Moving from reiserfs to xfs


Recommended Posts

I am trying to move from ReiserFS to xfs on my date drives and am following 

http://lime-technology.com/wiki/index.php/File_System_Conversion#Mirroring_procedure_to_convert_drives

 

when I add a swap drive as described in the above it does not give me the option to format the drive, but states "Clearing in Progress" does this mean that every time I use one of the "reiserfs" drives as the new swap it is going to clear the drive first? 

 

Clearing a 3Tb drive will take ages never mind transferring the data then repeating, I have 15 drives to do this could take months!

Link to comment

You need to have an empty drive in the array to begin the migration process. If you add a drive that has not been precleared, unRaid will clear it. You will then format it to xfs, and copy the data from one of your rfs drives to.this new drive. Afterwards, you will format that rfs drive (the one you copied to the new disk) to xfs. (Formatting simply writes the xfs format, which unRaid reflects in parity.) You can then copy the next rfs drive to the disk that you just reformatted. Repeat over and over until all disks are copied and you'll be left with an empty xfs disk. You should never have to clear another disk.

Link to comment
  • 6 years later...
  • 2 weeks later...

Where is the procedure witten for us that do not want to mirror the disk (do not have that much extra space) but just want the easy option of converting and then let the parity rebuilds the disk even if that means rebuilding the parity itself. Basically the easy and the long option.

 

We need straigtforward instructions for this because there will be a lot of users suddenly of a need for this conversion from unsupported ReiserFS.

@ljm42 @SpencerJ @limetech

Edited by snoopy86
Link to comment
55 minutes ago, snoopy86 said:

but just want the easy option of converting and then let the parity rebuilds the disk even if that means rebuilding the parity itself.

Parity will rebuild the same filesystem as it was before, parity cannot help with filesystem conversion.

  • Upvote 1
Link to comment
2 minutes ago, JorgeB said:

Parity will rebuild the same filesystem as it was before, parity cannot help with filesystem conversion.

So you are saying there is no way to keep the files from that disk you want to convert if you don't have the space to copy those files somewhere before you do the format and conversion?

Link to comment
Just now, snoopy86 said:

So you are saying there is no way to keep the files from that disk you want to convert if you don't have the space to copy those files somewhere before you do the format and conversion?

That is correct, and always will be, unless for example the XFS maintainers create a tool for automatic conversion.

 

P.S. there is a reiser to btrfs conversion tool, but it can be risky.

Link to comment
2 hours ago, snoopy86 said:

We need straigtforward instructions for this because there will be a lot of users suddenly of a need for this conversion from unsupported ReiserFS.

 

My take on this situation.  Understand that this is not a decision of the Unraid development team of that the Linux Kernel Team to drop the ReiserFS drivers!   Unfortunately, there is no way around the problem that to change the format of any disk from ReiserFS to any other file system requires that it be formatted.   Furthermore, if that drive is a part of the array when it is formatted, parity is updated as the same time as the physical drive is so that any emulated disk is now empty!

 

For those folks who are now finding themselves in this problem should have been thinking about it since the announcement was made back in August of 2022.    I can think of a couple of reasons why a wait-and-see attitude might have been considered a proper course of action at the time:

  1. The Linux people may change their position.
  2. I really don't care if they do, I will just never upgrade my Unraid software beyond the last version that supports ReiserFS.   (I still see requests for support for Unraid ver5.X...)

As I see it, there will be no simple procedure for those folks who can't do the Mirroring procedure!   (If that existed, many people would have avoided the Mirror process as it is a complex operation!  I know this as I have done it twice.)  It is now a major 'system' problem that might involve some hard choices.

 

Some choices:

  1. Look at what is on the array and see if some data is outdated and no longer required.  Delete that data.
  2. Investigate getting a external USB drive (and the Unassigned Devices plugin) and move enough data off the array to empty the largest array disk by moving the remaining files to the other drives.
  3. Invest in the hardware to upgrade the system to allow the mirror process to be used.  For some folks, this might only involve getting a large case.  Others might have to purchase a number of new larger hard drives
  4. Kludge together a hardware setup to temporarily add another hard drive to the array.  When the conversion is done, physically remove the empty drive.  Do a New Configuration, rebuild parity.

There are undoubtedly other approaches but each one will have to be tailored to each different set of circumstances. 

 

I will say that @limetech should include a ReiserFS file system check before offering an Unraid version without its support for downloading to any Unraid server.

 

 

 

Edited by Frank1940
Link to comment
43 minutes ago, Frank1940 said:

 

My take on this situation.  Understand that this is not a decision of the Unraid development team of that the Linux Kernel Team to drop the ReiserFS drivers!   Unfortunately, there is no way around the problem that to change the format of any disk from ReiserFS to any other file system requires that it be formatted.   Furthermore, if that drive is a part of the array when it is formatted, parity is updated as the same time as the physical drive is so that any emulated disk is now empty!

 

For those folks who are now finding themselves in this problem should have been thinking about it since the announcement was made back in August of 2022.    I can think of a couple of reasons why a wait-and-see attitude might have been considered a proper course of action at the time:

  1. The Linux people may change their position.
  2. I really don't care if they do, I will just never upgrade my Unraid software beyond the last version that supports ReiserFS.   (I still see requests for support for Unraid ver5.X...)

As I see it, there will be no simple procedure for those folks who can't do the Mirroring procedure!   (If that existed, many people would have avoided the Mirror process as it is a complex operation!  I know this as I have done it twice.)  It is now a major 'system' problem that might involve some hard choices.

 

Some choices:

  1. Look at what is on the array and see if some data is outdated and no longer required.  Delete that data.
  2. Investigate getting a external USB drive (and the Unassigned Devices plugin) and move enough data off the array to empty the largest array disk by moving the remaining files to the other drives.
  3. Invest in the hardware to upgrade the system to allow the mirror process to be used.  For some folks, this might only involve getting a large case.  Others might have to purchase a number of new larger hard drives
  4. Kludge together a hardware setup to temporarily add another hard drive to the array.  When the conversion is done, physically remove the empty drive.  Do a New Configuration, rebuild parity.

There are undoubtedly other approaches but each one will have to be tailored to each different set of circumstances. 

 

I will say that @limetech should include a ReiserFS file system check before offering an Unraid version without its support for downloading to any Unraid server.

 

 

 

You have to know that we don't have only us professional users but also users that are not comfortable with that kind of procedures. The steps written in one of the links that everyone is posting are for sure overwhelming for those users and would need to be polished.

Link to comment
19 minutes ago, snoopy86 said:

You have to know that we don't have only us professional users but also users that are not comfortable with that kind of procedures. The steps written in one of the links that everyone is posting are for sure overwhelming for those users and would need to be polished.

 

So volunteer to help out.   (I have become old enough that my physical and mental states have started to limit what I can now do.  My forum name provides a clue as to my age...) 

Link to comment
3 minutes ago, Frank1940 said:

 

So volunteer to help out.   (I have become old enough that my physical and mental states have started to limit what I can now do.  My forum name provides a clue as to my age...) 

You are joking right? This is a topic where the company can't just push to the forum. This is not open source product. Of course people here will help but this is secondary help.

Link to comment
26 minutes ago, snoopy86 said:

You are joking right? This is a topic where the company can't just push to the forum. This is not open source product. Of course people here will help but this is secondary help.


 

The process is not really that complicated and is covered here in the online documentation accessible via the ‘Manual’ link at the bottom of the GUI or the DOCS link at the top of each forum page.  It is mainly a case of being methodical as one works through the drives to avoid accidents.

 

What is useful is suggestions on improving the documentation from those who have tried to use the documentation but find the current documentation insufficient or confusing.  
 

it might be worth Limetech starting to emphasise the need to move away from ReiserFS in release notes as I am sure that there are many users who have put off the conversion because they were short of time and did not consider it urgent.    However the time when reiserfs support is likely to be dropped from the Linux kernel is starting to get a lot nearer.

  • Upvote 1
Link to comment
  • 2 weeks later...

@itimpiOne thing i'm not clear is why can't we treat this change from REISERFS to XFS the same as a failed drive.

 

When the drive fails (let's say this drive is REISERFS 1GB) you remove it and change it with another drive (2GB). The parity then does it's magic and the drive gets populated again where you choose the filesystem. Can't the same process be done with this?

Link to comment
17 minutes ago, snoopy86 said:

The parity then does it's magic and the drive gets populated again where you choose the filesystem. Can't the same process be done with this?

Parity has no idea of what files or a filesystem is. All it does is rebuild a drive exactly as it was, i.e. it'll still be reiserfs.

 

Edited by Kilrah
Link to comment
1 hour ago, snoopy86 said:

The parity then does it's magic

 

More to the point-- It is not magic!  It a very simple 'grunt' type of operation.  When Parity information is used to rebuild a drive, the rebuild operation duplicates the drive on a bit-by-bit operation starting with the first sector on the drive until it finishes with the last sector on the drive.  If you think about what is happening, this means that the existing file system area is also rebuilt.  So if the Parity information says that the emulated disk had ReiserFS  before the parity rebuild started, the new physical disk will have ReiserFS when the parity rebuild operation finishes. Any file system on that physical drive will be overwritten!  Also see here:

 

https://docs.unraid.net/unraid-os/manual/what-is-unraid/#parity-protected-array 

Edited by Frank1940
Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.