22 23:26:22 on that, the usb port and network are further apart Sep 22 23:26:33 usb is probably same place, but network moved Sep 22 23:27:55 mine has a gap like that picture Sep 22 23:28:41 hmm Sep 22 23:30:37 i'd upload a picture if i wasnt about to go to beed Sep 22 23:30:46 http://plugcomputer.org/plugforum/index.php?topic=771.0 Sep 22 23:30:53 posted it in the forums **** ENDING LOGGING AT Wed Sep 23 02:59:58 2009