You are viewing limited content. For full access, please sign in.

Discussion

Discussion

Input on plan to correct initial LF implementation using physical volumes rather than logical volumes

posted on November 6, 2023 Show version history

Our LF volumes were set up as physical volumes instead of logical volumes.  This needs to be corrected due to a disk storage limitation.  Here is my plan (specifics removed).  I'd appreciate any feedback.  Thanks, Jim.

 

  Maximum amount of storage on the D drive on SERVER1 has brought forth the need to alter and formalize the physical storage configuration of our repositories.

Current configuration

SERVER1

  1. REPOSITORY1 repository
    1. DEFAULT logical volume
      1. Unused read-only volume left over from initial installation
    2. VOLUME1 physical volume
      1. Current location of all contents of the REPOSITORY1 repository
      2. Mapped to the D drive on SERVER1 and at over 80% capacity of the virtual drive
      3. Virtual drive is limited to 2TB in size
      4. Volume security is set to Everyone has default access
  2. REPOSITORY1DEV repository
    1. DEFAULT logical volume
      1. Mapped to the D drive on SERVER1 which is at >80% capacity as noted above
      2. Volume security is set to Everyone has default access

 

SERVER2

  1. REPOSITORY2 repository
    1. VOLUME3 physical volume
      1. Current default volume for the REPOSITORY2 repository except for “F1” folder.
      2. Mapped to the F drive on SERVER2
      3. Virtual drive is at 1% maximum capacity
      4. Volume security is set to Everyone has default access
    2. VOLUME4 physical volume
      1. Current default volume for the “F1” folder
      2. Mapped to the F drive on SERVER2
      3. Virtual drive is at 1% maximum capacity
      4. Volume security is set to Everyone has default access

 

Issues

  1. About to run out of disk space on SERVER1 D drive.
  2. Unable to expand VOLUME1 volume due to it being a physical volume on a space limited virtual disk drive.
  3. Volume security would allow for a user to attempt to create a document on a full, or nearly full, volume.
  4. Want to ensure that the new default volume is propagated to the entire repository.

 

Plan

  For REPOSITORY1 repository on SERVER1

  1. Create new logical volume for REPOSITORY1 repository called VOLUME2 with path E:\LF_Volumes\LF_VOLUME2.  This is pointing to the new GPT virtual drive.
  2. Make physical volumes in VOLUME2 rollover at 1,000,000 MB.  Increase size of the new GPT drive in 1 or 2 TB increments.  Current size of this new GPT drive is 2 TB, with a maximum value of 18 TB.  Note that changing the physical path of a logical volume will only affect the creating of new child physical volumes, therefore when the new GPT drive becomes full, pointing the logical volume to a new disk drive will only affect the creation of new physical volumes.  It would probably be best to leave extra space available on the existing drive to accommodate the expansion of existing documents.
  3. Set default volume for REPOSITORY1 repository to VOLUME2.
  4. Set volume VOLUME2 access to standard access for “Everyone” and complete access for “Admins”.
  5. Set volume VOLUME1 access to default minus “Create documents” for “Everyone” and complete access for “Admins”.  This will allow for updating existing documents “in place,” while new documents will be created on the new volume.  If free space on the SERVER1 D drive starts to approach maximum (2 TB) then first remove the DEV volumes from that drive.

 

  For REPOSITORY1DEV on SERVER1

  1. No change until such time as the SERVER1 D drive starts to become full.
    1. Alternatively, use development repository as a test bed for moving the contents to the new drive by changing the logical and then the physical fixed path.

 

  For REPOSITORY2 repository on SERVER2

  1. Add complete access to both volumes for “R2 Admin” group.
  2. Restrict access to VOLUME3 volume to only “R2 Admin” group.
  3. Currently, leave access to the VOLUME4 volume as is.  In the future when other entities are added to the REPOSITORY2 repository we can tailor the access to specific volumes for specific groups representing specific entities.
    1. Alternatively, modify access to VOLUME4 to default access for ENTITY1 group and eliminate “Everyone” access
  4. When new scanned images for ENTITY1 are to be added to the “F1” top-level folder of the REPOSITORY2 repository by attaching an exported volume, attach the volume as a temporary convenience and migrate the documents to the VOLUME4 volume.

 

Questions

  1. Does modifying the default volume at the repository level change the default volume for all child folders in the repository?  Can I reasonably presume that I do not need to change the default volume on every folder?  Hypothetically, if a child folder has a default volume that does not match the default volume for the repository, would changing the default volume for the repository change the default volume for the unmatched folder?
  2. If I change the path on a logical volume, does the complete path have to already exist, or will LF create the necessary folder structure?
0 0
replied on November 8, 2023 Show version history
0 0
replied on November 8, 2023

As an added note based on our experience with large volumes; it may still be ideal to limit the size of your new drive even with a logical volume.

For example, when making backups for disaster recovery, things can get to be problematic when you have everything in one massive drive.

We have always used logical volumes; however, we still limit each drive to 4TB to ensure our backups and other things run smoothly.

We set up alerts for when a drive reaches about 80%, at which point we create a new drive rather than expanding the existing drive.

Then, we update the path of the existing logical volume without having to create a new one because a logical volume can span multiple drives:

  • Existing folders/sub-volumes will be unaffected because their full path is set upon creation.
  • The next time the volume rolls over, it will use the new path/drive.
2 0
replied on June 5, 2024 Show version history

@████████

Following up on your advice, i'm planning something very similar to your volume setup.

You mentioned you're limiting your drives to 4TB.

What size is your physical volume rollover set to?
Edit just saw a different post where you state you set physical volumes to 20GB.

Also is there a limit to how many drives a logical volume should use?

Or does your logical volume grow indefinitely to the limits of the VM?

If not then what happens when the current logical volume is deemed too big? Create a new logical volume and start over from scratch?

 

Thank you!

0 0
replied on June 5, 2024

@████████

Our volume rollover is set to 20GB, and it is important to note that logical volumes are basically an automated way of creating/managing a bunch of smaller physical volumes, so they are a lot more flexible.

As a result, there's not really a limit on how big a logical volume can or should be because under the hood they are already broken into smaller volumes; it's all a matter of what fits your needs and use case.

A logical volume will continue to grow until it runs out of space on the mapped drive, but the major benefit of using logical volumes is that you can just point it to a new drive before that happens, even across multiple servers.

The mapping for the subvolumes is set when they are created, so if you start with drive E, then switch to drive F when E starts to fill up, the existing ones will still be mapped to E and only the new ones will start using the F path.

The change won't break your existing subvolumes and you can grow the logical volume across as many drives or servers as you want/need; I've never found any reason to limit them to a set number of drives or anything. 

Basically, the main reason to use Logical volumes is they don't have those sorts of set limitations, so you don't have to create new ones all the time.

The only real "issue" I can think of is that opening the list of physical volumes in a logical volume's settings can be very slow when you have a lot of them, but that's not something you check often, and our repository is exceptionally large so that probably wouldn't be an issue for most people.

1 0
You are not allowed to follow up in this post.

Sign in to reply to this post.