Hi,

Has anyone tried out the 250 GB drives from Western Digital (like the WD2500BEVE)? I have generally used the 120GB Toshiba or Hitachi in sets of 2.

While reading comments about drive reliability (Newegg), the Western Digital comments were generally not complimetary.

Were there any difficulty in using the Bigdisk builder during the install or was it as silky smooth as I experienced with my multiple Empegs and 120GB drives?

Any recommendations?

Thanks,

Ross

ISTR you'll definitely need the hijack with LBA-48 support - the 120GB drives are under this limit and so don't need it.

Apart from that... good luck...!

Hugo

Yeah. My BigDisk builder images include Hijack for the LBA48 support, and should work (with a tad of luck).

The real question becomes, what new player software limitations will be encountered with these..

Hi,

Guess who's trying to install 250 GB drives?

Right, having a little trouble with it.


I have followed (I think) the big disk build in the FAQ.

I used your big disk builder (created 09/28/2006) that I had great success with ror multiple 120 GB drives.

This formatted (I assume it also partitions it too), the drive and stressed the disk successfully.

I then used the new hijack car2_v2.01_hijack (created 06/23/07) as directed in the FAQ.

After the reboot, it found the drive but not the file system (see attached). This occurred after several attempts.

==============
Start of list
==============

e000 v1.04for help):
Copying kernel...
Calling linux kernel...
ide_dat

Uncompressing Linux..................................... done, booting the
kernertition (1-4)080
p
Partition number (1-4): 2st
l.ro
Linux version 2.2.17-rmk5-np17-empeg52-hijack-v464 hijack@rtr.ca) (gcc
versionder (6-16709, default 6)

Command action
2.95.3 20010315 (release)) #2 Fri Oct 27 01:00:29 EDT 20060 36): Poll
result=0 on
Processor: Intel StrongARM-1100 revision 11 (1-4)

p
Par
Checking for extra DRAM:e 0xaaaa read 0x0080sp
c1000000: wrote ffffffff, read e28cc001
First cylinder (11-16709, defa
NetWinder Floating Point Emulator V0.94.1 (c) 1998 Corel Computer Corp.555
read 0x0080
Last cylinder or
empeg-car player (hardware revision 9, serial number 4010385

Page cache hash table entries: 4096 (order 2, 16k)ault
16709):0x000-0x007,0x038 on irq 6
POSIX conformance testing by UNIFIX

Linux NET4.0 for Linux 2.2
Using default value 167
Based upon Swansea University Computer Society NET3.039

Command (m for help): n
NET4: Linux TCP/IP 1.0 for NET4.0, default 1):
IP Protocols: ICMP, UDP, TCP212H9AT00,
TCP: Hash tables configured (ehash 16384 bhas


Hex code (type L to l
ttyS02 at 0xf8030000 (irq = 16) is a SA1100 UART
hda: hda1 < hda5 hda6 > hda2
Signature is 636f6972 'rioc'pe of partition 6 to 82 (Lin
Found custom animation at offset 0x9c388

Command (m for hel
Tuner: loopback=0, ID=-1db2
Scheduling custom logo.

Disk /dev/hda: 255 h
empeg display initialised.indersSK: ext2 filesystem
empeg dsp audio initialised
empeg dsp mixer initialised
empeg dsp initialised= cylinders of 16065
empeg single channel IDE

Calling ioctl() to r
Probing primary interface... music partition read-onlypo
ide_data_test: wrote 0x0000 read 0x8080

ide_data_test: wrote 0xffff read 0x8080th error 16: Device or resource busy.

ide_data_test: wrote 0xaaaa read 0x8080
Pr

ide_data_test: wrote 0x5555 read 0x8080page for additionalmezone:
Atlantic/Sou
ide_data_test: wrote 0x0000 read 0x8080

ide_data_test: wrote 0xffff read 0x8080t determine runlevel - doing soft
reboo
ide_data_test: wrote 0xaaaa read 0x8080

ide_data_test: wrote 0x5555 read 0x8080 instead of reboot from the command
lin
ide_data_test: wrote 0x0000 read 0x8080

ide_data_test: wrote 0xffff read 0x8080: cannot open /var/run/shutdown.pid

ide_data_test: wrote 0xaaaa read 0x8080
kftpd: listening on por
ide_data_test: wrote 0x5555 read 0x808ay.g dsp
ide_data_test: wrote 0x5555 read 0xdbfb -a: use
/etc/shutdown.allow...
hda: WDC WD2500BEVE-00WZT0, ATA DISK drive4231

ide0 at 0x000-0x007,0x038 on irq 6

hda: WDC WD2500BEVE-00WZT0, 238475MB w/8192kB Cache, CHS=30401/255/63, LBA48

-r: reboot afte
empeg-flash driver initialized
smc chip id/revision 0x3349 -n: do n
smc9194.c:v0.12 03/06/96 by Erik Stahlman (erik@vt.edu)


SMC9194: SMC91C94(r:9) at 0x4008000 IRQ:7 INTF:TP MEM:6144b MAC
00:02:d7:28:0f:0 -c: cancel a running shutdo

c
RAMDISK: Loading 320 blocks [1 disk] into ram disk... done. **
the "time" argument is mandatory! (try "now
VFS: Mounted root (ext2 filesystem).
empeg-pump v0.03 (19980601)
Press Ctrl-A to enter pump...VFS: Mounted root (ext2 filesystem) readonly.
change_root: old root has d_count=1
Trying to unmount old root ... okay
Freeing unused kernel memory: 4k initempeg init 0.8
I see this is a developer image!
Mounting proc
Mounting first music partition
Tried to mount /dev/hda4 as reiserfs but got error 19
VFS: Can't find an ext2 filesystem on dev ide0(3,4).
Tried to mount /dev/hda4 as ext2 but got error 22
Error mounting partitions (possibly already mounted)
Remounting first music partition read-only
No primary hard disk
Remounting second music partition read-only
No secondary hard disk
Press 'q' now to go into development mode. You Have Zero Seconds To
ComplùStarti
ng player
Timezone: US/Pacific
player.cpp : 385:empeg-car 2.01 2004/07/06.
! tags.cpp : 61:Failed to open tags (0xc0041002).
Prolux 4 empeg car - 2.1434 Jul 5 2004
Vcb: 0x4086d000

==============
End of list
==============


I also tried to partition it manually from another FAQ, which seemed to work, but I had trouble with formatting the drive. I eventually scribbled on the boot partition and started over with your programs.


Here's where I get embarrassed now, what am I missing? I know, I know you warned us all that there might be problems with this size drive...

Have others been successful with 250 GB drives?

Got any ideas?

Thanks,

Ross

You said you used:

- Builder (bigdisk)
- Hijack (latest)

Somewhere after using the builder, you're supposed to put on the actual *player software*, which is different from the builder. The serial output seems to indicate you did do that, but you didn't say which one you used. Anyway, Mark made a bigdisk version of the player software installation, make sure you used that.

It really looks as if the kernel currently installed isn't a (newish) Hijack one -- except that it *is*.

I'll look around and see if I can find a 250-300GB IDE drive here to connect up, and see if the LBA48 support works for me here with that.

I don't think it's been tested on anything other than a 160GB drive when I first implemented/tested it here.

Does anyone out there know of any empeg with a 160GB or larger drive installed ????

Cheers

Ross,

Connect up a serial cable, and hit Control^C to get a command prompt on the empeg. Then do this:

cat /proc/partitions

and this:

fdisk -l

And post the outputs of those.

Thanks

Hi,

Yes, I followed Mark's new instructions in the FAQ.

I installed hijack 488 prior to any install as the instructions indicate. Then the Big Disk Builder, then the player hijack that was recommended.

I thought the player software was the car2.v2.01_hijack.

I have tried this on 2 different known good Rio's and 2 different disk drive cables. I have completed the above install and tried v2b13 Hijack as well.

I started with a new 250 GB drive.

Here is the result of the above requests:

<CTRL C>

Restored terminal settings
Remounting first music partition read-only
No primary hard disk
Remounting second music partition read-only
No secondary hard disk
Abnormal player termination
Player received SIGINT, user interruption
Switching to shell-player loop
Starting bash.
empeg:/empeg/bin# cat /proc/partitions
major minor #blocks name

3 0 134217727 hda
3 1 1 hda1
3 2 40162 hda2
3 3 24097 hda3
3 4 134110620 hda4
3 5 24034 hda5
3 6 16033 hda6
empeg:/empeg/bin#
empeg:/empeg/bin# fdisk -l

Usage: fdisk [-l] [-b SSZ] [-u] device
E.g.: fdisk /dev/hda (for the first IDE disk)
or: fdisk /dev/sdc (for the third SCSI disk)
or: fdisk /dev/eda (for the first PS/2 ESDI drive)
or: fdisk /dev/rd/c0d0 or: fdisk /dev/ida/c0d0 (for RAID devices)
...

I assume this is what you really wanted.


empeg:/empeg/bin# fdisk /dev/hda -l

Disk /dev/hda: 255 heads, 63 sectors, 16709 cylinders
Units = cylinders of 16065 * 512 bytes

Device Boot Start End Blocks Id System
/dev/hda1 1 5 40131 5 Extended
/dev/hda2 6 10 40162+ 83 Linux
/dev/hda3 11 13 24097+ 10 OPUS
/dev/hda4 14 16709 134110620 83 Linux
/dev/hda5 1 3 24034+ 83 Linux
/dev/hda6 4 5 16033+ 82 Linux swap
empeg:/empeg/bin#


These were the same partition assignments that I got when I manually partitioned the drive too.

By the way, if anyone else is trying to upgrade to these drives, they seem to be in short supply.

Thanks for the help,

Ross

hi,

One other thing that might be worth mentioning is that when the player powers up, it looks like a blank music partition as in "End of playlist" and obviously, emplode doesn't see the player although the PC host recognizes the USB port cable being installed.

Thanks,

Ross

Yeah. The log messages indicate that the kernel doesn't see a formatted music partition on /dev/hda4. But the builder ought to have created/formatted that, so something is amiss.

No worries, though. We do have the expertise in house for this sort of thing.

At some point soon, I'll dig out a suitably large drive and run through the motions with it on one of the empegs here. And then fix whatever might be wrong.

It may take a week or three though, so do be patient.

But not too patient, as my memory isn't the greatest.

Cheers

Hi,

I'm not in a real hurry, and appreciate your "in-house" expertise.

Thanks,

Ross

Hi,

Some additional notes about this thread.

I was able to see the 250 GB drive (244 GB after partition & format) on the EMPEG Vital Signs, Boot log, and partition data with fdisk and cat.

I was able to manually partition & format the drive and using v2b13 in the Empeg and emplode would see the drive as 128 GB (no VBA-48). It showed up on the EMPEG Vital Signs, Partition table, and in emplode on USB.

When I manually partitioned and formatted the drive for 250 GB, it would show 244 GB on the EMPEG Vital Signs, the correct Partition Table (same as the big builder does), and emplode would see the player on USB. Emplode completed the file system check, and when rebuilding the database, the EMPEG rebooted and emplode crashed (asking the familiar "Do you want to send this crash dump to Microsoft for analysis" message). Emplode did however play music "ta-da" like it completed the process at the same time emplode crashed.

Does emplode need to be fixed too?

Ross

From what you are saying, maybe it does. Tried jemplode?

Do keep in mind, please, that *anything* you do with a drive larger than 128GB installed in an empeg/Rio car unit is meaningless unless Hijack is installed from start to finish.

Which I'm sure you've made a point of doing, but it does bear repeating here. Anything that writes to the disk (fdisk, emplode, database rebuild..) without a Hijack kernel will corrupt the drive.

Cheers

Hi,

Yes, I understood that. I just wanted to make sure the drive could be addresses within the limits of the available software.

It can be.

It was just a warm and fuzzy that the drive electronics were not the issue.

Thanks,

Ross

Hi,

Haven't tried jemplode yet. Wanted to get everything else correct with the partitioning (I think it looks good - not a software guy though), formatting, and the file system first.

Thanks,

Ross

In a general sense, yes it does. The odds of that happening are about as slim as a slim person who's lost a lot of weight recently.

In this specific case, I can't think of a reason why emplode would crash just because you've got that much space available. If it's reproducible, use JEmplode. If not, it's just one of those things.

Hi,

It's very reproducible. I was able to have it happen 3 times after each new partition, format, & system load.

I guess I need to load jemplode and see what it does.

Ross

I'm taking a short break from tweaking my Gray-Hoverman to begin diagnosis of the super large drive issue that Ross has reported here.

There's a 320GB desktop drive hooked up to beater here now, running the bigdisk builder. I'll follow it all the way through loading music and see if any problems can be identified for fixing.

-ml

Yep. The /dev/hda4 partition never got formatted. Something's wrong with the bigdisk_builder here. Still digging..

Mmm.. the /empeg/bin/scrawl command craps out for me here, causing the build script to abort:

empeg:/# /bin/make-drives
Setting up swapspace version 0, size = 24670208 bytes
Adding Swap: 24092k swap-space (priority -2)
Making first drive...
scrawl(91): user memory violation at pc=0x020046e0, lr=0x02004608 (bad address=0x4014a000, code 2)
pc : [<020046e0>] lr : [<02004608>]
sp : bffffd04 ip : bffffd04 fp : bffffd1c
r10: 40145128 r9 : 02002bc4 r8 : 02020538
r7 : 0200242c r6 : 4000c2a8 r5 : 0202aae0 r4 : 0202b5c8
r3 : 4014a000 r2 : f00f1bb1 r1 : 00000003 r0 : 00000001
Flags: nZCv IRQs on FIQs on Mode USER_32 Segment user
Control: C073D17D Table: C073D17D DAC: 00000015
Function entered at [<020045ec>] from [<02004c64>]
r4 = 0202B5C8
Function entered at [<02004c44>] from [<02004b88>]
r4 = BFFFFD70
Function entered at [<02004968>] from [<020048d8>]
Function entered at [<02004798>] from [<02004200>]
r4 = 020041EC
Function entered at [<020041ec>] from [<02020500>]
r4 = 020041EC
Function entered at [<020204dc>] from [<02002440>]
r5 = 00000002 r4 = BFFFFE04
Function entered at [<0200242c>] from [<4007cfec>]
Function entered at [<4007cee4>] from [<02002b38>]
r10 = 4001D858 r9 = 00000000 r8 = 00000000 r7 = 00000000
r6 = 02002B14 r5 = 00000000 r4 = 4001E5EC
/bin/make-drives: line 1: 91 Segmentation fault /empeg/bin/scrawl "Making $1 filesystem..."


I believe scrawl is just a program to dump progress messages to the front panel display, so we can nuke it, or write an Open Source replacement for it that actually works.

-ml

Here's my replacement for /empeg/bin/scrawl:

#!/bin/sh
echo "POPUP 9999 $*" > /proc/empeg_notify

Requires a Hijack kernel, of course, but we already have that in the new builders.

I'm still retesting things here. I'll post a new builder_bigdisk_v3.upgrade file once it's all testing and working.

-ml

Yeah, it is. As you point out, with Hijack, you can just use /proc/empeg_notify.

Gag.. this is taking forever, but that's okay -- I been going between the workshop (antenna stuff) and this for much of the afternoon now.

The 320GB drive I'm using here also requires more swap space for the mkfs to work, so the script now uses *both* hda3 and hda6 for swap. And I've updated it to zero hda3 afterwards, so that the dynamic data partition starts off clean instead of semi-random.

And other updates. Now running the (hopefully) final test from scratch.

-ml

Okay, I've updated the bigdisk builder on my web server. Could somebody please copy the new version over to riocar.org (I think there's a mirror of this stuff being kept there) ?

The new version has been tested on a 320GB drive here, and worked fine for me. I have not yet installed any tunes onto the drive, but I'll try that as well, just to confirm things.

Cheers

Mirrored here: http://riocar.org/upload/bigdisk/

Thanks, Tom!

EDIT: I just noticed a double-negative typo in my index.html file, so if you could pull that one again...

Thanks!

Fixed

Here's the new init script that I used for builder_bigdisk_v3. This also eliminates the scrawl and make-drives utilities, as they have been rolled into this one instead to save some execution time RAM.

Attached and inline (We really need a <pre> tag!!):

Code:

#!/bin/sh

log() {
        echo "$1"
        echo "POPUP 9999 $1" > /proc/empeg_notify
}

failed() {
        log "$1 failed"
        /bin/swapoff /dev/hda3
        /bin/swapoff /dev/hda6
        sync
        exec /bin/bash --login
}

makedrive() {
        dev="/dev/$1"
        log "Making $1 filesystem"
        /bin/mkfs.ext2 -v -s 1 -i 131072 -m 0 $dev || failed "mkfs $1"
        log "Tuning $1"
        /bin/tune2fs -c -1 -i0 $dev || failed "tune2fs $1"
        /bin/sync
        /bin/sync
        log "Mounting $1"
        /bin/mount -n -o rw,nocheck $dev $2 || failed "mount $1"
        /bin/sync
        /bin/sync
        log "Making directories"
        /bin/mkdir $2/fids || failed "mkdir /fids"
        /bin/mkdir $2/var || failed "mkdir /var"
        echo "[hijack]" > $2/var/config.ini || failed "config.ini"
        log "Remounting ro"
        /bin/mount -n -o remount,ro $dev $2 || failed "remount"
        /bin/sync
        /bin/sync
        log "$1 completed"
}

# echo everything to the serial console:
set -x

/bin/mount -n /proc
log "Builder image"

DRIVE1=""
if [ -e /proc/ide/hdb -a -e /proc/ide/hdc ]; then
        log 'hdb,hdc both exist?'
        exec /bin/bash --login
fi
[ -e /proc/ide/hdb ] && DRIVE1=hdb
[ -e /proc/ide/hdc ] && DRIVE1=hdc

/bin/mount -n -t ext2 -o ro,nocheck /dev/hda4 /drive0
if [ -d /drive0/fids ]; then
        log "Drives already built"
else
        /bin/umount /drive0

        /bin/mkswap  /dev/hda6
        /bin/swapon  /dev/hda6
        /bin/mkswap  /dev/hda3
        /bin/swapon  /dev/hda3

        log "Tuning hda5"
        /bin/tune2fs -c -1 -i0 /dev/hda5 || failed "tune2fs hda5"

        makedrive hda4 /drive0

        if [ "$DRIVE1" != "" ]; then
                log "partitioning second drive"
                init_drive /dev/$DRIVE1 || failed "init $DRIVE1"
                DRIVE1="${DRIVE1}4"
                makedrive $DRIVE1 /drive1 || failed "init $DRIVE1"
        fi

        /bin/swapoff /dev/hda3
        log "Zeroing hda3"
        /bin/cat /dev/zero > /dev/hda3

        /bin/swapoff /dev/hda6
        log 'Done!  Testing disk..'

        /sbin/hdstress
fi

exec /bin/bash --login