Now this makes sense. I think its safe to say that the boot portion is on the disk itself as well as the OS for the printing system. I just wonder if there is a BIOS that is accessible via front-panel that allows to prep the disk and to set it up for ingestion of a firmware file, similar to the larger 5500 series printers...but indeed it must be some sort of unix based OS on these drive, just for the fact that an FSCK is done on startup of the printer when powered down incorrectly.
http://h30499.www3.hp.com/t5/Printers-Designjet-Large-Format/Installing-a-new-blank-hard-drive-in-a-DesignJet-5500-or-5500PS/td-p/2350749#.VO1nQMYQKvI
(I was hoping someone was keen enough to have found a way, heheh)
Somehow, I think that these guys have found a way to rewrite the drives, either by using an alternate formatter in a printer that allows for firmware writes (maybe using a formatter for the designjet 5500 and others) I sure don't see a USB port option...and the service mode won't be accessible until the printer/OS is successfully loaded.
It's more for peace of mind. That's really the only reason I would want a clone, as drives have a finite timespan. These printers will eventually die once the drive gives up the ghost, and will render the printers useless. I think HP should have used a flash-based OS/firmware and a hard drive for job storage...
This article was promising as well...might have to try that, but need to figure out whether the IP is still accessible after the printer is powered back on.
http://h20565.www2.hp.com/hpsc/doc/public/display?sp4ts.oid=436089&docId=emr_na-c01768718
Thanks for the help and advice guys. If anyone has any clues on what the ATA password could be on these drives, that would go a long way to having access and being able to clone the drives.
The ATA password is "MartaLaiaDesiree".
Once you have that you can mount the drive on Linux by passing that via hdparm to the disk:
# hdparm --user-master u --security-unlock "MartaLaiaDesiree" /dev/sdb
and then you can use fdisk to get the partition table:
# fdisk /dev/sdb
Command (m for help): p
Disk /dev/sdb: 298.9 GiB, 320072933376 bytes, 625142448 sectors
Disk model: [redacted]
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: dos
Disk identifier: [redacted]
Device Boot Start End Sectors Size Id Type
/dev/sdb1 * 63 273104 273042 133.3M 83 Linux
/dev/sdb2 273105 2281229 2008125 980.5M 83 Linux
/dev/sdb3 2281230 18298034 16016805 7.7G 83 Linux
/dev/sdb4 18298035 156296384 137998350 65.8G 5 Extended
/dev/sdb5 18298098 20306159 2008062 980.5M 82 Linux swap / Solaris
/dev/sdb6 20306223 62316134 42009912 20G 83 Linux
/dev/sdb7 62316198 120326849 58010652 27.7G 83 Linux
/dev/sdb8 120326913 156296384 35969472 17.2G 83 Linux
As you can see, even if your formatter has a 300+GB drive, the machine only uses 80GB for partitions, the rest is unpartitioned space.
If you mount the third partition you can find an /etc/fstab which looks like this:
/dev/disk/disk0/p1 /boot ext3 async,atime,auto,dev,exec,rw,suid,nouser 1 1
/dev/disk/disk0/p3 / ext3 async,atime,auto,dev,exec,rw,suid,nouser 1 1
/dev/disk/disk0/p5 swap swap defaults 0 0
/dev/disk/disk0/p6 /vpm ext3 async,atime,auto,dev,exec,rw,suid,nouser 1 1
/dev/disk/disk0/p7 /data ext3 async,atime,auto,dev,exec,rw,suid,nouser,data=writeback 1 1
/dev/disk/disk0/p8 /plots ext3 async,atime,auto,dev,exec,rw,suid,nouser,data=journal 1 1
none /proc proc defaults 0 0
none /sys sysfs defaults 0 0
none /dev/shm tmpfs defaults 0 0
The bootloader is lilo, so altering partition size/location is somewhat problematic. I assume partition 2 is a recovery partition and the partitions you'd really want to tweak to do anything useful in terms of running bigger jobs or storing more jobs or whatever are p5 (swap), p6 (vpm), p7 (data), and p8 (plots).
Ideally you would first want to unlock the drive and then run a straight dd clone or clonezilla.
Note that the formatter is based off the Intel ICH4 southbridge which is natively PATA, there is a PATA to SATA converter chip on the board, so even if you go to a 7200rpm disk or a SSD you're not standing to gain much in performance because you're still limited to the max of ATA/133MB/s speed even though the slowest SATA spec is 150MB/s. So don't feel the need to put in the world's fastest SATA SSD thinking you'll save minutes off your next startup.
P.S.: This is for a Q5670-60021 Rev A formatter, but probably works for anything with the 50X15CR formatter, I would assume since the ATA password has to lurk in the BIOS and the BIOS ought to be the same for the boards that the password is the same.