Re: Thecus N2100 - problems after flashing
On Tue, Jan 08, 2008 at 06:45:04PM +, Martin Guy wrote: > 2008/1/8, Salvatore Iovene <[EMAIL PROTECTED]>: > > I booted it up with no HD inside, the FW was 2.1.0. Silly me, I decided > > that I didn't need to upgrade to 2.1.5 to have an IP address in RedBoot, > > > > Is there anything left to do or do I just have to start soldering? > > You can still talk to redboot via the serial port. If, like mine, the > serial port connector was not supplied you can still use it without > soldering if you wrench 3 serial port-type square gold pins off a dead > motherboard, clean them up and poke the soldery ends into pins 2, 3 > and 5 of the right type of DB9-to-2x5 connector: the sort where the > wires are awkwardly crossed over inside the plug shell so that pins > numbered 1-9 go to pins 1-9. You can then fit this into the serial > connector holes on the HD connector riser board without having to > solder anything. > > Good luck! Hi, and thanks for the answer. After all everything turned out to be fine. It seems that the Thecus leased a DHCP address after all, and I just insisted on probin 192.168.1.100. :) Now I'm in the middle of the installation. -- Salvatore Iovene http://www.iovene.com/ signature.asc Description: Digital signature
Problem with the 2.6.22-3-arm lenny kernel?
Hi I've run debian on my nslu2 for about 18 months with no problems (ever since the first debian installer rc for nslu2, infact). I've been looking to create a more robust to power failure solution by using a small thumb drive. To do so I used the latest etch installer (4.0r2?) to install clean (I like clean installs) and wanted to bump to lenny for the newer bacula there. Unfortunately the lenny kernel, 2.6.22-3-arm, doesn't seem to want to boot. The etch kernel is just fine, so I'm running an etch kernel with a lenny userspace, but I'd like to check out the latest kernel stuff too. When I flash the lenny kernel, the linksys does the usual apex lightscroll, then the status light flashes briefly red, before going yellow and then nothing more. (The ethernet light is the only one on). The USB stick is obviously not being found, (no Disk1 light, the light on the key doesn't come on). I've waited for about 1 hour for it to boot, but it's obvious the usb storage modules aren't being loaded. I've looked at the initramfs, and it seems that everything is in order. (Note that to switch back to etch from lenny is rather involved- i have to upslug the installer, fix the installer (it's getting some really weird network parameters from somewhere and ignoring dhcp), boot the installer to rescue mode, and finally, mount the root, and flash the etch kernel. Anyway, with the etch kernel, everything behaves fine (with the rootdelay parameter for LVM). Thoughts, suggestions? Thanks Christian -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Re: NSLU2 Debian Installer Erroneous asking for an IP address
So, basically even though the user has specified a static IP, the unit DHCP's anyway. Considering that a well-behaved DHCP server will probe the active addresses, it's *guaranteed* that not even by accident will the NSLU2 get the static IP that the user set. Hence the device is "lost on the ether". Debugging this situation is difficult for the novice. First, the user set a static IP -- there's no reason for them to go check their router to see if it issued a DHCP IP in the first place; they're not expecting that to happen. Secondly, many routers don't even provide a means to check what DHCP has done, so the user can't discover the IP even if a wiki or document told them to do so (I believe that Linksys, one of the most common routers in this area, is one such vendor). Personally, I find this behavior of the installer to be wrong, in the same way that I would be angry if my automobile took it upon itself to turn the steering wheel for me, because I happened to leave the turn-signal activated for too long. But I'm not a Debian user (I just happen to frequent the #nslu2-general IRC channel where this issue has become so commonly asked). IMO, if this behavior is retained, it needs a gigantic red box (flashing, preferably) on the web pages describing the installation process. Many of the users I encounter on that IRC channel are truely novices, so the text also should not just limit itself to outlining the behavior, but the implications of it as well (that the unit will DHCP and that many users may not have routers that offer the ability to see what the DHCP IP might have been, resulting in an NLSU2 on the network that is well and truely lost). And, no, running nmap to find it is not an option for most of these users! Regards, Mike (mwester) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Thecus N2100 - problems after flashing
Hi all, I just received a Thecus N2100. Here's a report of what I did, so that now I can't access the box. I booted it up with no HD inside, the FW was 2.1.0. Silly me, I decided that I didn't need to upgrade to 2.1.5 to have an IP address in RedBoot, and started flashing the debian installer's image. According to the webpage, the flashing went thru successfully, then I clicked on the Reboot button and waited 5 minutes. I tried to ping 192.168.1.100 (from my desktop: 192.168.1.150), but got no response. I pinged broadcast, and nothing. I nmapped: nothing. Notice that the original FW's web interface had the default GW set to a wrong address, but that shouldn't matter: I should still be able to ping it. After this, I put in a SATA disk on the slot 1, and rebooted, just in case. Still nothing. Is there anything left to do or do I just have to start soldering? Thanks in advance, Salvatore. -- Salvatore Iovene http://www.iovene.com/ signature.asc Description: Digital signature
Re: armel packages for linux-2.6.23-2 and install problem
Long agonising story about a Thecus brick and request for info on
increasing ramdisk size at boot
2008/1/7, Martin Michlmayr <[EMAIL PROTECTED]>:
> Did you modify the MTD partitions?
Oops. Make that a yes. I couldn't use the Thecus web interface, which
needs flash player, so I had loaded kernel and initrd in RedBoot with
load -v -r -b 0x80 -m http -h 192.168.1.1 /.../armel-installer-initrd.gz
fis create ramdisk
load -v -r -b 0x20 -m http -h 192.168.1.1 /.../vmlinuz-2.6.21-1-iop32x
fis create kernel
exec -c "console=ttyS0,115200 root=/dev/ram0 initrd=0xa080,42M mem"
without realising that that truncated the ramdisk partition size too.
Fixed in RedBoot with:
fis load ramdisk# at 0x8
fis create ramdisk -l 0xd0
That let it boot the old kernel and ramdisk, and the dpkg -i of the
new kernel package worked, but booting it failed because, although the
kernel flash partition was big enough, it was only loading the new
kernel up to the length of the old kernel:
RedBoot> exec -c "console=ttyS0,115200 root=/dev/ram0 initrd=0xa080,42M"
Build ATAG
ATAG_MEM: Overwrite ram_end with real_region_top=0x2000, memsize=512 M
[EMAIL PROTECTED], MACH_TYPE=1101
Using base address 0x0020 and length 0x0012fbd4
Fortunately I had backed up the mtd contents somewhere (but not the
new kernel and ramdisk images, managgia!) so I've restored those:
RedBoot> load -v -r -b 0x20 -m http -h 192.168.1.1 /martin/tech/n2100/debian
/mtd/mtdblock2
RedBoot> fis create kernel
RedBoot> load -v -r -b 0x80 -m http -h 192.168.1.1 /martin/tech/n2100/debian
/mtd/mtdblock1
RedBoot> fis create ramdisk -l 0xd0 -s 0x3e
and it now uncompresses and boots the old kernel again but clags saying:
RAMDISK: Compressed image found at block 0
RAMDISK: incomplete write (-28 != 32768) 8388608
List of all partitions:
1f00256 mtdblock0 (driver?)
1f01 13312 mtdblock1 (driver?)
1f02 1408 mtdblock2 (driver?)
1f03 1152 mtdblock3 (driver?)
1f04 4 mtdblock4 (driver?)
1f05128 mtdblock5 (driver?)
No filesystem could mount root, tried: cramfs
Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(1,0)
That seems to be because the initrd is set to 8MB and
"gzip -d < mtdblock/mtdblock1 | wc -c" says it's 9093120 bytes.
(gzip: stdin: decompression OK, trailing garbage ignored)
The kernel that's booting is 2.6.22-3, same as when it worked, and the
ramdisk contains the right image
("gzip -d < mtdblock1 | cpio -i -t"
lists lib/modules/2.6.22-3-iop32x/...)
and those mtdblock contents were taken from it when it booted ok before.
I've tried
RedBoot> fis load ramdisk
RedBoot> fis load kernel
RedBoot> exec -c "console=ttyS0,115200 root=/dev/ram0 initrd=0xa080,42M"
which I presume should set the ramdisk size, but it seems not to.
I'm stumped now. Anyone know how to up the kernel ramdisk size from
the bootparams?
The ",42M" above" doesn't do it, nor does the ramdisk_size=16384"
mentioned in "man bootparam" - all die saying "RAMDISK: incomplete
write (-28 != 32768) 8388608"
(and errno 28 is "No space left on device")
M
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: armel packages for linux-2.6.23-2 and install problem
* Colin Tuckley <[EMAIL PROTECTED]> [2008-01-08 11:54]: > Mine has: > [EMAIL PROTECTED]:~$ cat /proc/mtd > dev:size erasesize name > mtd0: 0004 0002 "RedBoot" > mtd1: 00d0 0002 "ramdisk" Yep, that's the same I have. -- Martin Michlmayr http://www.cyrius.com/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: armel packages for linux-2.6.23-2 and install problem
Martin Guy wrote: > Nope. The usual debian installer run on a from-the-factory "Yes Box" N2100. > Would someone with an N2100 forward the output of "cat /proc/mtd" please? > I have > mtd0: 0004 0002 "RedBoot" > mtd1: 003e 0002 "ramdisk" > mtd2: 0016 0002 "kernel" > mtd3: 0012 0002 "user" > mtd4: 1000 0002 "RedBoot config" > mtd5: 0002 0002 "FIS directory" Mine has: [EMAIL PROTECTED]:~$ cat /proc/mtd dev:size erasesize name mtd0: 0004 0002 "RedBoot" mtd1: 00d0 0002 "ramdisk" mtd2: 0016 0002 "kernel" mtd3: 0012 0002 "user" mtd4: 1000 0002 "RedBoot config" mtd5: 0002 0002 "FIS directory" Colin -- Colin Tuckley | +44(0)1903 236872 | PGP/GnuPG Key Id Debian Developer | +44(0)7799 143369 | 0x1B3045CE "Waiter, there's no fly in my soup!" -- Kermit the frog -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: armel packages for linux-2.6.23-2 and install problem
> > In fact, the N2100 nand initrd partition has size 0x003e or 4063232 > > bytes > > Did you modify the MTD partitions? > > Every N2100 I've seen so far has a ~13 MB ramdisk partition. Nope. The usual debian installer run on a from-the-factory "Yes Box" N2100. Would someone with an N2100 forward the output of "cat /proc/mtd" please? I have mtd0: 0004 0002 "RedBoot" mtd1: 003e 0002 "ramdisk" mtd2: 0016 0002 "kernel" mtd3: 0012 0002 "user" mtd4: 1000 0002 "RedBoot config" mtd5: 0002 0002 "FIS directory" Cheers M -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
