External Drive online but not recognized
-
@Joel said:
We can access data whilst Old NAS is on but its 6 years old and the client (without telling us) ordered a new NAS, plugged in and configured it and then pulled drives out of old NAS (which is a ReadyNas 312) and were trying to copy the data from the old drive across to the new!!!
Then tell them that it is time to restore from backup. They have intentionally thrown their data away it sounds. You would expect the data to be destroyed by that process.
If the old NAS still works, they have to go back to it.
-
I feel like I've asked this before, Scott, but why does RAID 1 need to do anything funky at all on the drive?
Assuming that most NASs today run some flavor of Linux, wouldn't they run a Linux compatible format, and therefore be readable in a Linux machine?
I suppose just to make things more difficult, the NAS could use a proprietary file system, or at the mimimum write something to the starting sectors tying it to the NAS to create artificial lock in. Am I missing something?
-
Just to add- they didnt plug in the OLD drives into the new Nas box...its completely separate. The new NAS is up and running on the network with new drives...We're left with 2 old NAS hard drives. Shall i tell them to plug it all back to how it was and see if we can view data on the nas?
-
@Dashrender said:
I feel like I've asked this before, Scott, but why does RAID 1 need to do anything funky at all on the drive?
Assuming that most NASs today run some flavor of Linux, wouldn't they run a Linux compatible format, and therefore be readable in a Linux machine?
I suppose just to make things more difficult, the NAS could use a proprietary file system, or at the mimimum write something to the starting sectors tying it to the NAS to create artificial lock in. Am I missing something?
Thats what I was thinking - ie if we have a Linux compatible PC (mac) then we can read the data?
-
@Breffni-Potter said:
Your client is a genius
They might have already broken the RAID-1 data on the old drives.
Very likely. This process would be expected to destroy the old drives. Possible that it did not, but you have to assume that it did.
This is definitely a moment where you at least have the conversation of "should this client be fired?" I'm not saying that they should, but this is a pretty epic amount of... not talking to their MSP, trying to bypass IT, doing things that are way over their head, taking huge risks with storage of all things, doing things that should get an intern level IT and/or business person fired, etc.
I mean seriously. This is a huge screw up.
-
@Dashrender said:
I feel like I've asked this before, Scott, but why does RAID 1 need to do anything funky at all on the drive?
Abstraction. Just like any partitioning or similar. RAID is an abstraction layer. It is not invisible.
-
@Joel said:
Thats what I was thinking - ie if we have a Linux compatible PC (mac) then we can read the data?
Yes, since this is a ReadyNAS we can figure out what filesystem it has. If it is quite old, the ReadyNAS were EXT3 and then EXT4. I have that same older model myself (but not where I am at the moment.) How much they have patched it will be a factor.
Any modern Linux system can read the disks from a ReadyNAS. You will need to install something like Mint or Fedora on a desktop and attach it to that and go from there.
-
@Joel said:
Just to add- they didnt plug in the OLD drives into the new Nas box...its completely separate. The new NAS is up and running on the network with new drives...We're left with 2 old NAS hard drives. Shall i tell them to plug it all back to how it was and see if we can view data on the nas?
Why did they remove the drives from the NAS then? Just to break it? I'm confused. It sounds like they were doing nothing but breaking the NAS then?
Yes... just put the drives back and get the NAS fixed
Like I said, there is no legitimate process by which NAS drives should be attached to a computer. All good processes involve them remaining in the NAS. -
@Dashrender said:
Assuming that most NASs today run some flavor of Linux, wouldn't they run a Linux compatible format, and therefore be readable in a Linux machine?
Yes, IF you use the RAID layer. You need to mount the abstracted block device before you can get to the filesystem. Since a ReadyNAS thankfully leveraged MD RAID, any Linux system can do this. If they used their own RAID instead of MD, you'd be screwed.
-
@scottalanmiller said:
@Dashrender said:
I feel like I've asked this before, Scott, but why does RAID 1 need to do anything funky at all on the drive?
Abstraction. Just like any partitioning or similar. RAID is an abstraction layer. It is not invisible.
I definitely understand abstraction in RAID 10, 5, 6, etc, but with RAID 1 it seems like unnecessary overhead.
-
@Dashrender said:
I suppose just to make things more difficult, the NAS could use a proprietary file system, or at the mimimum write something to the starting sectors tying it to the NAS to create artificial lock in. Am I missing something?
If could use a proprietary filesystem and/or RAID system. Thankfully this is ReadyNAS and we know what it is doing under the hood.
If this was, for example, a NetApp NAS, we would know the RAID and the filesystem (RAID-DP and WAFL) but we would be screwed as nothing can read those except for another NetApp.
-
@Dashrender said:
I definitely understand abstraction in RAID 10, 5, 6, etc, but with RAID 1 it seems like unnecessary overhead.
Why? Why do you feel it needs to exist for the others but feel that RAID 1 is unique? Abstraction is abstraction. You have RAID or you do not.
Could you do RAID 1 without any abstraction on the disks themselves? Yes. Could you do it and still have the expected features like the disk itself knowing about its RAID? No.
-
@scottalanmiller said:
Any modern Linux system can read the disks from a ReadyNAS. You will need to install something like Mint or Fedora on a desktop and attach it to that and go from there.
Are these free? I assume I can download software which will then allow me to mount the drive (assuming its readable)?
-
@Joel said:
@scottalanmiller said:
Any modern Linux system can read the disks from a ReadyNAS. You will need to install something like Mint or Fedora on a desktop and attach it to that and go from there.
Are these free? I assume I can download software which will then allow me to mount the drive (assuming its readable)?
Yes, both are free (as are tons of others.) I use Mint as my main desktop every day. Fedora is super great too.
We might have to do some MD RAID trickery to get it to mount, but maybe not as this is all native for these.
Let's start with Mint 17.3 (Rosa) and see how far we get. You should not even need to install it. Burn to DVD and fire it up. It will run live. We don't need to install for what we need to do.
But FIRST... get those drives back in that NAS. Let's do this the "right way" before we fool around with something weird.
-
@Joel said:
I assume I can download software which will then allow me to mount the drive (assuming its readable)?
No extra software needed... this isn't Windows you know.
All of the power and features are built in. -
At this point, it may be best to drop the drives back in the old NAS unit and see if it fires up at all. If it does, just copy data from the old NAS to the new.
If it does not, then try to do it with Linux. In the worst case scenario: Dead Nas & no backups, Paragon also has some software that will read EXT partitions in windows. Windows may not recognize the RAID blocks, but Paragon's software might know how to deal with it.
-
@scottalanmiller said:
@Dashrender said:
I definitely understand abstraction in RAID 10, 5, 6, etc, but with RAID 1 it seems like unnecessary overhead.
Why? Why do you feel it needs to exist for the others but feel that RAID 1 is unique? Abstraction is abstraction. You have RAID or you do not.
Could you do RAID 1 without any abstraction on the disks themselves? Yes. Could you do it and still have the expected features like the disk itself knowing about its RAID? No.
In RAID 1 does that matter? I definitely understand it's need in the other RAIDs.
-
@Dashrender said:
@scottalanmiller said:
@Dashrender said:
I definitely understand abstraction in RAID 10, 5, 6, etc, but with RAID 1 it seems like unnecessary overhead.
Why? Why do you feel it needs to exist for the others but feel that RAID 1 is unique? Abstraction is abstraction. You have RAID or you do not.
Could you do RAID 1 without any abstraction on the disks themselves? Yes. Could you do it and still have the expected features like the disk itself knowing about its RAID? No.
In RAID 1 does that matter? I definitely understand it's need in the other RAIDs.
RAID has an abstraction. That you are using RAID 1 doesn't change that. RAID 1 is not a special case in any way.
Are you thinking that RAID 1 implementations specifically drop the RAID abstraction at the disk layer and act differently than all other RAID levels, even within a single implementation?
-
@scottalanmiller said:
@Dashrender said:
@scottalanmiller said:
@Dashrender said:
I definitely understand abstraction in RAID 10, 5, 6, etc, but with RAID 1 it seems like unnecessary overhead.
Why? Why do you feel it needs to exist for the others but feel that RAID 1 is unique? Abstraction is abstraction. You have RAID or you do not.
Could you do RAID 1 without any abstraction on the disks themselves? Yes. Could you do it and still have the expected features like the disk itself knowing about its RAID? No.
In RAID 1 does that matter? I definitely understand it's need in the other RAIDs.
RAID has an abstraction. That you are using RAID 1 doesn't change that. RAID 1 is not a special case in any way.
Are you thinking that RAID 1 implementations specifically drop the RAID abstraction at the disk layer and act differently than all other RAID levels, even within a single implementation?
This has been my personal experience. The few times I've needed two, I could just remove one drive and slave it up to another computer and assuming the file system was known you could mount it and it just worked.
Granted this was probably 10+ years ago.
-
@Dashrender said:
@scottalanmiller said:
@Dashrender said:
@scottalanmiller said:
@Dashrender said:
I definitely understand abstraction in RAID 10, 5, 6, etc, but with RAID 1 it seems like unnecessary overhead.
Why? Why do you feel it needs to exist for the others but feel that RAID 1 is unique? Abstraction is abstraction. You have RAID or you do not.
Could you do RAID 1 without any abstraction on the disks themselves? Yes. Could you do it and still have the expected features like the disk itself knowing about its RAID? No.
In RAID 1 does that matter? I definitely understand it's need in the other RAIDs.
RAID has an abstraction. That you are using RAID 1 doesn't change that. RAID 1 is not a special case in any way.
Are you thinking that RAID 1 implementations specifically drop the RAID abstraction at the disk layer and act differently than all other RAID levels, even within a single implementation?
This has been my personal experience. The few times I've needed two, I could just remove one drive and slave it up to another computer and assuming the file system was known you could mount it and it just worked.
Granted this was probably 10+ years ago.
What RAID system was that? Remember all things like this are RAID implementation specific. There is nothing generic with RAID except for what the levels mean.
