BlackArmor NAS Network Storage Server

[Hajo]

It's christmas time ;-)

I've successfully replaced the original Seagate firmware with a customized Debian GNU/Linux 5.0.7 (Lenny) system. For your convience, you only have to upload a 'magic' service pack to your NAS and the rest of the installation gets executed automatically:

http://www.noerenberg.de/hajo/pub/seagate-blackarmor-nas.txt

WARNING: replacing the firmware is a dangerous process and may damage your NAS permanently! You have been warned!

Hajo

[netware5]

Merry Christmas Hajo! Thank you very much for your efforts, this is a real Christmas present for all owners of NAS 220! Before going to apply your service pack I would like to ask some questions:

1. The Web Interface still exists? If yes, then what kind of changes apply?

2. Samba server still exists? I am asking this question because of your previous statement that you use only NFS, not SAMBA.

3. What if the "Public" share does not exists or the user rights had been modified? In my case it exists, but I made some modifications on user rights.

4. What version of FTP server you apply? Does it support secure ftp?

5. Do you plan to provide us with some kind of "release notes" to explain main differences with the original FW?

[Hajo]

netware5,

please find my comments to your questions below:

1.) The web interface does not exist anymore. After applying the SP, the system starts with a Debian Lenny minimal system (basic system binaries and SSH daemon only). However, you are free to install additional packages via the Debian package management system. For configuration, you'll have to use your favourite text editor.

2.) apt-get install samba

3.) The SP searches for the root fs image at /DataVolume/Public/custom-rootfs_full.tgz(+.md5). If it is not able to open the image file for any reason, the SP only patches the preroot image (NAND flash), and does not replace the main system.

4.) apt-get install vsftpd/proftpd/...

5.) there are nearly no similarities between the original fw (Seagate/Wistron proprietary) and the Debian system (open source Linux distribution). During installation, all Seagate system files are removed (except the Linux kernel and a few supplementary files).

Installation of the Debian/GNU Linux system is only recommended for people with (at least) basic linux administration skills.

As of today, the installation has only been tested on one (my) BA220. I can not guarantee that it works for your system as well. If things go wrong, you'll need at least a custom-made serial cable to monitor/change the boot process (u-boot loader).

debian-armel:~# cat /etc/debian_version
5.0.7
debian-armel:~# uname -a
Linux debian-armel 2.6.22.18 #1 Mon Mar 15 14:19:05 CST 2010 v0.2.4 armv5tejl GNU/Linux
debian-armel:~# free
             total       used       free     shared    buffers     cached
Mem:        125492      70796      54696          0       6076      49360
-/+ buffers/cache:      15360     110132
Swap:      1044792       3812    1040980

Hajo

[netware5]

Hajo, thank you very much indeed for your answer. I'm planning now the migration

[CyberMaxX]

Hajo -

Just wanted to say thank you for the great Holiday Gift to us all. 

Nice Work!

Cheers

[ashantonio]

Wowwww

Any news about? transfer rates?

iscsi support?

Thxxxx

[Hajo]

ashantonio,

iscsi-target (IET) is working fine. I've compiled the necessary kernel module (iscsi_trgt.ko) last week. I'll include it in my next release of the Debian rootfs. Openvpn support (tun.ko) will be included as well.

I'll post some speed measurements later.

Hajo

[ashantonio]

Great!!

Let us informed about !!

Really good. Please do not delay!!

Thank youuuu

[FHWN12]

Thank you Hajo, I didn't try your solution yet. Before, I have a few questions: Is the transfer speed better? CIFS reed, write, FTP reed, write ? What about the problems with routers? Is this a hardware problem? Can I set ftp permissions for a subfolder (for dyndns access) ? I have basic linux administration skills, but I don't want to spend much time for it. Thanks very much … FHWN

[Hajo]

ashantonio,

Performance is about 20MB/s for writing and 24MB/s for reading (gbit network, RAID1 BA 220). Since my time is very limited these days, these are just some quick tests (tests have been repeated to suppress caching effects, though).

1. CIFS, linux smbclient to NAS samba server (default Debian install, no smb.conf tuning)

smb: \public\> mput movie*  
Put file movie-xy.avi? y
putting file movie-xy.avi as \public\movie-xy.avi (20738.5 kb/s) (average 20738.5 kb/s)                                

smb: \public\> mget movie*
getting file \public\movie-xy.avi of size 736686080 as movie-xy.avi (24601.4 kb/s) (average 24090.7 kb/s)              
smb: \all\>

2. iscsi, copy file to target (ext3 fileio image via IET-iscsi)

iscsi-client: ~/movies$ dd if=movie-b.avi of=/mnt/iscsi-target/movie-b.avi
4540812+0 records in
4540812+0 records out
2324895744 bytes (2.3 GB) copied, 111.962 s, 20.8 MB/s

3. iscsi, bonnie++ benchmark (ext3 fileio image via IET-iscsi)

Version 1.03d       ------Sequential Output------ --Sequential Input- --Random-
                    -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
iscsi-client    1G 19427  22 20352   2 20457   2 103531  99 +++++ +++ +++++ +++
                    ------Sequential Create------ --------Random Create--------
                    -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
              files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                 16 +++++ +++ +++++ +++ +++++ +++ +++++ +++ +++++ +++ +++++ +++

4. raw hdd speed (bonnie++ executed on NAS BA 220 RAID1)

Version 1.03d       ------Sequential Output------ --Sequential Input- --Random-
                    -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
debian-armel   300M  1777  97 46019  61 23283  38  1821  99 51580  56 381.9   5
                    ------Sequential Create------ --------Random Create--------
                    -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
              files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                 16  8584  94 +++++ +++ 12630  91  9280 100 +++++ +++ 14194 100

I'm currently checking various kernel/network-related issues, maybe there will be some news later. Maybe not

Hajo

[Hajo]

FHWN12,

if you have only basic linux admin skills and don't want to spend much time, you better do not try to install the Debian system. The package has not been tested extensively.

Hajo

[ashantonio]

Hi!!

The performance is good, but not enough to connect to my XenServer.
Anyway these days proves.
Your work is amazing!

Thankssssssssssssssss

[lnxmnky]

Hey all

Wanted to say thanks for the nice work Hajo. I managed to finally managed to get all my data off of the box, so that I could play with it.

Since I haven't seen to many people post about the 420(mainly the 440 and 220).

Your ssh/password solution worked just fine on the 420.

The debian install works as well; there are things I noticed but only one things to do with your custom firmware, but most of them had more to do with the generally layout of the default build of the offical firmware.

The only thing I ran into after installing your custom firmware was when I tried to install webmin, it complained a bit about some dependecies, even though they were on the box. I just did an apt-get -f install, which cleared it right up, and I have had no futher issues.

Here are a couple of notes that might help some others, but as you have noted Hajo this is not for the beginner.

1st thing I did was take and put the unit back to it's default state out of the box. This caused me an issue

right off the bat. I have 4 2TB drives in the unit as opposed to the standard 2 1TB drives is came with.

I tried setting it back to the default state with both the offical procedure, and with cleaning the mbr/gpt regions and letting the box just see them as blank drives.

WARNING DO NOT DO THIS IF YOU ARE NOT COMFORTABLE WITH EITHER OF THESE COMMANDS. THIS WILL REMOVE ALL INFORMATION ON THE DISKS!!!!!

I have found 2 ways to quickly get the drives back to a "blank" state. 1 is to connect them up to a linux box, and do:

dd if=/dev/zero of=/dev/<device name> bs=512 count=2

dd if=/dev/zero of=/dev/<device name> bs=512 seek=<put the total number of blocks, minus 200> *

*Note: This should only have to be done on larger drives, as most smaller drives will not use gpt.

2nd method(windows):

run diskpart in a command window

list disk (find # of disk you want to clean)

select disk <disk #>

clean

exit

This at 1st looked liked it was working, but would fail after about 8 hrs, and tell me that one of the drives were bad, after several days, and multiple attempts, in which I moved the disks around, and found that it would fail on different disks, but always the in the same place. I finally just put 2 of the 2TB drives into the unit, and let it rebuild, then I added the other disks in later(this does cause some extra steps later on to use them)

After finally getting them rebuilt back to the default state, I loaded up your firmware, and it went as it was written in the faq.

I have only put the 3 packages that you suggested and of course webmin, so far everything looks good, and other than one issue with accessing the unused portions of the disks(some manual steps required if you plan on using the disks in a raid 5)

As for the speed of cifs, or anything else, well I have only started playing with this build, this morning, so I haven't done a whole lot yet, as I am stilling getting it tweaked up, and general tests. Will report more later.

Cheers

[Hajo]

lnxmnky,

thank you for your report and congratulations for beeing BA Debian GNU/Linux user #2 :-)

Cleaning of the disks shouldn't be necessary, as long as you haven't modified the official firmware. Your shared data still exists after Debian installation (side note: I do not guarantee anything, this is alpha software).

Do you have any details on the dependency-issue? The Debian dpkg system is in a clean state after installation and I wonder what was going wrong there.

Hajo

[lnxmnky]

I don't have the logs as I forgot to turn them on, as I am just in the playing around stage, but as I am planning on doing a complete rebuild again in the next day or so, I will keep all of the logs, and share out with you anything I find.

From what I do remember however, I ran into the issue after installing the 3 packages you recommended. They installed cleanly, and I did not have any reason to check till I tried installing webmin.

It could just be that I caused the issue by something I did, as I was pretty much just trying tons of things out.

but like I said I will keep track of my steps, and let you know as soon as I get a chance to redo the install.

as for the cleaning of the drives, that was pretty much so that If anything went wrong , I could try and isolate where the problem was.

btw as of now, I have managed to not only get webmin running without any issues, with the exception of the raid, which I had to do manually, but also getting it to run a full instance of apache, dns, ldap, ampache, all of which installed without issue either from apt-get or installing right from webmin.

I will be doing tests on those later to see how the box handles all of these services while under load of a half dozen boxes.

[ashantonio]

Hajo,

It would be better performance with SSD drives?
24 MBs async mode is too poor to vhd's on XenServer ...
Thanks a lot

[Hajo]

lnxmnky,

if you install lots of packages you have to keep an eye on memory/cpu consumption. 128MB RAM (BA220) is not that much and if the system starts swapping, performance will decrease badly.

If you install packages via webmin and the Debian package manager, things probably get screwed up because you mix two installation sources. It might be a good idea to use Debian packages only. But I do not know anything about webmin, I've never used it.

ashantonio,

as you can see from the bonnie++ benchmark, local disk I/O speed is up to 50MB/s. I think the bottleneck is the CPU and network card. The network card driver seems to be unstable. It produces lots of kernel errors when stress-tested.

Hajo

[lnxmnky]

actually I ran into the problem when installing webmin(debian package), which I am pretty sure that I did right after the 3 packages you recommended(again I don't think these were the problem), but not 100% sure I remembered the order, or if I did anything before installing webmin.

The 420 has a bit more memory (256). I will pull a dump of that box and post it.

Other than some high wait states(which I attribute a lot to the fact that I was still syncing the raid 5), I really didn't see the box getting overwelmed. swap was barely touched, cpu was high, but still pretty resonable given what I was doing at the time. Although I will agree that the cpu on this box is weak. Would be better with a dual core.

[ashantonio]

There may be many mistakes along the support jumbo frames for the network card.
I put my network to support jumbo frames and I have hdparm more than 50MB / s.
I think the cpu is not the problem, but the sata controller and network interface.

Hope that helps.

[lnxmnky]

okay, I finally managed to do a full clean install of the firmware. The "problem" that I ran into the last time happened again. Turns out that I was not paying enough attention, and that the problems were just for dependencies that webmin needed to install, so there were no actual errors, just a case of lousy memory, and no logs.

One thing I did find out however is that the "datastorage" raid volume does not work after installing the firmware, I noticed that it happened the last time I installed this firmware as well. See below

debian-armel:/dev# df -k
Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/md0               1027416    481188    494036  50% /
tmpfs                   127676         0    127676   0% /lib/init/rw
udev                     10240       796      9444   8% /dev
tmpfs                   127676         0    127676   0% /dev/shm
/dev/md2                504932     10283    468579   3% /mnt/md2

debian-armel:/dev# parted -l
Model: Seagate ST32000542AS (scsi)
Disk /dev/sda: 2000GB
Sector size (logical/physical): 512B/512B
Partition Table: gpt

Number  Start   End     Size    File system  Name     Flags
 1      100MB   1169MB  1069MB  ext3                  raid
 2      1169MB  2239MB  1070MB  linux-swap            raid
 3      2239MB  2773MB  534MB   ext3                  raid
 4      2773MB  2000GB  1998GB               primary  raid


Model: Seagate ST32000542AS (scsi)
Disk /dev/sdb: 2000GB
Sector size (logical/physical): 512B/512B
Partition Table: gpt

Number  Start   End     Size    File system  Name     Flags
 1      100MB   1169MB  1069MB  ext3         primary  raid
 2      1169MB  2239MB  1070MB  linux-swap   primary  raid
 3      2239MB  2773MB  534MB   ext3         primary  raid
 4      2773MB  2000GB  1998GB               primary  raid


Model: Seagate ST32000542AS (scsi)
Disk /dev/sdc: 2000GB
Sector size (logical/physical): 512B/512B
Partition Table: gpt

Number  Start   End     Size    File system  Name     Flags
 1      100MB   1169MB  1069MB  ext3         primary  raid
 2      1169MB  2239MB  1070MB  linux-swap   primary  raid
 3      2239MB  2773MB  534MB   ext3         primary  raid
 4      2773MB  2000GB  1998GB               primary  raid


Model: Seagate ST32000542AS (scsi)
Disk /dev/sdd: 2000GB
Sector size (logical/physical): 512B/512B
Partition Table: gpt

Number  Start   End     Size    File system  Name     Flags
 1      100MB   1169MB  1069MB  ext3         primary  raid
 2      1169MB  2239MB  1070MB  linux-swap   primary  raid
 3      2239MB  2773MB  534MB   ext3         primary  raid
 4      2773MB  2000GB  1998GB               primary  raid

debian-armel:/proc# cat mdstat
Personalities : [linear] [raid0] [raid1] [raid10] [raid6] [raid5] [raid4]
md1 : active raid1 sdd2[1] sdc2[2] sdb2[3] sda2[0]
      1044800 blocks [4/4] [UUUU]

md2 : active raid1 sdd3[1] sdc3[2] sdb3[3] sda3[0]
      521408 blocks [4/4] [UUUU]

md0 : active raid1 sdd1[1] sdc1[2] sdb1[3] sda1[0]
      1043840 blocks [4/4] [UUUU]

unused devices: <none>

Installing the firmware was done after returning the nas box to it's default state(had to do this, as I renamed the volume, and the firmware specifically looks for /datastorage/public. removing and rebuilding the volume really didn't make much sense as I use raid5, so it would have taken the same amount of time either way, and I wanted to remove any chance that I introduced an anomaly into the mix.

the only things done to the box at this point were, root password reset, adding in packages (nfs-kernel-server samba vsftpd webmin openssl libauthen-pam-perl libio-pty-perl apt-show-versions)

I am going to recreate the raid5 device manually from the old partitions(sda4,sdb4,sdc4,sdd4), and see what happens.

[lnxmnky]

I was able to just re-assemble the raid5(mdadm --assemble /dev/sda4 /dev/sdb4 /dev/sdc4 /dev/sdd4), but will have to re-create everything on there, as /dev/md3 was being used as an lvm device, and I don't have access to the old configuration information.

Might be a good idea to log into the box and grab a copy of it prior to installing the firmware. I will test this out to see if that works.

[tuxmania]

Hi Hajo, and a really big warm thanks!

This installation works like a charm on the Blackarmor 110 NAS. The only thing i had to do was mounting /dev/md3

I now have a fogproject storage node running on a NAS, and man is it fast!

Again, thanks so very much =)

[Hajo]

lnxmnky, tuxmania,

my SP intentionally does not configure (reassemble, mount, fsck) the data partition. I would be too much work to reliably figure out the setup (raid1, raid5, crypto ...). Additionally, if things go wrong, the NAS might not boot at all.

It should not be a problem to manually set up the data volume.

Hajo

[stianiquniez]

Hello.. 

First, thanks for this wonderful 'hack' to blackarmor. But i have a few questions, that I hope you can answer me. 

1: I have enabled ssh, with you're howto. Do I have to restore it to default settings before installing Debian? 

2: Do you know about a bittorent client that works on the linux that is allredy installed? Or will it be easier to just install debian and use webmin?

3: I have never used webmin before, but have taken a look at it, and it looks kind of nice, so I wonder if it`s kind of simple to add users who can access my files.. 

I have tried to google the topic about torrent client on a Blackarmor, but can`t seem to find anything, do you guys in here know anything about this? 

[Hajo]

stianiquniez,

1.) No, you do not have to restore the default settings.

2.) Any bittorrent client compiled for the 'armel' architecture probably works.

3.) I've never used webmin. User 'tuxmania' recommened http://www.fogproject.org/ which might be a nice alternative.

Hajo

[stianiquniez]

How long does the installation normally takes? My 420 has been standing there for 2 hours now, and I still can't connect to it using ssh. I beginning to think that something went wrong. I have checked that it did not change IP.. On the screen it just says 'SystemBootingNow....' Some suggestion? I have not tried to reboot it manually, as I am afraid it still installing, therefore I am asking. 

Or will it be best to just restore it? 

Update: 

I restored the blackarmor, and tryed again, everything from scratch.. But I don't get any further.. So I don't think it works on 420... Hope you will continue the work on this Hajo, as it would be nice to get it working some day... 

[lnxmnky]

stianiquniez,

     This does indeed work on a 420. It should only take a few mins for this to work, as the nas device does get rebooted. Are you sure you have that port opened up? This is especially important if you are using windows vista or windows 7, and have their firewall enabled.

[stianiquniez]

Well, I have NO idea then what I am doing wrong. I am using mac, it's on the same network.

I only used recoverytools.exe or something to restore it.. Only took some minutes. Maybe I need to clean the disks to be 100% sure everything is deleted? 

[Hajo]

stianiquniez,

the IP of your NAS might have changed after the installation of the Debian system. Please try to scan your network for new IPs.

My Debian system does not change the contents of the LCD display (my BA220 does not have a display). You can use the /usr/bin/show_lcm command to send your own messages to the display after login via SSH.

If your problems persist, you have to attach a serial cable to your NAS to view the logs during system boot ($5 for a CA-42 3.3v mobile phone cable, and you need basic soldering skills for attaching a 2x5 pin-connector).

Hajo

[stianiquniez]

Hi, thanks for insisting on the IP change solution.. 

I have been watched the ip from my router, thought it would update as soon as the mac adress changed IP-adress. But it did not, so i used a IP scanner, and what do you know! 

One last question. When I was troubleshooting for why it 'did not work'. I took the power from the NAS, and booted it again. And the NAS loaded a file I think, and I got back to the web interface, but the nfs/afp servers were not available, so then i recovered the blackarmor using recovery tools. Anyway, I was thinking, will this happen when i reboot the NAS, or it loses the power supply? I was able to find nanddump.mtd2s4000.img in the /mnt/md2/ file.. witch is a partition on the first hardisk. So i was wondering if thats the file, it uses to recover to the orginal system when it looses power. What i DON'T want to happened. I was thinking of this just now, after some hours with configuring the server. 

[Hajo]

stianiquniez,

the preroot loader only reinstalls the system if there is no (correct) partition table on your hard disks (blank/new HDDs or crashed system). Thats the same behavior as with the original firmware.

The file nanddump.mtd2s4000.img contains the preroot loader backup (original Seagate version) from NAND flash partition 2 (CRAMFS image without CRC header). My SP has to patch the preroot loader to fix a bug there.

Did you try to use the show_lcm command? /usr/bin/show_lcm "message1" "message2"

Hajo

[Isaac_X]

Thanks Hajo, this is a great hack. The BlackArmor 440 finally works the way I want it to... it's still overpriced and much too slow compared to similar products, but at least now it can be configured to work properly with Linux and OS X.

Here's Webmin running on the BlackArmor 440, with all of my data intact, and a full-featured interface. There's Netatalk (AFP) for full OS X compatibility. Thanks again Hajo, this really is great.

[stianiquniez]

Hi hajo.. 

I could not find the /usr/bin/show_lcm.. 

I tried a find -name show* 

/usr/bin/showkey

/usr/bin/showcfont

/usr/bin/showkey/.usr/bin/showcfont 

was all i was able to find... So I have not been able to change the LCD. Not that I need that, but it would be nice. 

One question.. 

The / of the filesystem.. Is that on the first hardisk? I have been strugeling with low space now, and can't install some packages i would like.. So I was wondering if debian is on some flash drive on the disk, as the first disk is a part of a LVM.. I havent have the guts to change it to ext3. Would like some more space on my / disk. this is my fdisk -l 

WARNING: GPT (GUID Partition Table) detected on '/dev/sda'! The util fdisk doesn't support GPT. Use GNU Parted.

Disk /dev/sda: 1000.2 GB, 1000204886016 bytes

255 heads, 63 sectors/track, 121601 cylinders

Units = cylinders of 16065 * 512 = 8225280 bytes

Disk identifier: 0x06751000

   Device Boot      Start         End      Blocks   Id  System

/dev/sda1               1      121602   976762583+  ee  EFI GPT

Disk /dev/md0: 1068 MB, 1068892160 bytes

2 heads, 4 sectors/track, 260960 cylinders

Units = cylinders of 8 * 512 = 4096 bytes

Disk identifier: 0x00000000

Disk /dev/md0 doesn't contain a valid partition table

Disk /dev/md3: 997.3 GB, 997330911232 bytes

2 heads, 4 sectors/track, 243488992 cylinders

Units = cylinders of 8 * 512 = 4096 bytes

Disk identifier: 0x00000000

Disk /dev/md3 doesn't contain a valid partition table

Disk /dev/md2: 533 MB, 533921792 bytes

2 heads, 4 sectors/track, 130352 cylinders

Units = cylinders of 8 * 512 = 4096 bytes

Disk identifier: 0x00000000

Disk /dev/md2 doesn't contain a valid partition table

Disk /dev/md1: 1069 MB, 1069875200 bytes

2 heads, 4 sectors/track, 261200 cylinders

Units = cylinders of 8 * 512 = 4096 bytes

Disk identifier: 0x00000000

Disk /dev/md1 doesn't contain a valid partition table

Disk /dev/dm-0: 997.3 GB, 997329993728 bytes

255 heads, 63 sectors/track, 121251 cylinders

Units = cylinders of 16065 * 512 = 8225280 bytes

Disk identifier: 0x00000000

Disk /dev/dm-0 doesn't contain a valid partition table

WARNING: GPT (GUID Partition Table) detected on '/dev/sda'! The util fdisk doesn't support GPT. Use GNU Parted.

Disk /dev/sda: 1000.2 GB, 1000204886016 bytes255 heads, 63 sectors/track, 121601 cylindersUnits = cylinders of 16065 * 512 = 8225280 bytesDisk identifier: 0x06751000
   Device Boot      Start         End      Blocks   Id  System/dev/sda1               1      121602   976762583+  ee  EFI GPT
Disk /dev/md0: 1068 MB, 1068892160 bytes2 heads, 4 sectors/track, 260960 cylindersUnits = cylinders of 8 * 512 = 4096 bytesDisk identifier: 0x00000000
Disk /dev/md0 doesn't contain a valid partition table
Disk /dev/md3: 997.3 GB, 997330911232 bytes2 heads, 4 sectors/track, 243488992 cylindersUnits = cylinders of 8 * 512 = 4096 bytesDisk identifier: 0x00000000
Disk /dev/md3 doesn't contain a valid partition table
Disk /dev/md2: 533 MB, 533921792 bytes2 heads, 4 sectors/track, 130352 cylindersUnits = cylinders of 8 * 512 = 4096 bytesDisk identifier: 0x00000000
Disk /dev/md2 doesn't contain a valid partition table
Disk /dev/md1: 1069 MB, 1069875200 bytes2 heads, 4 sectors/track, 261200 cylindersUnits = cylinders of 8 * 512 = 4096 bytesDisk identifier: 0x00000000
Disk /dev/md1 doesn't contain a valid partition table
Disk /dev/dm-0: 997.3 GB, 997329993728 bytes255 heads, 63 sectors/track, 121251 cylindersUnits = cylinders of 16065 * 512 = 8225280 bytesDisk identifier: 0x00000000
Disk /dev/dm-0 doesn't contain a valid partition table

And then more disk.. comes up, but i don't think u need that. Thinking of just expand my /. Or is it not possible? 

[Hajo]

stianiquniez,

unfortunately my SP does not correctly copy the show_lcm command from the original Seagate fw to the Debian fs. I'll fix this within the next version.

Nevertheless you can copy the command from the preroot loader backup like this:

debian-armel:~# mkdir /tmp/foo
debian-armel:~# mount -o loop /mnt/md2/nanddump.mtd2s4000.img /tmp/foo
debian-armel:~# cp /tmp/foo/usr/bin/show_lcm /usr/sbin/
debian-armel:~# umount /tmp/foo

Please use GNU parted ('print' subcmd) to show the partition table. fdisk does not support GPT tables.

I can't believe that you installed >1GB of applications to the device It should be theoretically possible to resize all partitions ('/' ist the first partition) if you connect all drives to a Linux PC. But that's not a fun thing to do. Bind-mounting (mount --bind) some directories might be better option.

Hajo

[bill meyer]

Hi Hajo

Got my BA 220 up and running with Lenny. (I am not so familiar with Lenny - yet)

A few questions:

- How do I get the automounter for the USB ports up?

- I can mount USB devices with VFAT, but not NTFS.

OK - found the answer to this one. Had to  mount -t ntfs-3g /dev/sdc2 /mnt/USB2_2 -o force however

(http://beginlinux.wordpress.com/2009/03/18/mounting-an-ntfs-drive-in-debian/)

The gist of it is:

apt-get install libfuse2
apt-get install ntfs-3g

- Do you know of a DLNA server that works with BA NAS + Lenny?

[bill meyer]

Problem:

I share the NAS via a Netgear wireless router.

using cp to copy files gives little feedback on progress, so I use mc to manage the files... but...

ssh keeps on disconnecting during copying bulk files from USB device attached to NAS.

Solution:

1. Loaded a minimal x-server config, tightvncserver and midnight commander.

 apt-get install xserver-xorg-core xorg

 apt-get install xserver-xorg-core xorg

 apt-get install tightvncserver

 apt-get install mc

For extra goodness... install ice window manager. Very small.

 apt-get install icewm 

2. run vncserver

3. Connect to vncserver with remote desktop / vncviewer

4. Manage files with Midnight Commander

Now the connection can drop as much as it likes,  just reconnect via vnc. File copy is not interrupted. 

[Hajo]

Bill,

maybe you want to have a look at the 'screen' and 'rsync -v --progress' commands, which might be easier alternatives.

The BA NAS comes with a copy of the (commercial) ufsd driver. My SP copies the module to the Debian system. You can use it like this:

debian-armel:~# modprobe ufsd
debian-armel:~# dmesg | tail -n5
ufsd: module license 'Commercial product' taints kernel.
ufsd: driver loaded at bf013000 built on Jun  5 2009 15:42:37
UFSD version 7.04 (Jun  5 2009, 15:37:57)
NTFS read/write support included
$Id: ufsdvfs.c,v 1.207.4.1 2009/05/14 14:37:07 shura Exp $ (LBD=ON)
debian-armel:~#

I've never tried to mount an NTFS-filesystem, but 'mount -t ufsd ...' should do the work.

Hajo

[deltaray]

awesome news @Hajo! I have not tried this on my BlackArmor NS110 yet, but I am going to do.

Another question I have before I start installing my own Linux on this thing is, since the NS110 has only one Harddisk, it doesn't make much sense to me to run those partitions in raid 1 mode, which they do.

For example mdadm --detail /dev/md3 shows this:

[code]
/dev/md3:
Version : 0.90
Creation Time : Thu Oct 22 15:17:35 2009
Raid Level : raid1
Array Size : 1950707904 (1860.34 GiB 1997.52 GB)
Used Dev Size : 1950707904 (1860.34 GiB 1997.52 GB)
Raid Devices : 2
Total Devices : 1
Preferred Minor : 3
Persistence : Superblock is persistent

Update Time : Wed Feb 2 14:56:28 2011
State : clean, degraded
Active Devices : 1
Working Devices : 1
Failed Devices : 0
Spare Devices : 0

UUID : 1527aef2:79ba48fd:0b605f5f:46fb0873
Events : 0.223346

Number Major Minor RaidDevice State
0 8 4 0 active sync /dev/sda4
1 0 0 1 removed

[/code]

So would it be possible to mount /dev/md3 by using /dev/sda4 directly, without having it run in raid mode?

Has anybody tried this yet? Would it increase performance?

I have a simular device from called Seagate FreeAgent Dockstar, with Debian squeeze installed, and this device has much more performance but almost the same hardware as the NS110.

regards

[Hajo]

deltaray,

there should be no problem formatting and mounting /dev/sda4 directly. One owner of a BA110 mailed me by private email and told that he is using the NAS with XFS and no RAID. No word on performance, though.

Hajo

[deltaray]

Just one question before I try it, will the data on the sda4 (md3) partition remain untouched when installing debian using your update files?

Or should I create an backup of all my data just in case?

And if I understand your description correctly, if something goes wrong, I can open my NAS, plugin the HD to my PC, erase all content, plugin the HD back  int the NAS, and get the original system from seagate automatically restored?

best regards

[Hajo]

deltaray,

my SP does only modifies md0 and md2, the data partition remains untouched. If you insert blank disks into the NAS, the original Seagate FW gets installed.

Disclaimer: I do not guarantee anything, your NAS may get bricked, make backups, don't cross the street at a red light, be nice to your grandma.

Hajo

[eschumann]

Thanks for a lot of fun, Hajo! this is really great. I installed your SP yesterday evening around 19 and three hours later everything basic was running like a charm (samba, webmin, raid1 etc)

Now I want to  do some more things with the box, but as I understand it you're using the original kernel? Right?

Problem is that I would like to install some more kernel modules to be able to access cifs shares and to set up an ipsec vpn conection. Any chance to re-compile the kernel and get the needed modules or should I forget that project?

Any ideas?

Cheers, Erik

[deltaray]

hehe I am always nice to my grandma xD

The update on my Blackarmor NS110 went perfect, I had to recreate the "Public" share folder, because deleted it earlier, but then it worked.

Awesome thank you so much for this installation package!

[deltaray]

AWESOME!!!!!!!!

With the original Seagate Linux System, I had about 12-14MB/s transfer rate.

With your Debian adapted version and a default Samba installation (no tweaks at all), I already get 30-35MB/s transfer rate, stable!

Again, thanks so much for installatio package, I owe you a beer

[Autobahn]

First of all..thanks Hajo for the effort!

I have installed the Debian onto a 110. Installed Webmin and it worked, at least the installation. I´m not a Linux pro, what ever. This leads to my question.

Have anyone tried the rollback/fallback option to open the 110, inserting a blank drive, and got the NAS working with the old firmware/software again?

If so, can anyone advice how to proceed?

Thanks in advance!

Bahn

EDIT! Just dropped a blank Drive into it and it´s exactly like descriped by Hajo...It just took time! The normal Seagate WebGUI is available. Awesome!

EDIT2 :-) .....Can one of the Pro´s maybe describe how to set up the Samba share within webmin to get access to the "big data partition" to access it via Windows. The default folder I used before was named "Download".

[deltaray]

I have installed Webmin from this source, edit your /etc/apt/sources.list and add:

deb http://download.webmin.com/download/repository sarge contrib

This will install at least webmin version 1.530

Not sure if this version can mount large partitions, however you can do it yourself rather easily.

I mounted md3 to my /home directory by adding this into /etc/fstab:

/dev/md3                /home           ext3    defaults                0       2

Once you have mounted the data partition on /home, you can easily create linux users and configure shares for them.

[Hajo]

eschumann,

it is not possible to use the vanilla Linux kernel sources. You have to use the Linux kernel with Marvell patch (MV) for device driver (networking ...) support.The sources are available from the Seagate website (GPL downloads). So far, I've compiled two extra modules (tun.ko for openvpn and iscsi_trgt.ko for iscsi, both are shipped with my SP).

If you want to recompile and change the complete kernel image, I strongly to suggest to use a serial cable and netboot the kernel/system via network (see my webpage on how to do that).

Hajo

[ArrowP]

This is amazing, really.

Before I proceed, did anyone try with encrypted drives?

I'm using RAID5 with encryption, so I wonder how i'm going to authenticate to mount the drives.

Regards,

A.

[Hajo]

ArrowP,

untested, but this may work (dm-crypt):

cryptsetup luksOpen /dev/mapper/vg0-lv0 lv0dmcrypt < /path/to/usbstick/keyfile

mount /dev/mapper/lv0dmcrypt /mnt/your/mount/point

Hajo

[ArrowP]

thanks, i just installed ssh.

i'll run a backup of my most important stuff and get things ready for a smooth transition