Need help troubleshooting GFH that suddenly stopped booting from a SATA drive

So I have a GFH with a 3TB SATA drive that's been working like a champ for years now, running transmission and serving up files via samba. Awesome little device actually.

Anyway, I came back one day and it was off the network. So I rebooted it and it came up from in the rescue system, after flashing orange (for led_error) during the uboot process. So I fired up netconsole for the next reboot and see that it's complaining about a bad partition during sata_load_uimage:

Reset IDE: BUS 0: .......................................OK Bus1: not available
** Bad partition 1**
** Bad partition 1**
Wrong Image Format for bootm command
ERROR: can't get kernel image!

bootcmd_sata=run sata_init; run sata_load_uimage; run set_bootargs_sata; run sata_boot;

console=ttyS0,115200

mtdparts=mtdparts=orion_nand:1M(u-boot),4M(uImage),32M(rootfs),-(data)

set_bootargs_sata=setenv bootargs console=$console root=$sata_root rootdelay=$sata_rootdelay rootfstype=$sata_rootfstype $mtdparts

sata_boot=if ext2load ide $sata_device 0x1100000 /boot/uInitrd; then bootm 0x800000 0x1100000;else bootm 0x800000;fi;
sata_device=0:1
sata_init=ide reset
sata_load_uimage=mw 0x800000 0 1; ext2load ide $sata_device 0x800000 /boot/uImage
sata_root=/dev/sda1
sata_rootdelay=8
sata_rootfstype=ext3

I have full control over uboot and can boot into some older version of rescue or I assume from USB.

I'm wondering if I can investigate / fix the bad partition from rescue, or if i'll need to use some usb stick.
IIRC I have a 20gb ext3 linux partition and then the rest of the 3tb is an ext4 partition (might be ext3) with all my data.
should i be aiming to run fsck on that boot partition? is there anything else I can try?

I'm pretty sure that whatever happened to cause whatever problem i'm experiencing happened while the system was powered up and running as normal. Pretty sure there was no power outage or reboot or anything, so I'm wondering it the sata drive itself may be the issue.
One thing that makes me think that might be the case is that about half the time i run "ide reset" this happens:
Now that I think about it I'm not sure if the command always works the first time but then doesn't necessarily work after that, or it's really more random ... too tired to double check now :)

GoFlexNet> ide reset
Reset IDE: Bus0: ....................................** Timeout **

Anyway, any advice would be great. I've actually been thinking about trying to upgrade to the latest kernel and rescue system, because whatever I have is a year or two old, so maybe this can be a good opportunity to make that happen.
Definitely gotta get this thing runing again because it's been the perfect torrent machine for me, especially because of how little power it uses!!!

floodo1,

You could do fsck from inside rescue system to see if there is any error. If you have trouble doing that, then download the latest rootfs. Extract it a USB thumb drive formatted as Ext3 and plug that in the GFHome.

Power up, and stop u-boot at netconsole. Change the bootcmd to boot USB. You also need to change usb_root env, too, so that the correct drive is selected for rootfs (otherwise the rootfs on the HDD will be selected and you will not be able to fsck that). The HDD will be assigned sda1, so the USB thumb will be sdb1.

The 2 or 3TB drives from either Seagate or Western Digital are going to give you problem like this once in a while. It spins up so slowly that I've seen timeouts very often. However, my 3 GFHomes as media NAS servers boot with USB thumb Ext3 rootfs, so they never fail to boot. Once the GFHome boots into Linux, the HDD drive detection is solid (the driver is more robust than u-boot's driver). And it will help to update the kernel (it's 3.16.x now).

Also, it will help to update u-boot to the latest version so you don't have to use hybrid for 3TB HDD.

-bodhi
===========================
Forum Wiki
bodhi's corner (buy bodhi a beer)



Edited 1 time(s). Last edit at 09/06/2014 03:12AM by bodhi.

When I'm in rescue there are no sd* devices under dev. So I'm not sure how to run fsck on the drive if I can't just "fsck /dev/sda1"

Sounds like my best bet is to us a USB thumb drive because I get a full linux system that way. I'll try that now, because I have a thumb drive and uboot control

@Floodo1 the version of rescue I use doesn't have SATA support, so yes you are right a flash drive will be required.

As bodhi said, it might just be as simple as a fsck :D

Hope you get it sorted ...

@bodhi, will have a look at that new u-boot when I get a chance, was wondering about > 2tb drives ... thanks for posting :D

Cheers

DC